mirror of
https://github.com/langgenius/dify.git
synced 2026-04-14 07:56:31 +08:00
Merge branch 'main' into e-300
This commit is contained in:
NFish
commit
0301bd3ac1
@ -34,4 +34,4 @@ if you see such error message when you open this project in codespaces:
|
||||
|
||||
|
||||
a simple workaround is change `/signin` endpoint into another one, then login with GitHub account and close the tab, then change it back to `/signin` endpoint. Then all things will be fine.
|
||||
The reason is `signin` endpoint is not allowed in codespaces, details can be found [here](https://github.com/orgs/community/discussions/5204)
|
||||
The reason is `signin` endpoint is not allowed in codespaces, details can be found [here](https://github.com/orgs/community/discussions/5204)
|
||||
|
||||
@ -2,7 +2,7 @@
|
||||
// README at: https://github.com/devcontainers/templates/tree/main/src/anaconda
|
||||
{
|
||||
"name": "Python 3.12",
|
||||
"build": {
|
||||
"build": {
|
||||
"context": "..",
|
||||
"dockerfile": "Dockerfile"
|
||||
},
|
||||
|
||||
@ -1,3 +1,3 @@
|
||||
This file copied into the container along with environment.yml* from the parent
|
||||
folder. This file is included to prevents the Dockerfile COPY instruction from
|
||||
failing if no environment.yml is found.
|
||||
folder. This file is included to prevents the Dockerfile COPY instruction from
|
||||
failing if no environment.yml is found.
|
||||
|
||||
@ -5,18 +5,35 @@ root = true
|
||||
|
||||
# Unix-style newlines with a newline ending every file
|
||||
[*]
|
||||
charset = utf-8
|
||||
end_of_line = lf
|
||||
insert_final_newline = true
|
||||
trim_trailing_whitespace = true
|
||||
|
||||
[*.py]
|
||||
indent_size = 4
|
||||
indent_style = space
|
||||
|
||||
[*.{yml,yaml}]
|
||||
indent_style = space
|
||||
indent_size = 2
|
||||
|
||||
[*.toml]
|
||||
indent_size = 4
|
||||
indent_style = space
|
||||
|
||||
# Markdown and MDX are whitespace sensitive languages.
|
||||
# Do not remove trailing spaces.
|
||||
[*.{md,mdx}]
|
||||
trim_trailing_whitespace = false
|
||||
|
||||
# Matches multiple files with brace expansion notation
|
||||
# Set default charset
|
||||
[*.{js,tsx}]
|
||||
charset = utf-8
|
||||
indent_style = space
|
||||
indent_size = 2
|
||||
|
||||
|
||||
# Matches the exact files either package.json or .travis.yml
|
||||
[{package.json,.travis.yml}]
|
||||
# Matches the exact files package.json
|
||||
[package.json]
|
||||
indent_style = space
|
||||
indent_size = 2
|
||||
2
.gitattributes
vendored
2
.gitattributes
vendored
@ -1,5 +1,5 @@
|
||||
# Ensure that .sh scripts use LF as line separator, even if they are checked out
|
||||
# to Windows(NTFS) file-system, by a user of Docker for Windows.
|
||||
# to Windows(NTFS) file-system, by a user of Docker for Windows.
|
||||
# These .sh scripts will be run from the Container after `docker compose up -d`.
|
||||
# If they appear to be CRLF style, Dash from the Container will fail to execute
|
||||
# them.
|
||||
|
||||
22
.github/linters/editorconfig-checker.json
vendored
Normal file
22
.github/linters/editorconfig-checker.json
vendored
Normal file
@ -0,0 +1,22 @@
|
||||
{
|
||||
"Verbose": false,
|
||||
"Debug": false,
|
||||
"IgnoreDefaults": false,
|
||||
"SpacesAfterTabs": false,
|
||||
"NoColor": false,
|
||||
"Exclude": [
|
||||
"^web/public/vs/",
|
||||
"^web/public/pdf.worker.min.mjs$",
|
||||
"web/app/components/base/icons/src/vender/"
|
||||
],
|
||||
"AllowedContentTypes": [],
|
||||
"PassedFiles": [],
|
||||
"Disable": {
|
||||
"EndOfLine": false,
|
||||
"Indentation": false,
|
||||
"IndentSize": true,
|
||||
"InsertFinalNewline": false,
|
||||
"TrimTrailingWhitespace": false,
|
||||
"MaxLineLength": false
|
||||
}
|
||||
}
|
||||
3
.github/workflows/api-tests.yml
vendored
3
.github/workflows/api-tests.yml
vendored
@ -88,3 +88,6 @@ jobs:
|
||||
|
||||
- name: Run Workflow
|
||||
run: uv run --project api bash dev/pytest/pytest_workflow.sh
|
||||
|
||||
- name: Run Tool
|
||||
run: uv run --project api bash dev/pytest/pytest_tools.sh
|
||||
|
||||
21
.github/workflows/style.yml
vendored
21
.github/workflows/style.yml
vendored
@ -9,6 +9,12 @@ concurrency:
|
||||
group: style-${{ github.head_ref || github.run_id }}
|
||||
cancel-in-progress: true
|
||||
|
||||
permissions:
|
||||
checks: write
|
||||
statuses: write
|
||||
contents: read
|
||||
|
||||
|
||||
jobs:
|
||||
python-style:
|
||||
name: Python Style
|
||||
@ -43,8 +49,8 @@ jobs:
|
||||
if: steps.changed-files.outputs.any_changed == 'true'
|
||||
run: |
|
||||
uv run --directory api ruff --version
|
||||
uv run --directory api ruff check ./
|
||||
uv run --directory api ruff format --check ./
|
||||
uv run --directory api ruff check --diff ./
|
||||
uv run --directory api ruff format --check --diff ./
|
||||
|
||||
- name: Dotenv check
|
||||
if: steps.changed-files.outputs.any_changed == 'true'
|
||||
@ -163,3 +169,14 @@ jobs:
|
||||
VALIDATE_DOCKERFILE_HADOLINT: true
|
||||
VALIDATE_XML: true
|
||||
VALIDATE_YAML: true
|
||||
|
||||
- name: EditorConfig checks
|
||||
uses: super-linter/super-linter/slim@v7
|
||||
env:
|
||||
DEFAULT_BRANCH: main
|
||||
GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
|
||||
IGNORE_GENERATED_FILES: true
|
||||
IGNORE_GITIGNORED_FILES: true
|
||||
# EditorConfig validation
|
||||
VALIDATE_EDITORCONFIG: true
|
||||
EDITORCONFIG_FILE_NAME: editorconfig-checker.json
|
||||
|
||||
@ -90,4 +90,4 @@ Recomendamos revisar este documento cuidadosamente antes de proceder con la conf
|
||||
No dudes en contactarnos si encuentras algún problema durante el proceso de configuración.
|
||||
## Obteniendo Ayuda
|
||||
|
||||
Si alguna vez te quedas atascado o tienes una pregunta urgente mientras contribuyes, simplemente envíanos tus consultas a través del issue relacionado de GitHub, o únete a nuestro [Discord](https://discord.gg/8Tpq4AcN9c) para una charla rápida.
|
||||
Si alguna vez te quedas atascado o tienes una pregunta urgente mientras contribuyes, simplemente envíanos tus consultas a través del issue relacionado de GitHub, o únete a nuestro [Discord](https://discord.gg/8Tpq4AcN9c) para una charla rápida.
|
||||
|
||||
@ -90,4 +90,4 @@ Nous recommandons de revoir attentivement ce document avant de procéder à la c
|
||||
N'hésitez pas à nous contacter si vous rencontrez des problèmes pendant le processus de configuration.
|
||||
## Obtenir de l'aide
|
||||
|
||||
Si jamais vous êtes bloqué ou avez une question urgente en contribuant, envoyez-nous simplement vos questions via le problème GitHub concerné, ou rejoignez notre [Discord](https://discord.gg/8Tpq4AcN9c) pour une discussion rapide.
|
||||
Si jamais vous êtes bloqué ou avez une question urgente en contribuant, envoyez-nous simplement vos questions via le problème GitHub concerné, ou rejoignez notre [Discord](https://discord.gg/8Tpq4AcN9c) pour une discussion rapide.
|
||||
|
||||
@ -90,4 +90,4 @@ PR 설명에 기존 이슈를 연결하거나 새 이슈를 여는 것을 잊지
|
||||
설정 과정에서 문제가 발생하면 언제든지 연락해 주세요.
|
||||
## 도움 받기
|
||||
|
||||
기여하는 동안 막히거나 긴급한 질문이 있으면, 관련 GitHub 이슈를 통해 질문을 보내거나, 빠른 대화를 위해 우리의 [Discord](https://discord.gg/8Tpq4AcN9c)에 참여하세요.
|
||||
기여하는 동안 막히거나 긴급한 질문이 있으면, 관련 GitHub 이슈를 통해 질문을 보내거나, 빠른 대화를 위해 우리의 [Discord](https://discord.gg/8Tpq4AcN9c)에 참여하세요.
|
||||
|
||||
@ -90,4 +90,4 @@ Recomendamos revisar este documento cuidadosamente antes de prosseguir com a con
|
||||
Sinta-se à vontade para entrar em contato se encontrar quaisquer problemas durante o processo de configuração.
|
||||
## Obtendo Ajuda
|
||||
|
||||
Se você ficar preso ou tiver uma dúvida urgente enquanto contribui, simplesmente envie suas perguntas através do problema relacionado no GitHub, ou entre no nosso [Discord](https://discord.gg/8Tpq4AcN9c) para uma conversa rápida.
|
||||
Se você ficar preso ou tiver uma dúvida urgente enquanto contribui, simplesmente envie suas perguntas através do problema relacionado no GitHub, ou entre no nosso [Discord](https://discord.gg/8Tpq4AcN9c) para uma conversa rápida.
|
||||
|
||||
@ -90,4 +90,4 @@ Kuruluma geçmeden önce bu belgeyi dikkatlice incelemenizi öneririz, çünkü
|
||||
Kurulum süreci sırasında herhangi bir sorunla karşılaşırsanız bizimle iletişime geçmekten çekinmeyin.
|
||||
## Yardım Almak
|
||||
|
||||
Katkıda bulunurken takılırsanız veya yanıcı bir sorunuz olursa, sorularınızı ilgili GitHub sorunu aracılığıyla bize gönderin veya hızlı bir sohbet için [Discord'umuza](https://discord.gg/8Tpq4AcN9c) katılın.
|
||||
Katkıda bulunurken takılırsanız veya yanıcı bir sorunuz olursa, sorularınızı ilgili GitHub sorunu aracılığıyla bize gönderin veya hızlı bir sohbet için [Discord'umuza](https://discord.gg/8Tpq4AcN9c) katılın.
|
||||
|
||||
518
README_SI.md
518
README_SI.md
@ -1,259 +1,259 @@
|
||||
|
||||
|
||||
<p align="center">
|
||||
📌 <a href="https://dify.ai/blog/introducing-dify-workflow-file-upload-a-demo-on-ai-podcast">Predstavljamo nalaganje datotek Dify Workflow: znova ustvarite Google NotebookLM Podcast</a>
|
||||
</p>
|
||||
|
||||
<p align="center">
|
||||
<a href="https://cloud.dify.ai">Dify Cloud</a> ·
|
||||
<a href="https://docs.dify.ai/getting-started/install-self-hosted">Samostojno gostovanje</a> ·
|
||||
<a href="https://docs.dify.ai">Dokumentacija</a> ·
|
||||
<a href="https://dify.ai/pricing">Pregled ponudb izdelkov Dify</a>
|
||||
</p>
|
||||
|
||||
<p align="center">
|
||||
<a href="https://dify.ai" target="_blank">
|
||||
<img alt="Static Badge" src="https://img.shields.io/badge/Product-F04438"></a>
|
||||
<a href="https://dify.ai/pricing" target="_blank">
|
||||
<img alt="Static Badge" src="https://img.shields.io/badge/free-pricing?logo=free&color=%20%23155EEF&label=pricing&labelColor=%20%23528bff"></a>
|
||||
<a href="https://discord.gg/FngNHpbcY7" target="_blank">
|
||||
<img src="https://img.shields.io/discord/1082486657678311454?logo=discord&labelColor=%20%235462eb&logoColor=%20%23f5f5f5&color=%20%235462eb"
|
||||
alt="chat on Discord"></a>
|
||||
<a href="https://twitter.com/intent/follow?screen_name=dify_ai" target="_blank">
|
||||
<img src="https://img.shields.io/twitter/follow/dify_ai?logo=X&color=%20%23f5f5f5"
|
||||
alt="follow on X(Twitter)"></a>
|
||||
<a href="https://www.linkedin.com/company/langgenius/" target="_blank">
|
||||
<img src="https://custom-icon-badges.demolab.com/badge/LinkedIn-0A66C2?logo=linkedin-white&logoColor=fff"
|
||||
alt="follow on LinkedIn"></a>
|
||||
<a href="https://hub.docker.com/u/langgenius" target="_blank">
|
||||
<img alt="Docker Pulls" src="https://img.shields.io/docker/pulls/langgenius/dify-web?labelColor=%20%23FDB062&color=%20%23f79009"></a>
|
||||
<a href="https://github.com/langgenius/dify/graphs/commit-activity" target="_blank">
|
||||
<img alt="Commits last month" src="https://img.shields.io/github/commit-activity/m/langgenius/dify?labelColor=%20%2332b583&color=%20%2312b76a"></a>
|
||||
<a href="https://github.com/langgenius/dify/" target="_blank">
|
||||
<img alt="Issues closed" src="https://img.shields.io/github/issues-search?query=repo%3Alanggenius%2Fdify%20is%3Aclosed&label=issues%20closed&labelColor=%20%237d89b0&color=%20%235d6b98"></a>
|
||||
<a href="https://github.com/langgenius/dify/discussions/" target="_blank">
|
||||
<img alt="Discussion posts" src="https://img.shields.io/github/discussions/langgenius/dify?labelColor=%20%239b8afb&color=%20%237a5af8"></a>
|
||||
</p>
|
||||
|
||||
<p align="center">
|
||||
<a href="./README.md"><img alt="README in English" src="https://img.shields.io/badge/English-d9d9d9"></a>
|
||||
<a href="./README_CN.md"><img alt="简体中文版自述文件" src="https://img.shields.io/badge/简体中文-d9d9d9"></a>
|
||||
<a href="./README_JA.md"><img alt="日本語のREADME" src="https://img.shields.io/badge/日本語-d9d9d9"></a>
|
||||
<a href="./README_ES.md"><img alt="README en Español" src="https://img.shields.io/badge/Español-d9d9d9"></a>
|
||||
<a href="./README_FR.md"><img alt="README en Français" src="https://img.shields.io/badge/Français-d9d9d9"></a>
|
||||
<a href="./README_KL.md"><img alt="README tlhIngan Hol" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
|
||||
<a href="./README_KR.md"><img alt="README in Korean" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
|
||||
<a href="./README_AR.md"><img alt="README بالعربية" src="https://img.shields.io/badge/العربية-d9d9d9"></a>
|
||||
<a href="./README_TR.md"><img alt="Türkçe README" src="https://img.shields.io/badge/Türkçe-d9d9d9"></a>
|
||||
<a href="./README_VI.md"><img alt="README Tiếng Việt" src="https://img.shields.io/badge/Ti%E1%BA%BFng%20Vi%E1%BB%87t-d9d9d9"></a>
|
||||
<a href="./README_SI.md"><img alt="README Slovenščina" src="https://img.shields.io/badge/Sloven%C5%A1%C4%8Dina-d9d9d9"></a>
|
||||
<a href="./README_BN.md"><img alt="README in বাংলা" src="https://img.shields.io/badge/বাংলা-d9d9d9"></a>
|
||||
</p>
|
||||
|
||||
|
||||
Dify je odprtokodna platforma za razvoj aplikacij LLM. Njegov intuitivni vmesnik združuje agentski potek dela z umetno inteligenco, cevovod RAG, zmogljivosti agentov, upravljanje modelov, funkcije opazovanja in več, kar vam omogoča hiter prehod od prototipa do proizvodnje.
|
||||
|
||||
## Hitri začetek
|
||||
> Preden namestite Dify, se prepričajte, da vaša naprava izpolnjuje naslednje minimalne sistemske zahteve:
|
||||
>
|
||||
>- CPU >= 2 Core
|
||||
>- RAM >= 4 GiB
|
||||
|
||||
</br>
|
||||
|
||||
Najlažji način za zagon strežnika Dify je prek docker compose . Preden zaženete Dify z naslednjimi ukazi, se prepričajte, da sta Docker in Docker Compose nameščena na vašem računalniku:
|
||||
|
||||
```bash
|
||||
cd dify
|
||||
cd docker
|
||||
cp .env.example .env
|
||||
docker compose up -d
|
||||
```
|
||||
|
||||
Po zagonu lahko dostopate do nadzorne plošče Dify v brskalniku na [http://localhost/install](http://localhost/install) in začnete postopek inicializacije.
|
||||
|
||||
#### Iskanje pomoči
|
||||
Prosimo, glejte naša pogosta vprašanja [FAQ](https://docs.dify.ai/getting-started/install-self-hosted/faqs) če naletite na težave pri nastavitvi Dify. Če imate še vedno težave, se obrnite na [skupnost ali nas](#community--contact).
|
||||
|
||||
> Če želite prispevati k Difyju ali narediti dodaten razvoj, glejte naš vodnik za [uvajanje iz izvorne kode](https://docs.dify.ai/getting-started/install-self-hosted/local-source-code)
|
||||
|
||||
## Ključne značilnosti
|
||||
**1. Potek dela**:
|
||||
Zgradite in preizkusite zmogljive poteke dela AI na vizualnem platnu, pri čemer izkoristite vse naslednje funkcije in več.
|
||||
|
||||
|
||||
https://github.com/langgenius/dify/assets/13230914/356df23e-1604-483d-80a6-9517ece318aa
|
||||
|
||||
|
||||
|
||||
**2. Celovita podpora za modele**:
|
||||
Brezhibna integracija s stotinami lastniških/odprtokodnih LLM-jev ducatov ponudnikov sklepanja in samostojnih rešitev, ki pokrivajo GPT, Mistral, Llama3 in vse modele, združljive z API-jem OpenAI. Celoten seznam podprtih ponudnikov modelov najdete [tukaj](https://docs.dify.ai/getting-started/readme/model-providers).
|
||||
|
||||
|
||||
|
||||
|
||||
**3. Prompt IDE**:
|
||||
intuitivni vmesnik za ustvarjanje pozivov, primerjavo zmogljivosti modela in dodajanje dodatnih funkcij, kot je pretvorba besedila v govor, aplikaciji, ki temelji na klepetu.
|
||||
|
||||
**4. RAG Pipeline**:
|
||||
E Obsežne zmogljivosti RAG, ki pokrivajo vse od vnosa dokumenta do priklica, s podporo za ekstrakcijo besedila iz datotek PDF, PPT in drugih običajnih formatov dokumentov.
|
||||
|
||||
**5. Agent capabilities**:
|
||||
definirate lahko agente, ki temeljijo na klicanju funkcij LLM ali ReAct, in dodate vnaprej izdelana orodja ali orodja po meri za agenta. Dify ponuja več kot 50 vgrajenih orodij za agente AI, kot so Google Search, DALL·E, Stable Diffusion in WolframAlpha.
|
||||
|
||||
**6. LLMOps**:
|
||||
Spremljajte in analizirajte dnevnike aplikacij in učinkovitost skozi čas. Pozive, nabore podatkov in modele lahko nenehno izboljšujete na podlagi proizvodnih podatkov in opomb.
|
||||
|
||||
**7. Backend-as-a-Service**:
|
||||
AVse ponudbe Difyja so opremljene z ustreznimi API-ji, tako da lahko Dify brez težav integrirate v svojo poslovno logiko.
|
||||
|
||||
## Primerjava Funkcij
|
||||
|
||||
<table style="width: 100%;">
|
||||
<tr>
|
||||
<th align="center">Funkcija</th>
|
||||
<th align="center">Dify.AI</th>
|
||||
<th align="center">LangChain</th>
|
||||
<th align="center">Flowise</th>
|
||||
<th align="center">OpenAI Assistants API</th>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Programski pristop</td>
|
||||
<td align="center">API + usmerjeno v aplikacije</td>
|
||||
<td align="center">Python koda</td>
|
||||
<td align="center">Usmerjeno v aplikacije</td>
|
||||
<td align="center">Usmerjeno v API</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Podprti LLM-ji</td>
|
||||
<td align="center">Bogata izbira</td>
|
||||
<td align="center">Bogata izbira</td>
|
||||
<td align="center">Bogata izbira</td>
|
||||
<td align="center">Samo OpenAI</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">RAG pogon</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Agent</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
<td align="center">✅</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Potek dela</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Spremljanje</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
<td align="center">❌</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Funkcija za podjetja (SSO/nadzor dostopa)</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
<td align="center">❌</td>
|
||||
<td align="center">❌</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Lokalna namestitev</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
</tr>
|
||||
</table>
|
||||
|
||||
## Uporaba Dify
|
||||
|
||||
- **Cloud </br>**
|
||||
Gostimo storitev Dify Cloud za vsakogar, ki jo lahko preizkusite brez nastavitev. Zagotavlja vse zmožnosti različice za samostojno namestitev in vključuje 200 brezplačnih klicev GPT-4 v načrtu peskovnika.
|
||||
|
||||
- **Self-hosting Dify Community Edition</br>**
|
||||
Hitro zaženite Dify v svojem okolju s tem [začetnim vodnikom](#quick-start) . Za dodatne reference in podrobnejša navodila uporabite našo [dokumentacijo](https://docs.dify.ai) .
|
||||
|
||||
|
||||
- **Dify za podjetja/organizacije</br>**
|
||||
Ponujamo dodatne funkcije, osredotočene na podjetja. Zabeležite svoja vprašanja prek tega klepetalnega robota ali nam pošljite e-pošto, da se pogovorimo o potrebah podjetja. </br>
|
||||
> Za novoustanovljena podjetja in mala podjetja, ki uporabljajo AWS, si oglejte Dify Premium na AWS Marketplace in ga z enim klikom uvedite v svoj AWS VPC. To je cenovno ugodna ponudba AMI z možnostjo ustvarjanja aplikacij z logotipom in blagovno znamko po meri.
|
||||
|
||||
|
||||
## Staying ahead
|
||||
|
||||
Star Dify on GitHub and be instantly notified of new releases.
|
||||
|
||||
|
||||
|
||||
|
||||
## Napredne nastavitve
|
||||
|
||||
Če morate prilagoditi konfiguracijo, si oglejte komentarje v naši datoteki .env.example in posodobite ustrezne vrednosti v svoji .env datoteki. Poleg tega boste morda morali prilagoditi docker-compose.yamlsamo datoteko, na primer spremeniti različice slike, preslikave vrat ali namestitve nosilca, glede na vaše specifično okolje in zahteve za uvajanje. Po kakršnih koli spremembah ponovno zaženite docker-compose up -d. Celoten seznam razpoložljivih spremenljivk okolja najdete tukaj .
|
||||
|
||||
Če želite konfigurirati visoko razpoložljivo nastavitev, so na voljo Helm Charts in datoteke YAML, ki jih prispeva skupnost, ki omogočajo uvedbo Difyja v Kubernetes.
|
||||
|
||||
- [Helm Chart by @LeoQuote](https://github.com/douban/charts/tree/master/charts/dify)
|
||||
- [Helm Chart by @BorisPolonsky](https://github.com/BorisPolonsky/dify-helm)
|
||||
- [YAML file by @Winson-030](https://github.com/Winson-030/dify-kubernetes)
|
||||
- [YAML file by @wyy-holding](https://github.com/wyy-holding/dify-k8s)
|
||||
|
||||
#### Uporaba Terraform za uvajanje
|
||||
|
||||
namestite Dify v Cloud Platform z enim klikom z uporabo [terraform](https://www.terraform.io/)
|
||||
|
||||
##### Azure Global
|
||||
- [Azure Terraform by @nikawang](https://github.com/nikawang/dify-azure-terraform)
|
||||
|
||||
##### Google Cloud
|
||||
- [Google Cloud Terraform by @sotazum](https://github.com/DeNA/dify-google-cloud-terraform)
|
||||
|
||||
#### Uporaba AWS CDK za uvajanje
|
||||
|
||||
Uvedite Dify v AWS z uporabo [CDK](https://aws.amazon.com/cdk/)
|
||||
|
||||
##### AWS
|
||||
- [AWS CDK by @KevinZhao](https://github.com/aws-samples/solution-for-deploying-dify-on-aws)
|
||||
|
||||
## Prispevam
|
||||
|
||||
Za tiste, ki bi radi prispevali kodo, si oglejte naš vodnik za prispevke . Hkrati vas prosimo, da podprete Dify tako, da ga delite na družbenih medijih ter na dogodkih in konferencah.
|
||||
|
||||
|
||||
|
||||
> Iščemo sodelavce za pomoč pri prevajanju Difyja v jezike, ki niso mandarinščina ali angleščina. Če želite pomagati, si oglejte i18n README za več informacij in nam pustite komentar v global-userskanalu našega strežnika skupnosti Discord .
|
||||
|
||||
## Skupnost in stik
|
||||
|
||||
* [Github Discussion](https://github.com/langgenius/dify/discussions). Najboljše za: izmenjavo povratnih informacij in postavljanje vprašanj.
|
||||
* [GitHub Issues](https://github.com/langgenius/dify/issues). Najboljše za: hrošče, na katere naletite pri uporabi Dify.AI, in predloge funkcij. Oglejte si naš [vodnik za prispevke](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md).
|
||||
* [Discord](https://discord.gg/FngNHpbcY7). Najboljše za: deljenje vaših aplikacij in druženje s skupnostjo.
|
||||
* [X(Twitter)](https://twitter.com/dify_ai). Najboljše za: deljenje vaših aplikacij in druženje s skupnostjo.
|
||||
|
||||
**Contributors**
|
||||
|
||||
<a href="https://github.com/langgenius/dify/graphs/contributors">
|
||||
<img src="https://contrib.rocks/image?repo=langgenius/dify" />
|
||||
</a>
|
||||
|
||||
## Star history
|
||||
|
||||
[](https://star-history.com/#langgenius/dify&Date)
|
||||
|
||||
|
||||
## Varnostno razkritje
|
||||
|
||||
Zaradi zaščite vaše zasebnosti se izogibajte objavljanju varnostnih vprašanj na GitHub. Namesto tega pošljite vprašanja na security@dify.ai in zagotovili vam bomo podrobnejši odgovor.
|
||||
|
||||
## Licenca
|
||||
|
||||
To skladišče je na voljo pod [odprtokodno licenco Dify](LICENSE) , ki je v bistvu Apache 2.0 z nekaj dodatnimi omejitvami.
|
||||
|
||||
|
||||
<p align="center">
|
||||
📌 <a href="https://dify.ai/blog/introducing-dify-workflow-file-upload-a-demo-on-ai-podcast">Predstavljamo nalaganje datotek Dify Workflow: znova ustvarite Google NotebookLM Podcast</a>
|
||||
</p>
|
||||
|
||||
<p align="center">
|
||||
<a href="https://cloud.dify.ai">Dify Cloud</a> ·
|
||||
<a href="https://docs.dify.ai/getting-started/install-self-hosted">Samostojno gostovanje</a> ·
|
||||
<a href="https://docs.dify.ai">Dokumentacija</a> ·
|
||||
<a href="https://dify.ai/pricing">Pregled ponudb izdelkov Dify</a>
|
||||
</p>
|
||||
|
||||
<p align="center">
|
||||
<a href="https://dify.ai" target="_blank">
|
||||
<img alt="Static Badge" src="https://img.shields.io/badge/Product-F04438"></a>
|
||||
<a href="https://dify.ai/pricing" target="_blank">
|
||||
<img alt="Static Badge" src="https://img.shields.io/badge/free-pricing?logo=free&color=%20%23155EEF&label=pricing&labelColor=%20%23528bff"></a>
|
||||
<a href="https://discord.gg/FngNHpbcY7" target="_blank">
|
||||
<img src="https://img.shields.io/discord/1082486657678311454?logo=discord&labelColor=%20%235462eb&logoColor=%20%23f5f5f5&color=%20%235462eb"
|
||||
alt="chat on Discord"></a>
|
||||
<a href="https://twitter.com/intent/follow?screen_name=dify_ai" target="_blank">
|
||||
<img src="https://img.shields.io/twitter/follow/dify_ai?logo=X&color=%20%23f5f5f5"
|
||||
alt="follow on X(Twitter)"></a>
|
||||
<a href="https://www.linkedin.com/company/langgenius/" target="_blank">
|
||||
<img src="https://custom-icon-badges.demolab.com/badge/LinkedIn-0A66C2?logo=linkedin-white&logoColor=fff"
|
||||
alt="follow on LinkedIn"></a>
|
||||
<a href="https://hub.docker.com/u/langgenius" target="_blank">
|
||||
<img alt="Docker Pulls" src="https://img.shields.io/docker/pulls/langgenius/dify-web?labelColor=%20%23FDB062&color=%20%23f79009"></a>
|
||||
<a href="https://github.com/langgenius/dify/graphs/commit-activity" target="_blank">
|
||||
<img alt="Commits last month" src="https://img.shields.io/github/commit-activity/m/langgenius/dify?labelColor=%20%2332b583&color=%20%2312b76a"></a>
|
||||
<a href="https://github.com/langgenius/dify/" target="_blank">
|
||||
<img alt="Issues closed" src="https://img.shields.io/github/issues-search?query=repo%3Alanggenius%2Fdify%20is%3Aclosed&label=issues%20closed&labelColor=%20%237d89b0&color=%20%235d6b98"></a>
|
||||
<a href="https://github.com/langgenius/dify/discussions/" target="_blank">
|
||||
<img alt="Discussion posts" src="https://img.shields.io/github/discussions/langgenius/dify?labelColor=%20%239b8afb&color=%20%237a5af8"></a>
|
||||
</p>
|
||||
|
||||
<p align="center">
|
||||
<a href="./README.md"><img alt="README in English" src="https://img.shields.io/badge/English-d9d9d9"></a>
|
||||
<a href="./README_CN.md"><img alt="简体中文版自述文件" src="https://img.shields.io/badge/简体中文-d9d9d9"></a>
|
||||
<a href="./README_JA.md"><img alt="日本語のREADME" src="https://img.shields.io/badge/日本語-d9d9d9"></a>
|
||||
<a href="./README_ES.md"><img alt="README en Español" src="https://img.shields.io/badge/Español-d9d9d9"></a>
|
||||
<a href="./README_FR.md"><img alt="README en Français" src="https://img.shields.io/badge/Français-d9d9d9"></a>
|
||||
<a href="./README_KL.md"><img alt="README tlhIngan Hol" src="https://img.shields.io/badge/Klingon-d9d9d9"></a>
|
||||
<a href="./README_KR.md"><img alt="README in Korean" src="https://img.shields.io/badge/한국어-d9d9d9"></a>
|
||||
<a href="./README_AR.md"><img alt="README بالعربية" src="https://img.shields.io/badge/العربية-d9d9d9"></a>
|
||||
<a href="./README_TR.md"><img alt="Türkçe README" src="https://img.shields.io/badge/Türkçe-d9d9d9"></a>
|
||||
<a href="./README_VI.md"><img alt="README Tiếng Việt" src="https://img.shields.io/badge/Ti%E1%BA%BFng%20Vi%E1%BB%87t-d9d9d9"></a>
|
||||
<a href="./README_SI.md"><img alt="README Slovenščina" src="https://img.shields.io/badge/Sloven%C5%A1%C4%8Dina-d9d9d9"></a>
|
||||
<a href="./README_BN.md"><img alt="README in বাংলা" src="https://img.shields.io/badge/বাংলা-d9d9d9"></a>
|
||||
</p>
|
||||
|
||||
|
||||
Dify je odprtokodna platforma za razvoj aplikacij LLM. Njegov intuitivni vmesnik združuje agentski potek dela z umetno inteligenco, cevovod RAG, zmogljivosti agentov, upravljanje modelov, funkcije opazovanja in več, kar vam omogoča hiter prehod od prototipa do proizvodnje.
|
||||
|
||||
## Hitri začetek
|
||||
> Preden namestite Dify, se prepričajte, da vaša naprava izpolnjuje naslednje minimalne sistemske zahteve:
|
||||
>
|
||||
>- CPU >= 2 Core
|
||||
>- RAM >= 4 GiB
|
||||
|
||||
</br>
|
||||
|
||||
Najlažji način za zagon strežnika Dify je prek docker compose . Preden zaženete Dify z naslednjimi ukazi, se prepričajte, da sta Docker in Docker Compose nameščena na vašem računalniku:
|
||||
|
||||
```bash
|
||||
cd dify
|
||||
cd docker
|
||||
cp .env.example .env
|
||||
docker compose up -d
|
||||
```
|
||||
|
||||
Po zagonu lahko dostopate do nadzorne plošče Dify v brskalniku na [http://localhost/install](http://localhost/install) in začnete postopek inicializacije.
|
||||
|
||||
#### Iskanje pomoči
|
||||
Prosimo, glejte naša pogosta vprašanja [FAQ](https://docs.dify.ai/getting-started/install-self-hosted/faqs) če naletite na težave pri nastavitvi Dify. Če imate še vedno težave, se obrnite na [skupnost ali nas](#community--contact).
|
||||
|
||||
> Če želite prispevati k Difyju ali narediti dodaten razvoj, glejte naš vodnik za [uvajanje iz izvorne kode](https://docs.dify.ai/getting-started/install-self-hosted/local-source-code)
|
||||
|
||||
## Ključne značilnosti
|
||||
**1. Potek dela**:
|
||||
Zgradite in preizkusite zmogljive poteke dela AI na vizualnem platnu, pri čemer izkoristite vse naslednje funkcije in več.
|
||||
|
||||
|
||||
https://github.com/langgenius/dify/assets/13230914/356df23e-1604-483d-80a6-9517ece318aa
|
||||
|
||||
|
||||
|
||||
**2. Celovita podpora za modele**:
|
||||
Brezhibna integracija s stotinami lastniških/odprtokodnih LLM-jev ducatov ponudnikov sklepanja in samostojnih rešitev, ki pokrivajo GPT, Mistral, Llama3 in vse modele, združljive z API-jem OpenAI. Celoten seznam podprtih ponudnikov modelov najdete [tukaj](https://docs.dify.ai/getting-started/readme/model-providers).
|
||||
|
||||
|
||||
|
||||
|
||||
**3. Prompt IDE**:
|
||||
intuitivni vmesnik za ustvarjanje pozivov, primerjavo zmogljivosti modela in dodajanje dodatnih funkcij, kot je pretvorba besedila v govor, aplikaciji, ki temelji na klepetu.
|
||||
|
||||
**4. RAG Pipeline**:
|
||||
E Obsežne zmogljivosti RAG, ki pokrivajo vse od vnosa dokumenta do priklica, s podporo za ekstrakcijo besedila iz datotek PDF, PPT in drugih običajnih formatov dokumentov.
|
||||
|
||||
**5. Agent capabilities**:
|
||||
definirate lahko agente, ki temeljijo na klicanju funkcij LLM ali ReAct, in dodate vnaprej izdelana orodja ali orodja po meri za agenta. Dify ponuja več kot 50 vgrajenih orodij za agente AI, kot so Google Search, DALL·E, Stable Diffusion in WolframAlpha.
|
||||
|
||||
**6. LLMOps**:
|
||||
Spremljajte in analizirajte dnevnike aplikacij in učinkovitost skozi čas. Pozive, nabore podatkov in modele lahko nenehno izboljšujete na podlagi proizvodnih podatkov in opomb.
|
||||
|
||||
**7. Backend-as-a-Service**:
|
||||
AVse ponudbe Difyja so opremljene z ustreznimi API-ji, tako da lahko Dify brez težav integrirate v svojo poslovno logiko.
|
||||
|
||||
## Primerjava Funkcij
|
||||
|
||||
<table style="width: 100%;">
|
||||
<tr>
|
||||
<th align="center">Funkcija</th>
|
||||
<th align="center">Dify.AI</th>
|
||||
<th align="center">LangChain</th>
|
||||
<th align="center">Flowise</th>
|
||||
<th align="center">OpenAI Assistants API</th>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Programski pristop</td>
|
||||
<td align="center">API + usmerjeno v aplikacije</td>
|
||||
<td align="center">Python koda</td>
|
||||
<td align="center">Usmerjeno v aplikacije</td>
|
||||
<td align="center">Usmerjeno v API</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Podprti LLM-ji</td>
|
||||
<td align="center">Bogata izbira</td>
|
||||
<td align="center">Bogata izbira</td>
|
||||
<td align="center">Bogata izbira</td>
|
||||
<td align="center">Samo OpenAI</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">RAG pogon</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Agent</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
<td align="center">✅</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Potek dela</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Spremljanje</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
<td align="center">❌</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Funkcija za podjetja (SSO/nadzor dostopa)</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
<td align="center">❌</td>
|
||||
<td align="center">❌</td>
|
||||
</tr>
|
||||
<tr>
|
||||
<td align="center">Lokalna namestitev</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">✅</td>
|
||||
<td align="center">❌</td>
|
||||
</tr>
|
||||
</table>
|
||||
|
||||
## Uporaba Dify
|
||||
|
||||
- **Cloud </br>**
|
||||
Gostimo storitev Dify Cloud za vsakogar, ki jo lahko preizkusite brez nastavitev. Zagotavlja vse zmožnosti različice za samostojno namestitev in vključuje 200 brezplačnih klicev GPT-4 v načrtu peskovnika.
|
||||
|
||||
- **Self-hosting Dify Community Edition</br>**
|
||||
Hitro zaženite Dify v svojem okolju s tem [začetnim vodnikom](#quick-start) . Za dodatne reference in podrobnejša navodila uporabite našo [dokumentacijo](https://docs.dify.ai) .
|
||||
|
||||
|
||||
- **Dify za podjetja/organizacije</br>**
|
||||
Ponujamo dodatne funkcije, osredotočene na podjetja. Zabeležite svoja vprašanja prek tega klepetalnega robota ali nam pošljite e-pošto, da se pogovorimo o potrebah podjetja. </br>
|
||||
> Za novoustanovljena podjetja in mala podjetja, ki uporabljajo AWS, si oglejte Dify Premium na AWS Marketplace in ga z enim klikom uvedite v svoj AWS VPC. To je cenovno ugodna ponudba AMI z možnostjo ustvarjanja aplikacij z logotipom in blagovno znamko po meri.
|
||||
|
||||
|
||||
## Staying ahead
|
||||
|
||||
Star Dify on GitHub and be instantly notified of new releases.
|
||||
|
||||
|
||||
|
||||
|
||||
## Napredne nastavitve
|
||||
|
||||
Če morate prilagoditi konfiguracijo, si oglejte komentarje v naši datoteki .env.example in posodobite ustrezne vrednosti v svoji .env datoteki. Poleg tega boste morda morali prilagoditi docker-compose.yamlsamo datoteko, na primer spremeniti različice slike, preslikave vrat ali namestitve nosilca, glede na vaše specifično okolje in zahteve za uvajanje. Po kakršnih koli spremembah ponovno zaženite docker-compose up -d. Celoten seznam razpoložljivih spremenljivk okolja najdete tukaj .
|
||||
|
||||
Če želite konfigurirati visoko razpoložljivo nastavitev, so na voljo Helm Charts in datoteke YAML, ki jih prispeva skupnost, ki omogočajo uvedbo Difyja v Kubernetes.
|
||||
|
||||
- [Helm Chart by @LeoQuote](https://github.com/douban/charts/tree/master/charts/dify)
|
||||
- [Helm Chart by @BorisPolonsky](https://github.com/BorisPolonsky/dify-helm)
|
||||
- [YAML file by @Winson-030](https://github.com/Winson-030/dify-kubernetes)
|
||||
- [YAML file by @wyy-holding](https://github.com/wyy-holding/dify-k8s)
|
||||
|
||||
#### Uporaba Terraform za uvajanje
|
||||
|
||||
namestite Dify v Cloud Platform z enim klikom z uporabo [terraform](https://www.terraform.io/)
|
||||
|
||||
##### Azure Global
|
||||
- [Azure Terraform by @nikawang](https://github.com/nikawang/dify-azure-terraform)
|
||||
|
||||
##### Google Cloud
|
||||
- [Google Cloud Terraform by @sotazum](https://github.com/DeNA/dify-google-cloud-terraform)
|
||||
|
||||
#### Uporaba AWS CDK za uvajanje
|
||||
|
||||
Uvedite Dify v AWS z uporabo [CDK](https://aws.amazon.com/cdk/)
|
||||
|
||||
##### AWS
|
||||
- [AWS CDK by @KevinZhao](https://github.com/aws-samples/solution-for-deploying-dify-on-aws)
|
||||
|
||||
## Prispevam
|
||||
|
||||
Za tiste, ki bi radi prispevali kodo, si oglejte naš vodnik za prispevke . Hkrati vas prosimo, da podprete Dify tako, da ga delite na družbenih medijih ter na dogodkih in konferencah.
|
||||
|
||||
|
||||
|
||||
> Iščemo sodelavce za pomoč pri prevajanju Difyja v jezike, ki niso mandarinščina ali angleščina. Če želite pomagati, si oglejte i18n README za več informacij in nam pustite komentar v global-userskanalu našega strežnika skupnosti Discord .
|
||||
|
||||
## Skupnost in stik
|
||||
|
||||
* [Github Discussion](https://github.com/langgenius/dify/discussions). Najboljše za: izmenjavo povratnih informacij in postavljanje vprašanj.
|
||||
* [GitHub Issues](https://github.com/langgenius/dify/issues). Najboljše za: hrošče, na katere naletite pri uporabi Dify.AI, in predloge funkcij. Oglejte si naš [vodnik za prispevke](https://github.com/langgenius/dify/blob/main/CONTRIBUTING.md).
|
||||
* [Discord](https://discord.gg/FngNHpbcY7). Najboljše za: deljenje vaših aplikacij in druženje s skupnostjo.
|
||||
* [X(Twitter)](https://twitter.com/dify_ai). Najboljše za: deljenje vaših aplikacij in druženje s skupnostjo.
|
||||
|
||||
**Contributors**
|
||||
|
||||
<a href="https://github.com/langgenius/dify/graphs/contributors">
|
||||
<img src="https://contrib.rocks/image?repo=langgenius/dify" />
|
||||
</a>
|
||||
|
||||
## Star history
|
||||
|
||||
[](https://star-history.com/#langgenius/dify&Date)
|
||||
|
||||
|
||||
## Varnostno razkritje
|
||||
|
||||
Zaradi zaščite vaše zasebnosti se izogibajte objavljanju varnostnih vprašanj na GitHub. Namesto tega pošljite vprašanja na security@dify.ai in zagotovili vam bomo podrobnejši odgovor.
|
||||
|
||||
## Licenca
|
||||
|
||||
To skladišče je na voljo pod [odprtokodno licenco Dify](LICENSE) , ki je v bistvu Apache 2.0 z nekaj dodatnimi omejitvami.
|
||||
|
||||
@ -16,4 +16,4 @@ logs
|
||||
.ruff_cache
|
||||
|
||||
# venv
|
||||
.venv
|
||||
.venv
|
||||
|
||||
@ -90,3 +90,4 @@
|
||||
```bash
|
||||
uv run -P api bash dev/pytest/pytest_all_tests.sh
|
||||
```
|
||||
|
||||
|
||||
@ -444,13 +444,13 @@ def convert_to_agent_apps():
|
||||
WHERE a.mode = 'chat'
|
||||
AND am.agent_mode is not null
|
||||
AND (
|
||||
am.agent_mode like '%"strategy": "function_call"%'
|
||||
am.agent_mode like '%"strategy": "function_call"%'
|
||||
OR am.agent_mode like '%"strategy": "react"%'
|
||||
)
|
||||
)
|
||||
AND (
|
||||
am.agent_mode like '{"enabled": true%'
|
||||
am.agent_mode like '{"enabled": true%'
|
||||
OR am.agent_mode like '{"max_iteration": %'
|
||||
) ORDER BY a.created_at DESC LIMIT 1000
|
||||
) ORDER BY a.created_at DESC LIMIT 1000
|
||||
"""
|
||||
|
||||
with db.engine.begin() as conn:
|
||||
@ -818,8 +818,9 @@ def clear_free_plan_tenant_expired_logs(days: int, batch: int, tenant_ids: list[
|
||||
click.echo(click.style("Clear free plan tenant expired logs completed.", fg="green"))
|
||||
|
||||
|
||||
@click.option("-f", "--force", is_flag=True, help="Skip user confirmation and force the command to execute.")
|
||||
@click.command("clear-orphaned-file-records", help="Clear orphaned file records.")
|
||||
def clear_orphaned_file_records():
|
||||
def clear_orphaned_file_records(force: bool):
|
||||
"""
|
||||
Clear orphaned file records in the database.
|
||||
"""
|
||||
@ -845,7 +846,15 @@ def clear_orphaned_file_records():
|
||||
|
||||
# notify user and ask for confirmation
|
||||
click.echo(
|
||||
click.style("This command will find and delete orphaned file records in the following tables:", fg="yellow")
|
||||
click.style(
|
||||
"This command will first find and delete orphaned file records from the message_files table,", fg="yellow"
|
||||
)
|
||||
)
|
||||
click.echo(
|
||||
click.style(
|
||||
"and then it will find and delete orphaned file records in the following tables:",
|
||||
fg="yellow",
|
||||
)
|
||||
)
|
||||
for files_table in files_tables:
|
||||
click.echo(click.style(f"- {files_table['table']}", fg="yellow"))
|
||||
@ -878,11 +887,55 @@ def clear_orphaned_file_records():
|
||||
fg="yellow",
|
||||
)
|
||||
)
|
||||
click.confirm("Do you want to proceed?", abort=True)
|
||||
if not force:
|
||||
click.confirm("Do you want to proceed?", abort=True)
|
||||
|
||||
# start the cleanup process
|
||||
click.echo(click.style("Starting orphaned file records cleanup.", fg="white"))
|
||||
|
||||
# clean up the orphaned records in the message_files table where message_id doesn't exist in messages table
|
||||
try:
|
||||
click.echo(
|
||||
click.style("- Listing message_files records where message_id doesn't exist in messages table", fg="white")
|
||||
)
|
||||
query = (
|
||||
"SELECT mf.id, mf.message_id "
|
||||
"FROM message_files mf LEFT JOIN messages m ON mf.message_id = m.id "
|
||||
"WHERE m.id IS NULL"
|
||||
)
|
||||
orphaned_message_files = []
|
||||
with db.engine.begin() as conn:
|
||||
rs = conn.execute(db.text(query))
|
||||
for i in rs:
|
||||
orphaned_message_files.append({"id": str(i[0]), "message_id": str(i[1])})
|
||||
|
||||
if orphaned_message_files:
|
||||
click.echo(click.style(f"Found {len(orphaned_message_files)} orphaned message_files records:", fg="white"))
|
||||
for record in orphaned_message_files:
|
||||
click.echo(click.style(f" - id: {record['id']}, message_id: {record['message_id']}", fg="black"))
|
||||
|
||||
if not force:
|
||||
click.confirm(
|
||||
(
|
||||
f"Do you want to proceed "
|
||||
f"to delete all {len(orphaned_message_files)} orphaned message_files records?"
|
||||
),
|
||||
abort=True,
|
||||
)
|
||||
|
||||
click.echo(click.style("- Deleting orphaned message_files records", fg="white"))
|
||||
query = "DELETE FROM message_files WHERE id IN :ids"
|
||||
with db.engine.begin() as conn:
|
||||
conn.execute(db.text(query), {"ids": tuple([record["id"] for record in orphaned_message_files])})
|
||||
click.echo(
|
||||
click.style(f"Removed {len(orphaned_message_files)} orphaned message_files records.", fg="green")
|
||||
)
|
||||
else:
|
||||
click.echo(click.style("No orphaned message_files records found. There is nothing to delete.", fg="green"))
|
||||
except Exception as e:
|
||||
click.echo(click.style(f"Error deleting orphaned message_files records: {str(e)}", fg="red"))
|
||||
|
||||
# clean up the orphaned records in the rest of the *_files tables
|
||||
try:
|
||||
# fetch file id and keys from each table
|
||||
all_files_in_tables = []
|
||||
@ -964,7 +1017,8 @@ def clear_orphaned_file_records():
|
||||
click.echo(click.style(f"Found {len(orphaned_files)} orphaned file records.", fg="white"))
|
||||
for file in orphaned_files:
|
||||
click.echo(click.style(f"- orphaned file id: {file}", fg="black"))
|
||||
click.confirm(f"Do you want to proceed to delete all {len(orphaned_files)} orphaned file records?", abort=True)
|
||||
if not force:
|
||||
click.confirm(f"Do you want to proceed to delete all {len(orphaned_files)} orphaned file records?", abort=True)
|
||||
|
||||
# delete orphaned records for each file
|
||||
try:
|
||||
@ -979,8 +1033,9 @@ def clear_orphaned_file_records():
|
||||
click.echo(click.style(f"Removed {len(orphaned_files)} orphaned file records.", fg="green"))
|
||||
|
||||
|
||||
@click.option("-f", "--force", is_flag=True, help="Skip user confirmation and force the command to execute.")
|
||||
@click.command("remove-orphaned-files-on-storage", help="Remove orphaned files on the storage.")
|
||||
def remove_orphaned_files_on_storage():
|
||||
def remove_orphaned_files_on_storage(force: bool):
|
||||
"""
|
||||
Remove orphaned files on the storage.
|
||||
"""
|
||||
@ -1028,7 +1083,8 @@ def remove_orphaned_files_on_storage():
|
||||
fg="yellow",
|
||||
)
|
||||
)
|
||||
click.confirm("Do you want to proceed?", abort=True)
|
||||
if not force:
|
||||
click.confirm("Do you want to proceed?", abort=True)
|
||||
|
||||
# start the cleanup process
|
||||
click.echo(click.style("Starting orphaned files cleanup.", fg="white"))
|
||||
@ -1069,7 +1125,8 @@ def remove_orphaned_files_on_storage():
|
||||
click.echo(click.style(f"Found {len(orphaned_files)} orphaned files.", fg="white"))
|
||||
for file in orphaned_files:
|
||||
click.echo(click.style(f"- orphaned file: {file}", fg="black"))
|
||||
click.confirm(f"Do you want to proceed to remove all {len(orphaned_files)} orphaned files?", abort=True)
|
||||
if not force:
|
||||
click.confirm(f"Do you want to proceed to remove all {len(orphaned_files)} orphaned files?", abort=True)
|
||||
|
||||
# delete orphaned files
|
||||
removed_files = 0
|
||||
|
||||
@ -398,6 +398,11 @@ class InnerAPIConfig(BaseSettings):
|
||||
default=False,
|
||||
)
|
||||
|
||||
INNER_API_KEY: Optional[str] = Field(
|
||||
description="API key for accessing the internal API",
|
||||
default=None,
|
||||
)
|
||||
|
||||
|
||||
class LoggingConfig(BaseSettings):
|
||||
"""
|
||||
|
||||
@ -1,4 +1,5 @@
|
||||
from typing import Optional
|
||||
import enum
|
||||
from typing import Literal, Optional
|
||||
|
||||
from pydantic import Field, PositiveInt
|
||||
from pydantic_settings import BaseSettings
|
||||
@ -9,6 +10,14 @@ class OpenSearchConfig(BaseSettings):
|
||||
Configuration settings for OpenSearch
|
||||
"""
|
||||
|
||||
class AuthMethod(enum.StrEnum):
|
||||
"""
|
||||
Authentication method for OpenSearch
|
||||
"""
|
||||
|
||||
BASIC = "basic"
|
||||
AWS_MANAGED_IAM = "aws_managed_iam"
|
||||
|
||||
OPENSEARCH_HOST: Optional[str] = Field(
|
||||
description="Hostname or IP address of the OpenSearch server (e.g., 'localhost' or 'opensearch.example.com')",
|
||||
default=None,
|
||||
@ -19,6 +28,16 @@ class OpenSearchConfig(BaseSettings):
|
||||
default=9200,
|
||||
)
|
||||
|
||||
OPENSEARCH_SECURE: bool = Field(
|
||||
description="Whether to use SSL/TLS encrypted connection for OpenSearch (True for HTTPS, False for HTTP)",
|
||||
default=False,
|
||||
)
|
||||
|
||||
OPENSEARCH_AUTH_METHOD: AuthMethod = Field(
|
||||
description="Authentication method for OpenSearch connection (default is 'basic')",
|
||||
default=AuthMethod.BASIC,
|
||||
)
|
||||
|
||||
OPENSEARCH_USER: Optional[str] = Field(
|
||||
description="Username for authenticating with OpenSearch",
|
||||
default=None,
|
||||
@ -29,7 +48,11 @@ class OpenSearchConfig(BaseSettings):
|
||||
default=None,
|
||||
)
|
||||
|
||||
OPENSEARCH_SECURE: bool = Field(
|
||||
description="Whether to use SSL/TLS encrypted connection for OpenSearch (True for HTTPS, False for HTTP)",
|
||||
default=False,
|
||||
OPENSEARCH_AWS_REGION: Optional[str] = Field(
|
||||
description="AWS region for OpenSearch (e.g. 'us-west-2')",
|
||||
default=None,
|
||||
)
|
||||
|
||||
OPENSEARCH_AWS_SERVICE: Optional[Literal["es", "aoss"]] = Field(
|
||||
description="AWS service for OpenSearch (e.g. 'aoss' for OpenSearch Serverless)", default=None
|
||||
)
|
||||
|
||||
7
api/constants/mimetypes.py
Normal file
7
api/constants/mimetypes.py
Normal file
@ -0,0 +1,7 @@
|
||||
# The two constants below should keep in sync.
|
||||
# Default content type for files which have no explicit content type.
|
||||
|
||||
DEFAULT_MIME_TYPE = "application/octet-stream"
|
||||
# Default file extension for files which have no explicit content type, should
|
||||
# correspond to the `DEFAULT_MIME_TYPE` above.
|
||||
DEFAULT_EXTENSION = ".bin"
|
||||
@ -2,22 +2,22 @@ import uuid
|
||||
from typing import cast
|
||||
|
||||
from flask_login import current_user # type: ignore
|
||||
from flask_restful import (Resource, inputs, marshal, # type: ignore
|
||||
marshal_with, reqparse)
|
||||
from flask_restful import Resource, inputs, marshal, marshal_with, reqparse # type: ignore
|
||||
from sqlalchemy import select
|
||||
from sqlalchemy.orm import Session
|
||||
from werkzeug.exceptions import BadRequest, Forbidden, abort
|
||||
|
||||
from controllers.console import api
|
||||
from controllers.console.app.wraps import get_app_model
|
||||
from controllers.console.wraps import (account_initialization_required,
|
||||
cloud_edition_billing_resource_check,
|
||||
enterprise_license_required,
|
||||
setup_required)
|
||||
from controllers.console.wraps import (
|
||||
account_initialization_required,
|
||||
cloud_edition_billing_resource_check,
|
||||
enterprise_license_required,
|
||||
setup_required,
|
||||
)
|
||||
from core.ops.ops_trace_manager import OpsTraceManager
|
||||
from extensions.ext_database import db
|
||||
from fields.app_fields import (app_detail_fields, app_detail_fields_with_site,
|
||||
app_pagination_fields)
|
||||
from fields.app_fields import app_detail_fields, app_detail_fields_with_site, app_pagination_fields
|
||||
from libs.login import login_required
|
||||
from models import Account, App
|
||||
from services.app_dsl_service import AppDslService, ImportMode
|
||||
|
||||
@ -24,7 +24,7 @@ from libs.password import hash_password, valid_password
|
||||
from models.account import Account
|
||||
from services.account_service import AccountService, TenantService
|
||||
from services.errors.account import AccountRegisterError
|
||||
from services.errors.workspace import WorkSpaceNotAllowedCreateError, WorkspacesLimitExceededError
|
||||
from services.errors.workspace import WorkSpaceNotAllowedCreateError
|
||||
from services.feature_service import FeatureService
|
||||
|
||||
|
||||
|
||||
@ -4,15 +4,13 @@ from typing import Any
|
||||
|
||||
from flask import request
|
||||
from flask_login import current_user # type: ignore
|
||||
from flask_restful import (Resource, inputs, marshal_with, # type: ignore
|
||||
reqparse)
|
||||
from flask_restful import Resource, inputs, marshal_with, reqparse # type: ignore
|
||||
from sqlalchemy import and_
|
||||
from werkzeug.exceptions import BadRequest, Forbidden, NotFound
|
||||
|
||||
from controllers.console import api
|
||||
from controllers.console.explore.wraps import InstalledAppResource
|
||||
from controllers.console.wraps import (account_initialization_required,
|
||||
cloud_edition_billing_resource_check)
|
||||
from controllers.console.wraps import account_initialization_required, cloud_edition_billing_resource_check
|
||||
from extensions.ext_database import db
|
||||
from fields.installed_app_fields import installed_app_list_fields
|
||||
from libs.login import login_required
|
||||
|
||||
@ -70,12 +70,26 @@ class FilePreviewApi(Resource):
|
||||
direct_passthrough=True,
|
||||
headers={},
|
||||
)
|
||||
# add Accept-Ranges header for audio/video files
|
||||
if upload_file.mime_type in [
|
||||
"audio/mpeg",
|
||||
"audio/wav",
|
||||
"audio/mp4",
|
||||
"audio/ogg",
|
||||
"audio/flac",
|
||||
"audio/aac",
|
||||
"video/mp4",
|
||||
"video/webm",
|
||||
"video/quicktime",
|
||||
"audio/x-m4a",
|
||||
]:
|
||||
response.headers["Accept-Ranges"] = "bytes"
|
||||
if upload_file.size > 0:
|
||||
response.headers["Content-Length"] = str(upload_file.size)
|
||||
if args["as_attachment"]:
|
||||
encoded_filename = quote(upload_file.name)
|
||||
response.headers["Content-Disposition"] = f"attachment; filename*=UTF-8''{encoded_filename}"
|
||||
response.headers["Content-Type"] = "application/octet-stream"
|
||||
response.headers["Content-Type"] = "application/octet-stream"
|
||||
|
||||
return response
|
||||
|
||||
|
||||
@ -1,10 +1,14 @@
|
||||
from urllib.parse import quote
|
||||
|
||||
from flask import Response
|
||||
from flask_restful import Resource, reqparse # type: ignore
|
||||
from werkzeug.exceptions import Forbidden, NotFound
|
||||
|
||||
from controllers.files import api
|
||||
from controllers.files.error import UnsupportedFileTypeError
|
||||
from core.tools.signature import verify_tool_file_signature
|
||||
from core.tools.tool_file_manager import ToolFileManager
|
||||
from models import db as global_db
|
||||
|
||||
|
||||
class ToolFilePreviewApi(Resource):
|
||||
@ -19,17 +23,14 @@ class ToolFilePreviewApi(Resource):
|
||||
parser.add_argument("as_attachment", type=bool, required=False, default=False, location="args")
|
||||
|
||||
args = parser.parse_args()
|
||||
|
||||
if not ToolFileManager.verify_file(
|
||||
file_id=file_id,
|
||||
timestamp=args["timestamp"],
|
||||
nonce=args["nonce"],
|
||||
sign=args["sign"],
|
||||
if not verify_tool_file_signature(
|
||||
file_id=file_id, timestamp=args["timestamp"], nonce=args["nonce"], sign=args["sign"]
|
||||
):
|
||||
raise Forbidden("Invalid request.")
|
||||
|
||||
try:
|
||||
stream, tool_file = ToolFileManager.get_file_generator_by_tool_file_id(
|
||||
tool_file_manager = ToolFileManager(engine=global_db.engine)
|
||||
stream, tool_file = tool_file_manager.get_file_generator_by_tool_file_id(
|
||||
file_id,
|
||||
)
|
||||
|
||||
@ -47,7 +48,8 @@ class ToolFilePreviewApi(Resource):
|
||||
if tool_file.size > 0:
|
||||
response.headers["Content-Length"] = str(tool_file.size)
|
||||
if args["as_attachment"]:
|
||||
response.headers["Content-Disposition"] = f"attachment; filename={tool_file.name}"
|
||||
encoded_filename = quote(tool_file.name)
|
||||
response.headers["Content-Disposition"] = f"attachment; filename*=UTF-8''{encoded_filename}"
|
||||
|
||||
return response
|
||||
|
||||
|
||||
@ -53,7 +53,7 @@ class PluginUploadFileApi(Resource):
|
||||
raise Forbidden("Invalid request.")
|
||||
|
||||
try:
|
||||
tool_file = ToolFileManager.create_file_by_raw(
|
||||
tool_file = ToolFileManager().create_file_by_raw(
|
||||
user_id=user.id,
|
||||
tenant_id=tenant_id,
|
||||
file_binary=file.read(),
|
||||
|
||||
@ -5,6 +5,6 @@ from libs.external_api import ExternalApi
|
||||
bp = Blueprint("inner_api", __name__, url_prefix="/inner/api")
|
||||
api = ExternalApi(bp)
|
||||
|
||||
from .plugin import plugin
|
||||
from . import mail
|
||||
from .plugin import plugin
|
||||
from .workspace import workspace
|
||||
|
||||
@ -18,7 +18,7 @@ def enterprise_inner_api_only(view):
|
||||
|
||||
# get header 'X-Inner-Api-Key'
|
||||
inner_api_key = request.headers.get("X-Inner-Api-Key")
|
||||
if not inner_api_key or inner_api_key != dify_config.INNER_API_KEY_FOR_PLUGIN:
|
||||
if not inner_api_key or inner_api_key != dify_config.INNER_API_KEY:
|
||||
abort(401)
|
||||
|
||||
return view(*args, **kwargs)
|
||||
|
||||
@ -69,6 +69,13 @@ class CotAgentRunner(BaseAgentRunner, ABC):
|
||||
tool_instances, prompt_messages_tools = self._init_prompt_tools()
|
||||
self._prompt_messages_tools = prompt_messages_tools
|
||||
|
||||
# fix metadata filter not work
|
||||
if app_config.dataset is not None:
|
||||
metadata_filtering_conditions = app_config.dataset.retrieve_config.metadata_filtering_conditions
|
||||
for key, dataset_retriever_tool in tool_instances.items():
|
||||
if hasattr(dataset_retriever_tool, "retrieval_tool"):
|
||||
dataset_retriever_tool.retrieval_tool.metadata_filtering_conditions = metadata_filtering_conditions
|
||||
|
||||
function_call_state = True
|
||||
llm_usage: dict[str, Optional[LLMUsage]] = {"usage": None}
|
||||
final_answer = ""
|
||||
|
||||
@ -45,6 +45,13 @@ class FunctionCallAgentRunner(BaseAgentRunner):
|
||||
# convert tools into ModelRuntime Tool format
|
||||
tool_instances, prompt_messages_tools = self._init_prompt_tools()
|
||||
|
||||
# fix metadata filter not work
|
||||
if app_config.dataset is not None:
|
||||
metadata_filtering_conditions = app_config.dataset.retrieve_config.metadata_filtering_conditions
|
||||
for key, dataset_retriever_tool in tool_instances.items():
|
||||
if hasattr(dataset_retriever_tool, "retrieval_tool"):
|
||||
dataset_retriever_tool.retrieval_tool.metadata_filtering_conditions = metadata_filtering_conditions
|
||||
|
||||
assert app_config.agent
|
||||
|
||||
iteration_step = 1
|
||||
|
||||
@ -1,4 +1,4 @@
|
||||
ENGLISH_REACT_COMPLETION_PROMPT_TEMPLATES = """Respond to the human as helpfully and accurately as possible.
|
||||
ENGLISH_REACT_COMPLETION_PROMPT_TEMPLATES = """Respond to the human as helpfully and accurately as possible.
|
||||
|
||||
{{instruction}}
|
||||
|
||||
@ -47,7 +47,7 @@ Thought:""" # noqa: E501
|
||||
ENGLISH_REACT_COMPLETION_AGENT_SCRATCHPAD_TEMPLATES = """Observation: {{observation}}
|
||||
Thought:"""
|
||||
|
||||
ENGLISH_REACT_CHAT_PROMPT_TEMPLATES = """Respond to the human as helpfully and accurately as possible.
|
||||
ENGLISH_REACT_CHAT_PROMPT_TEMPLATES = """Respond to the human as helpfully and accurately as possible.
|
||||
|
||||
{{instruction}}
|
||||
|
||||
|
||||
@ -21,7 +21,6 @@ from core.model_runtime.entities.message_entities import (
|
||||
ImagePromptMessageContent,
|
||||
PromptMessage,
|
||||
)
|
||||
from core.model_runtime.entities.model_entities import ModelPropertyKey
|
||||
from core.model_runtime.errors.invoke import InvokeBadRequestError
|
||||
from core.moderation.input_moderation import InputModeration
|
||||
from core.prompt.advanced_prompt_transform import AdvancedPromptTransform
|
||||
|
||||
@ -24,7 +24,7 @@ from core.app.entities.task_entities import (
|
||||
WorkflowTaskState,
|
||||
)
|
||||
from core.llm_generator.llm_generator import LLMGenerator
|
||||
from core.tools.tool_file_manager import ToolFileManager
|
||||
from core.tools.signature import sign_tool_file
|
||||
from extensions.ext_database import db
|
||||
from models.model import AppMode, Conversation, MessageAnnotation, MessageFile
|
||||
from services.annotation_service import AppAnnotationService
|
||||
@ -154,7 +154,7 @@ class MessageCycleManage:
|
||||
if message_file.url.startswith("http"):
|
||||
url = message_file.url
|
||||
else:
|
||||
url = ToolFileManager.sign_file(tool_file_id=tool_file_id, extension=extension)
|
||||
url = sign_tool_file(tool_file_id=tool_file_id, extension=extension)
|
||||
|
||||
return MessageFileStreamResponse(
|
||||
task_id=self._application_generate_entity.task_id,
|
||||
|
||||
@ -381,6 +381,8 @@ class WorkflowCycleManage:
|
||||
workflow_node_execution.elapsed_time = elapsed_time
|
||||
workflow_node_execution.execution_metadata = execution_metadata
|
||||
|
||||
self._workflow_node_execution_repository.update(workflow_node_execution)
|
||||
|
||||
return workflow_node_execution
|
||||
|
||||
def _handle_workflow_node_execution_retried(
|
||||
|
||||
@ -1 +1 @@
|
||||
1
|
||||
1
|
||||
|
||||
@ -10,12 +10,12 @@ from core.model_runtime.entities import (
|
||||
VideoPromptMessageContent,
|
||||
)
|
||||
from core.model_runtime.entities.message_entities import PromptMessageContentUnionTypes
|
||||
from core.tools.signature import sign_tool_file
|
||||
from extensions.ext_storage import storage
|
||||
|
||||
from . import helpers
|
||||
from .enums import FileAttribute
|
||||
from .models import File, FileTransferMethod, FileType
|
||||
from .tool_file_parser import ToolFileParser
|
||||
|
||||
|
||||
def get_attr(*, file: File, attr: FileAttribute):
|
||||
@ -130,6 +130,6 @@ def _to_url(f: File, /):
|
||||
# add sign url
|
||||
if f.related_id is None or f.extension is None:
|
||||
raise ValueError("Missing file related_id or extension")
|
||||
return ToolFileParser.get_tool_file_manager().sign_file(tool_file_id=f.related_id, extension=f.extension)
|
||||
return sign_tool_file(tool_file_id=f.related_id, extension=f.extension)
|
||||
else:
|
||||
raise ValueError(f"Unsupported transfer method: {f.transfer_method}")
|
||||
|
||||
@ -4,11 +4,11 @@ from typing import Any, Optional
|
||||
from pydantic import BaseModel, Field, model_validator
|
||||
|
||||
from core.model_runtime.entities.message_entities import ImagePromptMessageContent
|
||||
from core.tools.signature import sign_tool_file
|
||||
|
||||
from . import helpers
|
||||
from .constants import FILE_MODEL_IDENTITY
|
||||
from .enums import FileTransferMethod, FileType
|
||||
from .tool_file_parser import ToolFileParser
|
||||
|
||||
|
||||
class ImageConfig(BaseModel):
|
||||
@ -34,13 +34,21 @@ class FileUploadConfig(BaseModel):
|
||||
|
||||
|
||||
class File(BaseModel):
|
||||
# NOTE: dify_model_identity is a special identifier used to distinguish between
|
||||
# new and old data formats during serialization and deserialization.
|
||||
dify_model_identity: str = FILE_MODEL_IDENTITY
|
||||
|
||||
id: Optional[str] = None # message file id
|
||||
tenant_id: str
|
||||
type: FileType
|
||||
transfer_method: FileTransferMethod
|
||||
# If `transfer_method` is `FileTransferMethod.remote_url`, the
|
||||
# `remote_url` attribute must not be `None`.
|
||||
remote_url: Optional[str] = None # remote url
|
||||
# If `transfer_method` is `FileTransferMethod.local_file` or
|
||||
# `FileTransferMethod.tool_file`, the `related_id` attribute must not be `None`.
|
||||
#
|
||||
# It should be set to `ToolFile.id` when `transfer_method` is `tool_file`.
|
||||
related_id: Optional[str] = None
|
||||
filename: Optional[str] = None
|
||||
extension: Optional[str] = Field(default=None, description="File extension, should contains dot")
|
||||
@ -110,9 +118,7 @@ class File(BaseModel):
|
||||
elif self.transfer_method == FileTransferMethod.TOOL_FILE:
|
||||
assert self.related_id is not None
|
||||
assert self.extension is not None
|
||||
return ToolFileParser.get_tool_file_manager().sign_file(
|
||||
tool_file_id=self.related_id, extension=self.extension
|
||||
)
|
||||
return sign_tool_file(tool_file_id=self.related_id, extension=self.extension)
|
||||
|
||||
def to_plugin_parameter(self) -> dict[str, Any]:
|
||||
return {
|
||||
|
||||
@ -1,12 +1,19 @@
|
||||
from typing import TYPE_CHECKING, Any, cast
|
||||
from collections.abc import Callable
|
||||
from typing import TYPE_CHECKING
|
||||
|
||||
if TYPE_CHECKING:
|
||||
from core.tools.tool_file_manager import ToolFileManager
|
||||
|
||||
tool_file_manager: dict[str, Any] = {"manager": None}
|
||||
_tool_file_manager_factory: Callable[[], "ToolFileManager"] | None = None
|
||||
|
||||
|
||||
class ToolFileParser:
|
||||
@staticmethod
|
||||
def get_tool_file_manager() -> "ToolFileManager":
|
||||
return cast("ToolFileManager", tool_file_manager["manager"])
|
||||
assert _tool_file_manager_factory is not None
|
||||
return _tool_file_manager_factory()
|
||||
|
||||
|
||||
def set_tool_file_manager_factory(factory: Callable[[], "ToolFileManager"]) -> None:
|
||||
global _tool_file_manager_factory
|
||||
_tool_file_manager_factory = factory
|
||||
|
||||
@ -10,13 +10,13 @@ class NodeJsTemplateTransformer(TemplateTransformer):
|
||||
f"""
|
||||
// declare main function
|
||||
{cls._code_placeholder}
|
||||
|
||||
|
||||
// decode and prepare input object
|
||||
var inputs_obj = JSON.parse(Buffer.from('{cls._inputs_placeholder}', 'base64').toString('utf-8'))
|
||||
|
||||
|
||||
// execute main function
|
||||
var output_obj = main(inputs_obj)
|
||||
|
||||
|
||||
// convert output to json and print
|
||||
var output_json = JSON.stringify(output_obj)
|
||||
var result = `<<RESULT>>${{output_json}}<<RESULT>>`
|
||||
|
||||
@ -21,20 +21,20 @@ class Jinja2TemplateTransformer(TemplateTransformer):
|
||||
import jinja2
|
||||
template = jinja2.Template('''{cls._code_placeholder}''')
|
||||
return template.render(**inputs)
|
||||
|
||||
|
||||
import json
|
||||
from base64 import b64decode
|
||||
|
||||
|
||||
# decode and prepare input dict
|
||||
inputs_obj = json.loads(b64decode('{cls._inputs_placeholder}').decode('utf-8'))
|
||||
|
||||
|
||||
# execute main function
|
||||
output = main(**inputs_obj)
|
||||
|
||||
|
||||
# convert output and print
|
||||
result = f'''<<RESULT>>{{output}}<<RESULT>>'''
|
||||
print(result)
|
||||
|
||||
|
||||
""")
|
||||
return runner_script
|
||||
|
||||
@ -43,15 +43,15 @@ class Jinja2TemplateTransformer(TemplateTransformer):
|
||||
preload_script = dedent("""
|
||||
import jinja2
|
||||
from base64 import b64decode
|
||||
|
||||
|
||||
def _jinja2_preload_():
|
||||
# prepare jinja2 environment, load template and render before to avoid sandbox issue
|
||||
template = jinja2.Template('{{s}}')
|
||||
template.render(s='a')
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
_jinja2_preload_()
|
||||
|
||||
|
||||
""")
|
||||
|
||||
return preload_script
|
||||
|
||||
@ -9,16 +9,16 @@ class Python3TemplateTransformer(TemplateTransformer):
|
||||
runner_script = dedent(f"""
|
||||
# declare main function
|
||||
{cls._code_placeholder}
|
||||
|
||||
|
||||
import json
|
||||
from base64 import b64decode
|
||||
|
||||
|
||||
# decode and prepare input dict
|
||||
inputs_obj = json.loads(b64decode('{cls._inputs_placeholder}').decode('utf-8'))
|
||||
|
||||
|
||||
# execute main function
|
||||
output_obj = main(**inputs_obj)
|
||||
|
||||
|
||||
# convert output to json and print
|
||||
output_json = json.dumps(output_obj, indent=4)
|
||||
result = f'''<<RESULT>>{{output_json}}<<RESULT>>'''
|
||||
|
||||
@ -3,6 +3,8 @@ import logging
|
||||
import re
|
||||
from typing import Optional, cast
|
||||
|
||||
import json_repair
|
||||
|
||||
from core.llm_generator.output_parser.rule_config_generator import RuleConfigGeneratorOutputParser
|
||||
from core.llm_generator.output_parser.suggested_questions_after_answer import SuggestedQuestionsAfterAnswerOutputParser
|
||||
from core.llm_generator.prompts import (
|
||||
@ -366,7 +368,20 @@ class LLMGenerator:
|
||||
),
|
||||
)
|
||||
|
||||
generated_json_schema = cast(str, response.message.content)
|
||||
raw_content = response.message.content
|
||||
|
||||
if not isinstance(raw_content, str):
|
||||
raise ValueError(f"LLM response content must be a string, got: {type(raw_content)}")
|
||||
|
||||
try:
|
||||
parsed_content = json.loads(raw_content)
|
||||
except json.JSONDecodeError:
|
||||
parsed_content = json_repair.loads(raw_content)
|
||||
|
||||
if not isinstance(parsed_content, dict | list):
|
||||
raise ValueError(f"Failed to parse structured output from llm: {raw_content}")
|
||||
|
||||
generated_json_schema = json.dumps(parsed_content, indent=2, ensure_ascii=False)
|
||||
return {"output": generated_json_schema, "error": ""}
|
||||
|
||||
except InvokeError as e:
|
||||
|
||||
@ -1,5 +1,5 @@
|
||||
# Written by YORKI MINAKO🤡, Edited by Xiaoyi
|
||||
CONVERSATION_TITLE_PROMPT = """You need to decompose the user's input into "subject" and "intention" in order to accurately figure out what the user's input language actually is.
|
||||
CONVERSATION_TITLE_PROMPT = """You need to decompose the user's input into "subject" and "intention" in order to accurately figure out what the user's input language actually is.
|
||||
Notice: the language type user uses could be diverse, which can be English, Chinese, Italian, Español, Arabic, Japanese, French, and etc.
|
||||
ENSURE your output is in the SAME language as the user's input!
|
||||
Your output is restricted only to: (Input language) Intention + Subject(short as possible)
|
||||
@ -58,7 +58,7 @@ User Input: yo, 你今天咋样？
|
||||
"Your Output": "查询今日我的状态☺️"
|
||||
}
|
||||
|
||||
User Input:
|
||||
User Input:
|
||||
""" # noqa: E501
|
||||
|
||||
PYTHON_CODE_GENERATOR_PROMPT_TEMPLATE = (
|
||||
@ -163,11 +163,11 @@ Here is a task description for which I would like you to create a high-quality p
|
||||
{{TASK_DESCRIPTION}}
|
||||
</task_description>
|
||||
Based on task description, please create a well-structured prompt template that another AI could use to consistently complete the task. The prompt template should include:
|
||||
- Do not include <input> or <output> section and variables in the prompt, assume user will add them at their own will.
|
||||
- Clear instructions for the AI that will be using this prompt, demarcated with <instruction> tags. The instructions should provide step-by-step directions on how to complete the task using the input variables. Also Specifies in the instructions that the output should not contain any xml tag.
|
||||
- Relevant examples if needed to clarify the task further, demarcated with <example> tags. Do not include variables in the prompt. Give three pairs of input and output examples.
|
||||
- Do not include <input> or <output> section and variables in the prompt, assume user will add them at their own will.
|
||||
- Clear instructions for the AI that will be using this prompt, demarcated with <instruction> tags. The instructions should provide step-by-step directions on how to complete the task using the input variables. Also Specifies in the instructions that the output should not contain any xml tag.
|
||||
- Relevant examples if needed to clarify the task further, demarcated with <example> tags. Do not include variables in the prompt. Give three pairs of input and output examples.
|
||||
- Include other relevant sections demarcated with appropriate XML tags like <examples>, <instruction>.
|
||||
- Use the same language as task description.
|
||||
- Use the same language as task description.
|
||||
- Output in ``` xml ``` and start with <instruction>
|
||||
Please generate the full prompt template with at least 300 words and output only the prompt template.
|
||||
""" # noqa: E501
|
||||
@ -178,28 +178,28 @@ Here is a task description for which I would like you to create a high-quality p
|
||||
{{TASK_DESCRIPTION}}
|
||||
</task_description>
|
||||
Based on task description, please create a well-structured prompt template that another AI could use to consistently complete the task. The prompt template should include:
|
||||
- Descriptive variable names surrounded by {{ }} (two curly brackets) to indicate where the actual values will be substituted in. Choose variable names that clearly indicate the type of value expected. Variable names have to be composed of number, english alphabets and underline and nothing else.
|
||||
- Clear instructions for the AI that will be using this prompt, demarcated with <instruction> tags. The instructions should provide step-by-step directions on how to complete the task using the input variables. Also Specifies in the instructions that the output should not contain any xml tag.
|
||||
- Relevant examples if needed to clarify the task further, demarcated with <example> tags. Do not use curly brackets any other than in <instruction> section.
|
||||
- Descriptive variable names surrounded by {{ }} (two curly brackets) to indicate where the actual values will be substituted in. Choose variable names that clearly indicate the type of value expected. Variable names have to be composed of number, english alphabets and underline and nothing else.
|
||||
- Clear instructions for the AI that will be using this prompt, demarcated with <instruction> tags. The instructions should provide step-by-step directions on how to complete the task using the input variables. Also Specifies in the instructions that the output should not contain any xml tag.
|
||||
- Relevant examples if needed to clarify the task further, demarcated with <example> tags. Do not use curly brackets any other than in <instruction> section.
|
||||
- Any other relevant sections demarcated with appropriate XML tags like <input>, <output>, etc.
|
||||
- Use the same language as task description.
|
||||
- Use the same language as task description.
|
||||
- Output in ``` xml ``` and start with <instruction>
|
||||
Please generate the full prompt template and output only the prompt template.
|
||||
""" # noqa: E501
|
||||
|
||||
RULE_CONFIG_PARAMETER_GENERATE_TEMPLATE = """
|
||||
I need to extract the following information from the input text. The <information to be extracted> tag specifies the 'type', 'description' and 'required' of the information to be extracted.
|
||||
I need to extract the following information from the input text. The <information to be extracted> tag specifies the 'type', 'description' and 'required' of the information to be extracted.
|
||||
<information to be extracted>
|
||||
variables name bounded two double curly brackets. Variable name has to be composed of number, english alphabets and underline and nothing else.
|
||||
variables name bounded two double curly brackets. Variable name has to be composed of number, english alphabets and underline and nothing else.
|
||||
</information to be extracted>
|
||||
|
||||
Step 1: Carefully read the input and understand the structure of the expected output.
|
||||
Step 2: Extract relevant parameters from the provided text based on the name and description of object.
|
||||
Step 2: Extract relevant parameters from the provided text based on the name and description of object.
|
||||
Step 3: Structure the extracted parameters to JSON object as specified in <structure>.
|
||||
Step 4: Ensure that the list of variable_names is properly formatted and valid. The output should not contain any XML tags. Output an empty list if there is no valid variable name in input text.
|
||||
Step 4: Ensure that the list of variable_names is properly formatted and valid. The output should not contain any XML tags. Output an empty list if there is no valid variable name in input text.
|
||||
|
||||
### Structure
|
||||
Here is the structure of the expected output, I should always follow the output structure.
|
||||
Here is the structure of the expected output, I should always follow the output structure.
|
||||
["variable_name_1", "variable_name_2"]
|
||||
|
||||
### Input Text
|
||||
@ -214,13 +214,13 @@ I should always output a valid list. Output nothing other than the list of varia
|
||||
|
||||
RULE_CONFIG_STATEMENT_GENERATE_TEMPLATE = """
|
||||
<instruction>
|
||||
Step 1: Identify the purpose of the chatbot from the variable {{TASK_DESCRIPTION}} and infer chatbot's tone (e.g., friendly, professional, etc.) to add personality traits.
|
||||
Step 1: Identify the purpose of the chatbot from the variable {{TASK_DESCRIPTION}} and infer chatbot's tone (e.g., friendly, professional, etc.) to add personality traits.
|
||||
Step 2: Create a coherent and engaging opening statement.
|
||||
Step 3: Ensure the output is welcoming and clearly explains what the chatbot is designed to do. Do not include any XML tags in the output.
|
||||
Please use the same language as the user's input language. If user uses chinese then generate opening statement in chinese, if user uses english then generate opening statement in english.
|
||||
Example Input:
|
||||
Please use the same language as the user's input language. If user uses chinese then generate opening statement in chinese, if user uses english then generate opening statement in english.
|
||||
Example Input:
|
||||
Provide customer support for an e-commerce website
|
||||
Example Output:
|
||||
Example Output:
|
||||
Welcome! I'm here to assist you with any questions or issues you might have with your shopping experience. Whether you're looking for product information, need help with your order, or have any other inquiries, feel free to ask. I'm friendly, helpful, and ready to support you in any way I can.
|
||||
<Task>
|
||||
Here is the task description: {{INPUT_TEXT}}
|
||||
@ -276,15 +276,15 @@ Your task is to convert simple user descriptions into properly formatted JSON Sc
|
||||
{
|
||||
"type": "object",
|
||||
"properties": {
|
||||
"email": {
|
||||
"email": {
|
||||
"type": "string",
|
||||
"format": "email"
|
||||
},
|
||||
"password": {
|
||||
"password": {
|
||||
"type": "string",
|
||||
"minLength": 8
|
||||
},
|
||||
"age": {
|
||||
"age": {
|
||||
"type": "integer",
|
||||
"minimum": 18
|
||||
}
|
||||
|
||||
@ -101,7 +101,7 @@ class ModelInstance:
|
||||
@overload
|
||||
def invoke_llm(
|
||||
self,
|
||||
prompt_messages: list[PromptMessage],
|
||||
prompt_messages: Sequence[PromptMessage],
|
||||
model_parameters: Optional[dict] = None,
|
||||
tools: Sequence[PromptMessageTool] | None = None,
|
||||
stop: Optional[list[str]] = None,
|
||||
|
||||
@ -307,4 +307,4 @@ Runtime Errors:
|
||||
"""
|
||||
```
|
||||
|
||||
For interface method details, see: [Interfaces](./interfaces.md). For specific implementations, refer to: [llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py).
|
||||
For interface method details, see: [Interfaces](./interfaces.md). For specific implementations, refer to: [llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py).
|
||||
|
||||
@ -170,4 +170,4 @@ Runtime Errors:
|
||||
"""
|
||||
```
|
||||
|
||||
For interface method explanations, see: [Interfaces](./interfaces.md). For detailed implementation, refer to: [llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py).
|
||||
For interface method explanations, see: [Interfaces](./interfaces.md). For detailed implementation, refer to: [llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py).
|
||||
|
||||
@ -294,4 +294,4 @@ provider_credential_schema:
|
||||
"""
|
||||
```
|
||||
|
||||
接口方法说明见：[Interfaces](./interfaces.md)，具体实现可参考：[llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py)。
|
||||
接口方法说明见：[Interfaces](./interfaces.md)，具体实现可参考：[llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py)。
|
||||
|
||||
@ -169,4 +169,4 @@ pricing: # 价格信息
|
||||
"""
|
||||
```
|
||||
|
||||
接口方法说明见：[Interfaces](./interfaces.md)，具体实现可参考：[llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py)。
|
||||
接口方法说明见：[Interfaces](./interfaces.md)，具体实现可参考：[llm.py](https://github.com/langgenius/dify-runtime/blob/main/lib/model_providers/anthropic/llm/llm.py)。
|
||||
|
||||
@ -1,4 +1,5 @@
|
||||
from collections.abc import Sequence
|
||||
from abc import ABC
|
||||
from collections.abc import Mapping, Sequence
|
||||
from enum import Enum, StrEnum
|
||||
from typing import Annotated, Any, Literal, Optional, Union
|
||||
|
||||
@ -60,8 +61,12 @@ class PromptMessageContentType(StrEnum):
|
||||
DOCUMENT = "document"
|
||||
|
||||
|
||||
class PromptMessageContent(BaseModel):
|
||||
pass
|
||||
class PromptMessageContent(ABC, BaseModel):
|
||||
"""
|
||||
Model class for prompt message content.
|
||||
"""
|
||||
|
||||
type: PromptMessageContentType
|
||||
|
||||
|
||||
class TextPromptMessageContent(PromptMessageContent):
|
||||
@ -125,7 +130,16 @@ PromptMessageContentUnionTypes = Annotated[
|
||||
]
|
||||
|
||||
|
||||
class PromptMessage(BaseModel):
|
||||
CONTENT_TYPE_MAPPING: Mapping[PromptMessageContentType, type[PromptMessageContent]] = {
|
||||
PromptMessageContentType.TEXT: TextPromptMessageContent,
|
||||
PromptMessageContentType.IMAGE: ImagePromptMessageContent,
|
||||
PromptMessageContentType.AUDIO: AudioPromptMessageContent,
|
||||
PromptMessageContentType.VIDEO: VideoPromptMessageContent,
|
||||
PromptMessageContentType.DOCUMENT: DocumentPromptMessageContent,
|
||||
}
|
||||
|
||||
|
||||
class PromptMessage(ABC, BaseModel):
|
||||
"""
|
||||
Model class for prompt message.
|
||||
"""
|
||||
@ -142,6 +156,23 @@ class PromptMessage(BaseModel):
|
||||
"""
|
||||
return not self.content
|
||||
|
||||
@field_validator("content", mode="before")
|
||||
@classmethod
|
||||
def validate_content(cls, v):
|
||||
if isinstance(v, list):
|
||||
prompts = []
|
||||
for prompt in v:
|
||||
if isinstance(prompt, PromptMessageContent):
|
||||
if not isinstance(prompt, TextPromptMessageContent | MultiModalPromptMessageContent):
|
||||
prompt = CONTENT_TYPE_MAPPING[prompt.type].model_validate(prompt.model_dump())
|
||||
elif isinstance(prompt, dict):
|
||||
prompt = CONTENT_TYPE_MAPPING[prompt["type"]].model_validate(prompt)
|
||||
else:
|
||||
raise ValueError(f"invalid prompt message {prompt}")
|
||||
prompts.append(prompt)
|
||||
return prompts
|
||||
return v
|
||||
|
||||
@field_serializer("content")
|
||||
def serialize_content(
|
||||
self, content: Optional[Union[str, Sequence[PromptMessageContent]]]
|
||||
|
||||
@ -24,7 +24,6 @@ from core.model_runtime.errors.invoke import (
|
||||
InvokeRateLimitError,
|
||||
InvokeServerUnavailableError,
|
||||
)
|
||||
from core.model_runtime.model_providers.__base.tokenizers.gpt2_tokenzier import GPT2Tokenizer
|
||||
from core.plugin.entities.plugin_daemon import PluginDaemonInnerError, PluginModelProviderEntity
|
||||
from core.plugin.impl.model import PluginModelClient
|
||||
|
||||
@ -253,15 +252,3 @@ class AIModel(BaseModel):
|
||||
raise Exception(f"Invalid model parameter rule name {name}")
|
||||
|
||||
return default_parameter_rule
|
||||
|
||||
def _get_num_tokens_by_gpt2(self, text: str) -> int:
|
||||
"""
|
||||
Get number of tokens for given prompt messages by gpt2
|
||||
Some provider models do not provide an interface for obtaining the number of tokens.
|
||||
Here, the gpt2 tokenizer is used to calculate the number of tokens.
|
||||
This method can be executed offline, and the gpt2 tokenizer has been cached in the project.
|
||||
|
||||
:param text: plain text of prompt. You need to convert the original message to plain text
|
||||
:return: number of tokens
|
||||
"""
|
||||
return GPT2Tokenizer.get_num_tokens(text)
|
||||
|
||||
@ -2,7 +2,7 @@ import logging
|
||||
import time
|
||||
import uuid
|
||||
from collections.abc import Generator, Sequence
|
||||
from typing import Optional, Union, cast
|
||||
from typing import Optional, Union
|
||||
|
||||
from pydantic import ConfigDict
|
||||
|
||||
@ -13,14 +13,15 @@ from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk,
|
||||
from core.model_runtime.entities.message_entities import (
|
||||
AssistantPromptMessage,
|
||||
PromptMessage,
|
||||
PromptMessageContentUnionTypes,
|
||||
PromptMessageTool,
|
||||
TextPromptMessageContent,
|
||||
)
|
||||
from core.model_runtime.entities.model_entities import (
|
||||
ModelType,
|
||||
PriceType,
|
||||
)
|
||||
from core.model_runtime.model_providers.__base.ai_model import AIModel
|
||||
from core.model_runtime.utils.helper import convert_llm_result_chunk_to_str
|
||||
from core.plugin.impl.model import PluginModelClient
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
@ -238,7 +239,7 @@ class LargeLanguageModel(AIModel):
|
||||
def _invoke_result_generator(
|
||||
self,
|
||||
model: str,
|
||||
result: Generator,
|
||||
result: Generator[LLMResultChunk, None, None],
|
||||
credentials: dict,
|
||||
prompt_messages: Sequence[PromptMessage],
|
||||
model_parameters: dict,
|
||||
@ -255,11 +256,21 @@ class LargeLanguageModel(AIModel):
|
||||
:return: result generator
|
||||
"""
|
||||
callbacks = callbacks or []
|
||||
assistant_message = AssistantPromptMessage(content="")
|
||||
message_content: list[PromptMessageContentUnionTypes] = []
|
||||
usage = None
|
||||
system_fingerprint = None
|
||||
real_model = model
|
||||
|
||||
def _update_message_content(content: str | list[PromptMessageContentUnionTypes] | None):
|
||||
if not content:
|
||||
return
|
||||
if isinstance(content, list):
|
||||
message_content.extend(content)
|
||||
return
|
||||
if isinstance(content, str):
|
||||
message_content.append(TextPromptMessageContent(data=content))
|
||||
return
|
||||
|
||||
try:
|
||||
for chunk in result:
|
||||
# Following https://github.com/langgenius/dify/issues/17799,
|
||||
@ -281,9 +292,8 @@ class LargeLanguageModel(AIModel):
|
||||
callbacks=callbacks,
|
||||
)
|
||||
|
||||
text = convert_llm_result_chunk_to_str(chunk.delta.message.content)
|
||||
current_content = cast(str, assistant_message.content)
|
||||
assistant_message.content = current_content + text
|
||||
_update_message_content(chunk.delta.message.content)
|
||||
|
||||
real_model = chunk.model
|
||||
if chunk.delta.usage:
|
||||
usage = chunk.delta.usage
|
||||
@ -293,6 +303,7 @@ class LargeLanguageModel(AIModel):
|
||||
except Exception as e:
|
||||
raise self._transform_invoke_error(e)
|
||||
|
||||
assistant_message = AssistantPromptMessage(content=message_content)
|
||||
self._trigger_after_invoke_callbacks(
|
||||
model=model,
|
||||
result=LLMResult(
|
||||
|
||||
@ -30,6 +30,8 @@ class GPT2Tokenizer:
|
||||
@staticmethod
|
||||
def get_encoder() -> Any:
|
||||
global _tokenizer, _lock
|
||||
if _tokenizer is not None:
|
||||
return _tokenizer
|
||||
with _lock:
|
||||
if _tokenizer is None:
|
||||
# Try to use tiktoken to get the tokenizer because it is faster
|
||||
|
||||
@ -1,8 +1,6 @@
|
||||
import pydantic
|
||||
from pydantic import BaseModel
|
||||
|
||||
from core.model_runtime.entities.message_entities import PromptMessageContentUnionTypes
|
||||
|
||||
|
||||
def dump_model(model: BaseModel) -> dict:
|
||||
if hasattr(pydantic, "model_dump"):
|
||||
@ -10,18 +8,3 @@ def dump_model(model: BaseModel) -> dict:
|
||||
return pydantic.model_dump(model) # type: ignore
|
||||
else:
|
||||
return model.model_dump()
|
||||
|
||||
|
||||
def convert_llm_result_chunk_to_str(content: None | str | list[PromptMessageContentUnionTypes]) -> str:
|
||||
if content is None:
|
||||
message_text = ""
|
||||
elif isinstance(content, str):
|
||||
message_text = content
|
||||
elif isinstance(content, list):
|
||||
# Assuming the list contains PromptMessageContent objects with a "data" attribute
|
||||
message_text = "".join(
|
||||
item.data if hasattr(item, "data") and isinstance(item.data, str) else str(item) for item in content
|
||||
)
|
||||
else:
|
||||
message_text = str(content)
|
||||
return message_text
|
||||
|
||||
@ -1 +1 @@
|
||||
3
|
||||
3
|
||||
|
||||
@ -1 +1 @@
|
||||
2
|
||||
2
|
||||
|
||||
@ -1 +1 @@
|
||||
1
|
||||
1
|
||||
|
||||
@ -239,8 +239,8 @@ class PluginModelBackwardsInvocation(BaseBackwardsInvocation):
|
||||
content = payload.text
|
||||
|
||||
SUMMARY_PROMPT = """You are a professional language researcher, you are interested in the language
|
||||
and you can quickly aimed at the main point of an webpage and reproduce it in your own words but
|
||||
retain the original meaning and keep the key points.
|
||||
and you can quickly aimed at the main point of an webpage and reproduce it in your own words but
|
||||
retain the original meaning and keep the key points.
|
||||
however, the text you got is too long, what you got is possible a part of the text.
|
||||
Please summarize the text you got.
|
||||
|
||||
|
||||
@ -10,4 +10,4 @@
|
||||
],
|
||||
"query_prompt": "\n\n用户：{{#query#}}",
|
||||
"stops": ["用户："]
|
||||
}
|
||||
}
|
||||
|
||||
@ -6,4 +6,4 @@
|
||||
],
|
||||
"query_prompt": "{{#query#}}",
|
||||
"stops": null
|
||||
}
|
||||
}
|
||||
|
||||
@ -6,4 +6,4 @@
|
||||
],
|
||||
"query_prompt": "{{#query#}}",
|
||||
"stops": null
|
||||
}
|
||||
}
|
||||
|
||||
@ -156,8 +156,8 @@ class AnalyticdbVectorBySql:
|
||||
values = []
|
||||
id_prefix = str(uuid.uuid4()) + "_"
|
||||
sql = f"""
|
||||
INSERT INTO {self.table_name}
|
||||
(id, ref_doc_id, vector, page_content, metadata_, to_tsvector)
|
||||
INSERT INTO {self.table_name}
|
||||
(id, ref_doc_id, vector, page_content, metadata_, to_tsvector)
|
||||
VALUES (%s, %s, %s, %s, %s, to_tsvector('zh_cn', %s));
|
||||
"""
|
||||
for i, doc in enumerate(documents):
|
||||
@ -242,7 +242,7 @@ class AnalyticdbVectorBySql:
|
||||
where_clause += f"AND metadata_->>'document_id' IN ({document_ids})"
|
||||
with self._get_cursor() as cur:
|
||||
cur.execute(
|
||||
f"""SELECT id, vector, page_content, metadata_,
|
||||
f"""SELECT id, vector, page_content, metadata_,
|
||||
ts_rank(to_tsvector, to_tsquery_from_text(%s, 'zh_cn'), 32) AS score
|
||||
FROM {self.table_name}
|
||||
WHERE to_tsvector@@to_tsquery_from_text(%s, 'zh_cn') {where_clause}
|
||||
|
||||
@ -27,8 +27,8 @@ class MilvusConfig(BaseModel):
|
||||
|
||||
uri: str # Milvus server URI
|
||||
token: Optional[str] = None # Optional token for authentication
|
||||
user: str # Username for authentication
|
||||
password: str # Password for authentication
|
||||
user: Optional[str] = None # Username for authentication
|
||||
password: Optional[str] = None # Password for authentication
|
||||
batch_size: int = 100 # Batch size for operations
|
||||
database: str = "default" # Database name
|
||||
enable_hybrid_search: bool = False # Flag to enable hybrid search
|
||||
@ -43,10 +43,11 @@ class MilvusConfig(BaseModel):
|
||||
"""
|
||||
if not values.get("uri"):
|
||||
raise ValueError("config MILVUS_URI is required")
|
||||
if not values.get("user"):
|
||||
raise ValueError("config MILVUS_USER is required")
|
||||
if not values.get("password"):
|
||||
raise ValueError("config MILVUS_PASSWORD is required")
|
||||
if not values.get("token"):
|
||||
if not values.get("user"):
|
||||
raise ValueError("config MILVUS_USER is required")
|
||||
if not values.get("password"):
|
||||
raise ValueError("config MILVUS_PASSWORD is required")
|
||||
return values
|
||||
|
||||
def to_milvus_params(self):
|
||||
@ -356,11 +357,14 @@ class MilvusVector(BaseVector):
|
||||
)
|
||||
redis_client.set(collection_exist_cache_key, 1, ex=3600)
|
||||
|
||||
def _init_client(self, config) -> MilvusClient:
|
||||
def _init_client(self, config: MilvusConfig) -> MilvusClient:
|
||||
"""
|
||||
Initialize and return a Milvus client.
|
||||
"""
|
||||
client = MilvusClient(uri=config.uri, user=config.user, password=config.password, db_name=config.database)
|
||||
if config.token:
|
||||
client = MilvusClient(uri=config.uri, token=config.token, db_name=config.database)
|
||||
else:
|
||||
client = MilvusClient(uri=config.uri, user=config.user, password=config.password, db_name=config.database)
|
||||
return client
|
||||
|
||||
|
||||
|
||||
@ -203,7 +203,7 @@ class OceanBaseVector(BaseVector):
|
||||
|
||||
full_sql = f"""SELECT metadata, text, MATCH (text) AGAINST (:query) AS score
|
||||
FROM {self._collection_name}
|
||||
WHERE MATCH (text) AGAINST (:query) > 0
|
||||
WHERE MATCH (text) AGAINST (:query) > 0
|
||||
{where_clause}
|
||||
ORDER BY score DESC
|
||||
LIMIT {top_k}"""
|
||||
|
||||
@ -59,12 +59,12 @@ CREATE TABLE IF NOT EXISTS {table_name} (
|
||||
"""
|
||||
|
||||
SQL_CREATE_INDEX_PQ = """
|
||||
CREATE INDEX IF NOT EXISTS embedding_{table_name}_pq_idx ON {table_name}
|
||||
CREATE INDEX IF NOT EXISTS embedding_{table_name}_pq_idx ON {table_name}
|
||||
USING hnsw (embedding vector_cosine_ops) WITH (m = 16, ef_construction = 64, enable_pq=on, pq_m={pq_m});
|
||||
"""
|
||||
|
||||
SQL_CREATE_INDEX = """
|
||||
CREATE INDEX IF NOT EXISTS embedding_cosine_{table_name}_idx ON {table_name}
|
||||
CREATE INDEX IF NOT EXISTS embedding_cosine_{table_name}_idx ON {table_name}
|
||||
USING hnsw (embedding vector_cosine_ops) WITH (m = 16, ef_construction = 64);
|
||||
"""
|
||||
|
||||
|
||||
@ -1,10 +1,9 @@
|
||||
import json
|
||||
import logging
|
||||
import ssl
|
||||
from typing import Any, Optional
|
||||
from typing import Any, Literal, Optional
|
||||
from uuid import uuid4
|
||||
|
||||
from opensearchpy import OpenSearch, helpers
|
||||
from opensearchpy import OpenSearch, Urllib3AWSV4SignerAuth, Urllib3HttpConnection, helpers
|
||||
from opensearchpy.helpers import BulkIndexError
|
||||
from pydantic import BaseModel, model_validator
|
||||
|
||||
@ -24,9 +23,12 @@ logger = logging.getLogger(__name__)
|
||||
class OpenSearchConfig(BaseModel):
|
||||
host: str
|
||||
port: int
|
||||
secure: bool = False
|
||||
auth_method: Literal["basic", "aws_managed_iam"] = "basic"
|
||||
user: Optional[str] = None
|
||||
password: Optional[str] = None
|
||||
secure: bool = False
|
||||
aws_region: Optional[str] = None
|
||||
aws_service: Optional[str] = None
|
||||
|
||||
@model_validator(mode="before")
|
||||
@classmethod
|
||||
@ -35,24 +37,40 @@ class OpenSearchConfig(BaseModel):
|
||||
raise ValueError("config OPENSEARCH_HOST is required")
|
||||
if not values.get("port"):
|
||||
raise ValueError("config OPENSEARCH_PORT is required")
|
||||
if values.get("auth_method") == "aws_managed_iam":
|
||||
if not values.get("aws_region"):
|
||||
raise ValueError("config OPENSEARCH_AWS_REGION is required for AWS_MANAGED_IAM auth method")
|
||||
if not values.get("aws_service"):
|
||||
raise ValueError("config OPENSEARCH_AWS_SERVICE is required for AWS_MANAGED_IAM auth method")
|
||||
return values
|
||||
|
||||
def create_ssl_context(self) -> ssl.SSLContext:
|
||||
ssl_context = ssl.create_default_context()
|
||||
ssl_context.check_hostname = False
|
||||
ssl_context.verify_mode = ssl.CERT_NONE # Disable Certificate Validation
|
||||
return ssl_context
|
||||
def create_aws_managed_iam_auth(self) -> Urllib3AWSV4SignerAuth:
|
||||
import boto3 # type: ignore
|
||||
|
||||
return Urllib3AWSV4SignerAuth(
|
||||
credentials=boto3.Session().get_credentials(),
|
||||
region=self.aws_region,
|
||||
service=self.aws_service, # type: ignore[arg-type]
|
||||
)
|
||||
|
||||
def to_opensearch_params(self) -> dict[str, Any]:
|
||||
params = {
|
||||
"hosts": [{"host": self.host, "port": self.port}],
|
||||
"use_ssl": self.secure,
|
||||
"verify_certs": self.secure,
|
||||
"connection_class": Urllib3HttpConnection,
|
||||
"pool_maxsize": 20,
|
||||
}
|
||||
if self.user and self.password:
|
||||
|
||||
if self.auth_method == "basic":
|
||||
logger.info("Using basic authentication for OpenSearch Vector DB")
|
||||
|
||||
params["http_auth"] = (self.user, self.password)
|
||||
if self.secure:
|
||||
params["ssl_context"] = self.create_ssl_context()
|
||||
elif self.auth_method == "aws_managed_iam":
|
||||
logger.info("Using AWS managed IAM role for OpenSearch Vector DB")
|
||||
|
||||
params["http_auth"] = self.create_aws_managed_iam_auth()
|
||||
|
||||
return params
|
||||
|
||||
|
||||
@ -76,16 +94,23 @@ class OpenSearchVector(BaseVector):
|
||||
action = {
|
||||
"_op_type": "index",
|
||||
"_index": self._collection_name.lower(),
|
||||
"_id": uuid4().hex,
|
||||
"_source": {
|
||||
Field.CONTENT_KEY.value: documents[i].page_content,
|
||||
Field.VECTOR.value: embeddings[i], # Make sure you pass an array here
|
||||
Field.METADATA_KEY.value: documents[i].metadata,
|
||||
},
|
||||
}
|
||||
# See https://github.com/langchain-ai/langchainjs/issues/4346#issuecomment-1935123377
|
||||
if self._client_config.aws_service not in ["aoss"]:
|
||||
action["_id"] = uuid4().hex
|
||||
actions.append(action)
|
||||
|
||||
helpers.bulk(self._client, actions)
|
||||
helpers.bulk(
|
||||
client=self._client,
|
||||
actions=actions,
|
||||
timeout=30,
|
||||
max_retries=3,
|
||||
)
|
||||
|
||||
def get_ids_by_metadata_field(self, key: str, value: str):
|
||||
query = {"query": {"term": {f"{Field.METADATA_KEY.value}.{key}": value}}}
|
||||
@ -234,6 +259,7 @@ class OpenSearchVector(BaseVector):
|
||||
},
|
||||
}
|
||||
|
||||
logger.info(f"Creating OpenSearch index {self._collection_name.lower()}")
|
||||
self._client.indices.create(index=self._collection_name.lower(), body=index_body)
|
||||
|
||||
redis_client.set(collection_exist_cache_key, 1, ex=3600)
|
||||
@ -252,9 +278,12 @@ class OpenSearchVectorFactory(AbstractVectorFactory):
|
||||
open_search_config = OpenSearchConfig(
|
||||
host=dify_config.OPENSEARCH_HOST or "localhost",
|
||||
port=dify_config.OPENSEARCH_PORT,
|
||||
secure=dify_config.OPENSEARCH_SECURE,
|
||||
auth_method=dify_config.OPENSEARCH_AUTH_METHOD.value,
|
||||
user=dify_config.OPENSEARCH_USER,
|
||||
password=dify_config.OPENSEARCH_PASSWORD,
|
||||
secure=dify_config.OPENSEARCH_SECURE,
|
||||
aws_region=dify_config.OPENSEARCH_AWS_REGION,
|
||||
aws_service=dify_config.OPENSEARCH_AWS_SERVICE,
|
||||
)
|
||||
|
||||
return OpenSearchVector(collection_name=collection_name, config=open_search_config)
|
||||
|
||||
@ -59,8 +59,8 @@ CREATE TABLE IF NOT EXISTS {table_name} (
|
||||
)
|
||||
"""
|
||||
SQL_CREATE_INDEX = """
|
||||
CREATE INDEX IF NOT EXISTS idx_docs_{table_name} ON {table_name}(text)
|
||||
INDEXTYPE IS CTXSYS.CONTEXT PARAMETERS
|
||||
CREATE INDEX IF NOT EXISTS idx_docs_{table_name} ON {table_name}(text)
|
||||
INDEXTYPE IS CTXSYS.CONTEXT PARAMETERS
|
||||
('FILTER CTXSYS.NULL_FILTER SECTION GROUP CTXSYS.HTML_SECTION_GROUP LEXER world_lexer')
|
||||
"""
|
||||
|
||||
@ -164,7 +164,7 @@ class OracleVector(BaseVector):
|
||||
with conn.cursor() as cur:
|
||||
try:
|
||||
cur.execute(
|
||||
f"""INSERT INTO {self.table_name} (id, text, meta, embedding)
|
||||
f"""INSERT INTO {self.table_name} (id, text, meta, embedding)
|
||||
VALUES (:1, :2, :3, :4)""",
|
||||
value,
|
||||
)
|
||||
@ -227,8 +227,8 @@ class OracleVector(BaseVector):
|
||||
conn.outputtypehandler = self.output_type_handler
|
||||
with conn.cursor() as cur:
|
||||
cur.execute(
|
||||
f"""SELECT meta, text, vector_distance(embedding,(select to_vector(:1) from dual),cosine)
|
||||
AS distance FROM {self.table_name}
|
||||
f"""SELECT meta, text, vector_distance(embedding,(select to_vector(:1) from dual),cosine)
|
||||
AS distance FROM {self.table_name}
|
||||
{where_clause} ORDER BY distance fetch first {top_k} rows only""",
|
||||
[numpy.array(query_vector)],
|
||||
)
|
||||
@ -290,7 +290,7 @@ class OracleVector(BaseVector):
|
||||
document_ids = ", ".join(f"'{id}'" for id in document_ids_filter)
|
||||
where_clause = f" AND metadata->>'document_id' in ({document_ids}) "
|
||||
cur.execute(
|
||||
f"""select meta, text, embedding FROM {self.table_name}
|
||||
f"""select meta, text, embedding FROM {self.table_name}
|
||||
WHERE CONTAINS(text, :kk, 1) > 0 {where_clause}
|
||||
order by score(1) desc fetch first {top_k} rows only""",
|
||||
kk=" ACCUM ".join(entities),
|
||||
|
||||
@ -61,7 +61,7 @@ CREATE TABLE IF NOT EXISTS {table_name} (
|
||||
"""
|
||||
|
||||
SQL_CREATE_INDEX = """
|
||||
CREATE INDEX IF NOT EXISTS embedding_cosine_v1_idx ON {table_name}
|
||||
CREATE INDEX IF NOT EXISTS embedding_cosine_v1_idx ON {table_name}
|
||||
USING hnsw (embedding vector_cosine_ops) WITH (m = 16, ef_construction = 64);
|
||||
"""
|
||||
|
||||
|
||||
@ -58,7 +58,7 @@ CREATE TABLE IF NOT EXISTS {table_name} (
|
||||
"""
|
||||
|
||||
SQL_CREATE_INDEX = """
|
||||
CREATE INDEX IF NOT EXISTS embedding_cosine_v1_idx ON {table_name}
|
||||
CREATE INDEX IF NOT EXISTS embedding_cosine_v1_idx ON {table_name}
|
||||
USING hnsw (embedding floatvector_cosine_ops) WITH (m = 16, ef_construction = 64);
|
||||
"""
|
||||
|
||||
|
||||
@ -205,9 +205,9 @@ class TiDBVector(BaseVector):
|
||||
|
||||
with Session(self._engine) as session:
|
||||
select_statement = sql_text(f"""
|
||||
SELECT meta, text, distance
|
||||
SELECT meta, text, distance
|
||||
FROM (
|
||||
SELECT
|
||||
SELECT
|
||||
meta,
|
||||
text,
|
||||
{tidb_dist_func}(vector, :query_vector_str) AS distance
|
||||
|
||||
@ -52,14 +52,16 @@ class RerankModelRunner(BaseRerankRunner):
|
||||
rerank_documents = []
|
||||
|
||||
for result in rerank_result.docs:
|
||||
# format document
|
||||
rerank_document = Document(
|
||||
page_content=result.text,
|
||||
metadata=documents[result.index].metadata,
|
||||
provider=documents[result.index].provider,
|
||||
)
|
||||
if rerank_document.metadata is not None:
|
||||
rerank_document.metadata["score"] = result.score
|
||||
rerank_documents.append(rerank_document)
|
||||
if score_threshold is None or result.score >= score_threshold:
|
||||
# format document
|
||||
rerank_document = Document(
|
||||
page_content=result.text,
|
||||
metadata=documents[result.index].metadata,
|
||||
provider=documents[result.index].provider,
|
||||
)
|
||||
if rerank_document.metadata is not None:
|
||||
rerank_document.metadata["score"] = result.score
|
||||
rerank_documents.append(rerank_document)
|
||||
|
||||
return rerank_documents
|
||||
rerank_documents.sort(key=lambda x: x.metadata.get("score", 0.0), reverse=True)
|
||||
return rerank_documents[:top_n] if top_n else rerank_documents
|
||||
|
||||
@ -7,7 +7,7 @@ from collections.abc import Generator, Mapping
|
||||
from typing import Any, Optional, Union, cast
|
||||
|
||||
from flask import Flask, current_app
|
||||
from sqlalchemy import Integer, and_, or_, text
|
||||
from sqlalchemy import Float, and_, or_, text
|
||||
from sqlalchemy import cast as sqlalchemy_cast
|
||||
|
||||
from core.app.app_config.entities import (
|
||||
@ -1005,28 +1005,24 @@ class DatasetRetrieval:
|
||||
if isinstance(value, str):
|
||||
filters.append(DatasetDocument.doc_metadata[metadata_name] == f'"{value}"')
|
||||
else:
|
||||
filters.append(
|
||||
sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Integer) == value
|
||||
)
|
||||
filters.append(sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Float) == value)
|
||||
case "is not" | "≠":
|
||||
if isinstance(value, str):
|
||||
filters.append(DatasetDocument.doc_metadata[metadata_name] != f'"{value}"')
|
||||
else:
|
||||
filters.append(
|
||||
sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Integer) != value
|
||||
)
|
||||
filters.append(sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Float) != value)
|
||||
case "empty":
|
||||
filters.append(DatasetDocument.doc_metadata[metadata_name].is_(None))
|
||||
case "not empty":
|
||||
filters.append(DatasetDocument.doc_metadata[metadata_name].isnot(None))
|
||||
case "before" | "<":
|
||||
filters.append(sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Integer) < value)
|
||||
filters.append(sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Float) < value)
|
||||
case "after" | ">":
|
||||
filters.append(sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Integer) > value)
|
||||
filters.append(sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Float) > value)
|
||||
case "≤" | "<=":
|
||||
filters.append(sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Integer) <= value)
|
||||
filters.append(sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Float) <= value)
|
||||
case "≥" | ">=":
|
||||
filters.append(sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Integer) >= value)
|
||||
filters.append(sqlalchemy_cast(DatasetDocument.doc_metadata[metadata_name].astext, Float) >= value)
|
||||
case _:
|
||||
pass
|
||||
return filters
|
||||
|
||||
@ -2,7 +2,7 @@ METADATA_FILTER_SYSTEM_PROMPT = """
|
||||
### Job Description',
|
||||
You are a text metadata extract engine that extract text's metadata based on user input and set the metadata value
|
||||
### Task
|
||||
Your task is to ONLY extract the metadatas that exist in the input text from the provided metadata list and Use the following operators ["=", "!=", ">", "<", ">=", "<="] to express logical relationships, then return result in JSON format with the key "metadata_fields" and value "metadata_field_value" and comparison operator "comparison_operator".
|
||||
Your task is to ONLY extract the metadatas that exist in the input text from the provided metadata list and Use the following operators ["contains", "not contains", "start with", "end with", "is", "is not", "empty", "not empty", "=", "≠", ">", "<", "≥", "≤", "before", "after"] to express logical relationships, then return result in JSON format with the key "metadata_fields" and value "metadata_field_value" and comparison operator "comparison_operator".
|
||||
### Format
|
||||
The input text is in the variable input_text. Metadata are specified as a list in the variable metadata_fields.
|
||||
### Constraint
|
||||
@ -50,7 +50,7 @@ You are a text metadata extract engine that extract text's metadata based on use
|
||||
# Your task is to ONLY extract the metadatas that exist in the input text from the provided metadata list and Use the following operators ["=", "!=", ">", "<", ">=", "<="] to express logical relationships, then return result in JSON format with the key "metadata_fields" and value "metadata_field_value" and comparison operator "comparison_operator".
|
||||
### Format
|
||||
The input text is in the variable input_text. Metadata are specified as a list in the variable metadata_fields.
|
||||
### Constraint
|
||||
### Constraint
|
||||
DO NOT include anything other than the JSON array in your response.
|
||||
### Example
|
||||
Here is the chat example between human and assistant, inside <example></example> XML tags.
|
||||
@ -59,7 +59,7 @@ User:{{"input_text": ["I want to know which company’s email address test@examp
|
||||
Assistant:{{"metadata_map": [{{"metadata_field_name": "email", "metadata_field_value": "test@example.com", "comparison_operator": "="}}]}}
|
||||
User:{{"input_text": "What are the movies with a score of more than 9 in 2024?", "metadata_fields": ["name", "year", "rating", "country"]}}
|
||||
Assistant:{{"metadata_map": [{{"metadata_field_name": "year", "metadata_field_value": "2024", "comparison_operator": "="}, {{"metadata_field_name": "rating", "metadata_field_value": "9", "comparison_operator": ">"}}]}}
|
||||
</example>
|
||||
</example>
|
||||
### User Input
|
||||
{{"input_text" : "{input_text}", "metadata_fields" : {metadata_fields}}}
|
||||
### Assistant Output
|
||||
|
||||
@ -159,50 +159,6 @@ class TextSplitter(BaseDocumentTransformer, ABC):
|
||||
)
|
||||
return cls(length_function=lambda x: [_huggingface_tokenizer_length(text) for text in x], **kwargs)
|
||||
|
||||
@classmethod
|
||||
def from_tiktoken_encoder(
|
||||
cls: type[TS],
|
||||
encoding_name: str = "gpt2",
|
||||
model_name: Optional[str] = None,
|
||||
allowed_special: Union[Literal["all"], Set[str]] = set(),
|
||||
disallowed_special: Union[Literal["all"], Collection[str]] = "all",
|
||||
**kwargs: Any,
|
||||
) -> TS:
|
||||
"""Text splitter that uses tiktoken encoder to count length."""
|
||||
try:
|
||||
import tiktoken
|
||||
except ImportError:
|
||||
raise ImportError(
|
||||
"Could not import tiktoken python package. "
|
||||
"This is needed in order to calculate max_tokens_for_prompt. "
|
||||
"Please install it with `pip install tiktoken`."
|
||||
)
|
||||
|
||||
if model_name is not None:
|
||||
enc = tiktoken.encoding_for_model(model_name)
|
||||
else:
|
||||
enc = tiktoken.get_encoding(encoding_name)
|
||||
|
||||
def _tiktoken_encoder(text: str) -> int:
|
||||
return len(
|
||||
enc.encode(
|
||||
text,
|
||||
allowed_special=allowed_special,
|
||||
disallowed_special=disallowed_special,
|
||||
)
|
||||
)
|
||||
|
||||
if issubclass(cls, TokenTextSplitter):
|
||||
extra_kwargs = {
|
||||
"encoding_name": encoding_name,
|
||||
"model_name": model_name,
|
||||
"allowed_special": allowed_special,
|
||||
"disallowed_special": disallowed_special,
|
||||
}
|
||||
kwargs = {**kwargs, **extra_kwargs}
|
||||
|
||||
return cls(length_function=lambda x: [_tiktoken_encoder(text) for text in x], **kwargs)
|
||||
|
||||
def transform_documents(self, documents: Sequence[Document], **kwargs: Any) -> Sequence[Document]:
|
||||
"""Transform sequence of documents by splitting them."""
|
||||
return self.split_documents(list(documents))
|
||||
|
||||
@ -6,8 +6,8 @@ from core.tools.entities.tool_entities import ToolProviderType
|
||||
from core.tools.utils.model_invocation_utils import ModelInvocationUtils
|
||||
|
||||
_SUMMARY_PROMPT = """You are a professional language researcher, you are interested in the language
|
||||
and you can quickly aimed at the main point of an webpage and reproduce it in your own words but
|
||||
retain the original meaning and keep the key points.
|
||||
and you can quickly aimed at the main point of an webpage and reproduce it in your own words but
|
||||
retain the original meaning and keep the key points.
|
||||
however, the text you got is too long, what you got is possible a part of the text.
|
||||
Please summarize the text you got.
|
||||
"""
|
||||
|
||||
41
api/core/tools/signature.py
Normal file
41
api/core/tools/signature.py
Normal file
@ -0,0 +1,41 @@
|
||||
import base64
|
||||
import hashlib
|
||||
import hmac
|
||||
import os
|
||||
import time
|
||||
|
||||
from configs import dify_config
|
||||
|
||||
|
||||
def sign_tool_file(tool_file_id: str, extension: str) -> str:
|
||||
"""
|
||||
sign file to get a temporary url
|
||||
"""
|
||||
base_url = dify_config.FILES_URL
|
||||
file_preview_url = f"{base_url}/files/tools/{tool_file_id}{extension}"
|
||||
|
||||
timestamp = str(int(time.time()))
|
||||
nonce = os.urandom(16).hex()
|
||||
data_to_sign = f"file-preview|{tool_file_id}|{timestamp}|{nonce}"
|
||||
secret_key = dify_config.SECRET_KEY.encode() if dify_config.SECRET_KEY else b""
|
||||
sign = hmac.new(secret_key, data_to_sign.encode(), hashlib.sha256).digest()
|
||||
encoded_sign = base64.urlsafe_b64encode(sign).decode()
|
||||
|
||||
return f"{file_preview_url}?timestamp={timestamp}&nonce={nonce}&sign={encoded_sign}"
|
||||
|
||||
|
||||
def verify_tool_file_signature(file_id: str, timestamp: str, nonce: str, sign: str) -> bool:
|
||||
"""
|
||||
verify signature
|
||||
"""
|
||||
data_to_sign = f"file-preview|{file_id}|{timestamp}|{nonce}"
|
||||
secret_key = dify_config.SECRET_KEY.encode() if dify_config.SECRET_KEY else b""
|
||||
recalculated_sign = hmac.new(secret_key, data_to_sign.encode(), hashlib.sha256).digest()
|
||||
recalculated_encoded_sign = base64.urlsafe_b64encode(recalculated_sign).decode()
|
||||
|
||||
# verify signature
|
||||
if sign != recalculated_encoded_sign:
|
||||
return False
|
||||
|
||||
current_time = int(time.time())
|
||||
return current_time - int(timestamp) <= dify_config.FILES_ACCESS_TIMEOUT
|
||||
@ -4,23 +4,34 @@ import hmac
|
||||
import logging
|
||||
import os
|
||||
import time
|
||||
from collections.abc import Generator
|
||||
from mimetypes import guess_extension, guess_type
|
||||
from typing import Optional, Union
|
||||
from uuid import uuid4
|
||||
|
||||
import httpx
|
||||
from sqlalchemy.orm import Session
|
||||
|
||||
from configs import dify_config
|
||||
from core.helper import ssrf_proxy
|
||||
from extensions.ext_database import db
|
||||
from extensions.ext_database import db as global_db
|
||||
from extensions.ext_storage import storage
|
||||
from models.model import MessageFile
|
||||
from models.tools import ToolFile
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
from sqlalchemy.engine import Engine
|
||||
|
||||
|
||||
class ToolFileManager:
|
||||
_engine: Engine
|
||||
|
||||
def __init__(self, engine: Engine | None = None):
|
||||
if engine is None:
|
||||
engine = global_db.engine
|
||||
self._engine = engine
|
||||
|
||||
@staticmethod
|
||||
def sign_file(tool_file_id: str, extension: str) -> str:
|
||||
"""
|
||||
@ -55,8 +66,8 @@ class ToolFileManager:
|
||||
current_time = int(time.time())
|
||||
return current_time - int(timestamp) <= dify_config.FILES_ACCESS_TIMEOUT
|
||||
|
||||
@staticmethod
|
||||
def create_file_by_raw(
|
||||
self,
|
||||
*,
|
||||
user_id: str,
|
||||
tenant_id: str,
|
||||
@ -77,24 +88,25 @@ class ToolFileManager:
|
||||
filepath = f"tools/{tenant_id}/{unique_filename}"
|
||||
storage.save(filepath, file_binary)
|
||||
|
||||
tool_file = ToolFile(
|
||||
user_id=user_id,
|
||||
tenant_id=tenant_id,
|
||||
conversation_id=conversation_id,
|
||||
file_key=filepath,
|
||||
mimetype=mimetype,
|
||||
name=present_filename,
|
||||
size=len(file_binary),
|
||||
)
|
||||
with Session(self._engine, expire_on_commit=False) as session:
|
||||
tool_file = ToolFile(
|
||||
user_id=user_id,
|
||||
tenant_id=tenant_id,
|
||||
conversation_id=conversation_id,
|
||||
file_key=filepath,
|
||||
mimetype=mimetype,
|
||||
name=present_filename,
|
||||
size=len(file_binary),
|
||||
)
|
||||
|
||||
db.session.add(tool_file)
|
||||
db.session.commit()
|
||||
db.session.refresh(tool_file)
|
||||
session.add(tool_file)
|
||||
session.commit()
|
||||
session.refresh(tool_file)
|
||||
|
||||
return tool_file
|
||||
|
||||
@staticmethod
|
||||
def create_file_by_url(
|
||||
self,
|
||||
user_id: str,
|
||||
tenant_id: str,
|
||||
file_url: str,
|
||||
@ -119,24 +131,24 @@ class ToolFileManager:
|
||||
filepath = f"tools/{tenant_id}/{filename}"
|
||||
storage.save(filepath, blob)
|
||||
|
||||
tool_file = ToolFile(
|
||||
user_id=user_id,
|
||||
tenant_id=tenant_id,
|
||||
conversation_id=conversation_id,
|
||||
file_key=filepath,
|
||||
mimetype=mimetype,
|
||||
original_url=file_url,
|
||||
name=filename,
|
||||
size=len(blob),
|
||||
)
|
||||
with Session(self._engine, expire_on_commit=False) as session:
|
||||
tool_file = ToolFile(
|
||||
user_id=user_id,
|
||||
tenant_id=tenant_id,
|
||||
conversation_id=conversation_id,
|
||||
file_key=filepath,
|
||||
mimetype=mimetype,
|
||||
original_url=file_url,
|
||||
name=filename,
|
||||
size=len(blob),
|
||||
)
|
||||
|
||||
db.session.add(tool_file)
|
||||
db.session.commit()
|
||||
session.add(tool_file)
|
||||
session.commit()
|
||||
|
||||
return tool_file
|
||||
|
||||
@staticmethod
|
||||
def get_file_binary(id: str) -> Union[tuple[bytes, str], None]:
|
||||
def get_file_binary(self, id: str) -> Union[tuple[bytes, str], None]:
|
||||
"""
|
||||
get file binary
|
||||
|
||||
@ -144,13 +156,14 @@ class ToolFileManager:
|
||||
|
||||
:return: the binary of the file, mime type
|
||||
"""
|
||||
tool_file: ToolFile | None = (
|
||||
db.session.query(ToolFile)
|
||||
.filter(
|
||||
ToolFile.id == id,
|
||||
with Session(self._engine, expire_on_commit=False) as session:
|
||||
tool_file: ToolFile | None = (
|
||||
session.query(ToolFile)
|
||||
.filter(
|
||||
ToolFile.id == id,
|
||||
)
|
||||
.first()
|
||||
)
|
||||
.first()
|
||||
)
|
||||
|
||||
if not tool_file:
|
||||
return None
|
||||
@ -159,8 +172,7 @@ class ToolFileManager:
|
||||
|
||||
return blob, tool_file.mimetype
|
||||
|
||||
@staticmethod
|
||||
def get_file_binary_by_message_file_id(id: str) -> Union[tuple[bytes, str], None]:
|
||||
def get_file_binary_by_message_file_id(self, id: str) -> Union[tuple[bytes, str], None]:
|
||||
"""
|
||||
get file binary
|
||||
|
||||
@ -168,33 +180,34 @@ class ToolFileManager:
|
||||
|
||||
:return: the binary of the file, mime type
|
||||
"""
|
||||
message_file: MessageFile | None = (
|
||||
db.session.query(MessageFile)
|
||||
.filter(
|
||||
MessageFile.id == id,
|
||||
with Session(self._engine, expire_on_commit=False) as session:
|
||||
message_file: MessageFile | None = (
|
||||
session.query(MessageFile)
|
||||
.filter(
|
||||
MessageFile.id == id,
|
||||
)
|
||||
.first()
|
||||
)
|
||||
.first()
|
||||
)
|
||||
|
||||
# Check if message_file is not None
|
||||
if message_file is not None:
|
||||
# get tool file id
|
||||
if message_file.url is not None:
|
||||
tool_file_id = message_file.url.split("/")[-1]
|
||||
# trim extension
|
||||
tool_file_id = tool_file_id.split(".")[0]
|
||||
# Check if message_file is not None
|
||||
if message_file is not None:
|
||||
# get tool file id
|
||||
if message_file.url is not None:
|
||||
tool_file_id = message_file.url.split("/")[-1]
|
||||
# trim extension
|
||||
tool_file_id = tool_file_id.split(".")[0]
|
||||
else:
|
||||
tool_file_id = None
|
||||
else:
|
||||
tool_file_id = None
|
||||
else:
|
||||
tool_file_id = None
|
||||
|
||||
tool_file: ToolFile | None = (
|
||||
db.session.query(ToolFile)
|
||||
.filter(
|
||||
ToolFile.id == tool_file_id,
|
||||
tool_file: ToolFile | None = (
|
||||
session.query(ToolFile)
|
||||
.filter(
|
||||
ToolFile.id == tool_file_id,
|
||||
)
|
||||
.first()
|
||||
)
|
||||
.first()
|
||||
)
|
||||
|
||||
if not tool_file:
|
||||
return None
|
||||
@ -203,8 +216,7 @@ class ToolFileManager:
|
||||
|
||||
return blob, tool_file.mimetype
|
||||
|
||||
@staticmethod
|
||||
def get_file_generator_by_tool_file_id(tool_file_id: str):
|
||||
def get_file_generator_by_tool_file_id(self, tool_file_id: str) -> tuple[Optional[Generator], Optional[ToolFile]]:
|
||||
"""
|
||||
get file binary
|
||||
|
||||
@ -212,13 +224,14 @@ class ToolFileManager:
|
||||
|
||||
:return: the binary of the file, mime type
|
||||
"""
|
||||
tool_file: ToolFile | None = (
|
||||
db.session.query(ToolFile)
|
||||
.filter(
|
||||
ToolFile.id == tool_file_id,
|
||||
with Session(self._engine, expire_on_commit=False) as session:
|
||||
tool_file: ToolFile | None = (
|
||||
session.query(ToolFile)
|
||||
.filter(
|
||||
ToolFile.id == tool_file_id,
|
||||
)
|
||||
.first()
|
||||
)
|
||||
.first()
|
||||
)
|
||||
|
||||
if not tool_file:
|
||||
return None, None
|
||||
@ -229,6 +242,11 @@ class ToolFileManager:
|
||||
|
||||
|
||||
# init tool_file_parser
|
||||
from core.file.tool_file_parser import tool_file_manager
|
||||
from core.file.tool_file_parser import set_tool_file_manager_factory
|
||||
|
||||
tool_file_manager["manager"] = ToolFileManager
|
||||
|
||||
def _factory() -> ToolFileManager:
|
||||
return ToolFileManager()
|
||||
|
||||
|
||||
set_tool_file_manager_factory(_factory)
|
||||
|
||||
@ -4,6 +4,7 @@ from pydantic import BaseModel, Field
|
||||
|
||||
from core.rag.datasource.retrieval_service import RetrievalService
|
||||
from core.rag.entities.context_entities import DocumentContext
|
||||
from core.rag.entities.metadata_entities import MetadataCondition
|
||||
from core.rag.models.document import Document as RetrievalDocument
|
||||
from core.rag.retrieval.retrieval_methods import RetrievalMethod
|
||||
from core.tools.utils.dataset_retriever.dataset_retriever_base_tool import DatasetRetrieverBaseTool
|
||||
@ -33,6 +34,7 @@ class DatasetRetrieverTool(DatasetRetrieverBaseTool):
|
||||
args_schema: type[BaseModel] = DatasetRetrieverToolInput
|
||||
description: str = "use this to retrieve a dataset. "
|
||||
dataset_id: str
|
||||
metadata_filtering_conditions: MetadataCondition
|
||||
|
||||
@classmethod
|
||||
def from_dataset(cls, dataset: Dataset, **kwargs):
|
||||
@ -46,6 +48,7 @@ class DatasetRetrieverTool(DatasetRetrieverBaseTool):
|
||||
tenant_id=dataset.tenant_id,
|
||||
dataset_id=dataset.id,
|
||||
description=description,
|
||||
metadata_filtering_conditions=MetadataCondition(),
|
||||
**kwargs,
|
||||
)
|
||||
|
||||
@ -65,6 +68,7 @@ class DatasetRetrieverTool(DatasetRetrieverBaseTool):
|
||||
dataset_id=dataset.id,
|
||||
query=query,
|
||||
external_retrieval_parameters=dataset.retrieval_model,
|
||||
metadata_condition=self.metadata_filtering_conditions,
|
||||
)
|
||||
for external_document in external_documents:
|
||||
document = RetrievalDocument(
|
||||
|
||||
@ -31,8 +31,8 @@ class ToolFileMessageTransformer:
|
||||
# try to download image
|
||||
try:
|
||||
assert isinstance(message.message, ToolInvokeMessage.TextMessage)
|
||||
|
||||
file = ToolFileManager.create_file_by_url(
|
||||
tool_file_manager = ToolFileManager()
|
||||
file = tool_file_manager.create_file_by_url(
|
||||
user_id=user_id,
|
||||
tenant_id=tenant_id,
|
||||
file_url=message.message.text,
|
||||
@ -60,7 +60,7 @@ class ToolFileMessageTransformer:
|
||||
|
||||
mimetype = meta.get("mime_type", "application/octet-stream")
|
||||
# get filename from meta
|
||||
filename = meta.get("file_name", None)
|
||||
filename = meta.get("filename", None)
|
||||
# if message is str, encode it to bytes
|
||||
|
||||
if not isinstance(message.message, ToolInvokeMessage.BlobMessage):
|
||||
@ -68,7 +68,8 @@ class ToolFileMessageTransformer:
|
||||
|
||||
# FIXME: should do a type check here.
|
||||
assert isinstance(message.message.blob, bytes)
|
||||
file = ToolFileManager.create_file_by_raw(
|
||||
tool_file_manager = ToolFileManager()
|
||||
file = tool_file_manager.create_file_by_raw(
|
||||
user_id=user_id,
|
||||
tenant_id=tenant_id,
|
||||
conversation_id=conversation_id,
|
||||
|
||||
@ -223,8 +223,8 @@ def _extract_text_from_doc(file_content: bytes) -> str:
|
||||
"""
|
||||
from unstructured.partition.api import partition_via_api
|
||||
|
||||
if not (dify_config.UNSTRUCTURED_API_URL and dify_config.UNSTRUCTURED_API_KEY):
|
||||
raise TextExtractionError("UNSTRUCTURED_API_URL and UNSTRUCTURED_API_KEY must be set")
|
||||
if not dify_config.UNSTRUCTURED_API_URL:
|
||||
raise TextExtractionError("UNSTRUCTURED_API_URL must be set")
|
||||
|
||||
try:
|
||||
with tempfile.NamedTemporaryFile(suffix=".doc", delete=False) as temp_file:
|
||||
@ -235,7 +235,7 @@ def _extract_text_from_doc(file_content: bytes) -> str:
|
||||
file=file,
|
||||
metadata_filename=temp_file.name,
|
||||
api_url=dify_config.UNSTRUCTURED_API_URL,
|
||||
api_key=dify_config.UNSTRUCTURED_API_KEY,
|
||||
api_key=dify_config.UNSTRUCTURED_API_KEY, # type: ignore
|
||||
)
|
||||
os.unlink(temp_file.name)
|
||||
return "\n".join([getattr(element, "text", "") for element in elements])
|
||||
|
||||
@ -262,7 +262,10 @@ class Executor:
|
||||
headers[authorization.config.header] = f"Bearer {authorization.config.api_key}"
|
||||
elif self.auth.config.type == "basic":
|
||||
credentials = authorization.config.api_key
|
||||
encoded_credentials = base64.b64encode(credentials.encode("utf-8")).decode("utf-8")
|
||||
if ":" in credentials:
|
||||
encoded_credentials = base64.b64encode(credentials.encode("utf-8")).decode("utf-8")
|
||||
else:
|
||||
encoded_credentials = credentials
|
||||
headers[authorization.config.header] = f"Basic {encoded_credentials}"
|
||||
elif self.auth.config.type == "custom":
|
||||
headers[authorization.config.header] = authorization.config.api_key or ""
|
||||
|
||||
@ -191,8 +191,9 @@ class HttpRequestNode(BaseNode[HttpRequestNodeData]):
|
||||
mime_type = (
|
||||
content_disposition_type or content_type or mimetypes.guess_type(filename)[0] or "application/octet-stream"
|
||||
)
|
||||
tool_file_manager = ToolFileManager()
|
||||
|
||||
tool_file = ToolFileManager.create_file_by_raw(
|
||||
tool_file = tool_file_manager.create_file_by_raw(
|
||||
user_id=self.user_id,
|
||||
tenant_id=self.tenant_id,
|
||||
conversation_id=None,
|
||||
|
||||
@ -6,7 +6,7 @@ from collections import defaultdict
|
||||
from collections.abc import Mapping, Sequence
|
||||
from typing import Any, Optional, cast
|
||||
|
||||
from sqlalchemy import Integer, and_, func, or_, text
|
||||
from sqlalchemy import Float, and_, func, or_, text
|
||||
from sqlalchemy import cast as sqlalchemy_cast
|
||||
|
||||
from core.app.app_config.entities import DatasetRetrieveConfigEntity
|
||||
@ -32,11 +32,11 @@ from core.workflow.nodes.knowledge_retrieval.template_prompts import (
|
||||
METADATA_FILTER_COMPLETION_PROMPT,
|
||||
METADATA_FILTER_SYSTEM_PROMPT,
|
||||
METADATA_FILTER_USER_PROMPT_1,
|
||||
METADATA_FILTER_USER_PROMPT_2,
|
||||
METADATA_FILTER_USER_PROMPT_3,
|
||||
)
|
||||
from core.workflow.nodes.llm.entities import LLMNodeChatModelMessage, LLMNodeCompletionModelPromptTemplate
|
||||
from core.workflow.nodes.llm.node import LLMNode
|
||||
from core.workflow.nodes.question_classifier.template_prompts import QUESTION_CLASSIFIER_USER_PROMPT_2
|
||||
from extensions.ext_database import db
|
||||
from extensions.ext_redis import redis_client
|
||||
from libs.json_in_md_parser import parse_and_check_json_markdown
|
||||
@ -493,24 +493,24 @@ class KnowledgeRetrievalNode(LLMNode):
|
||||
if isinstance(value, str):
|
||||
filters.append(Document.doc_metadata[metadata_name] == f'"{value}"')
|
||||
else:
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Integer) == value)
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Float) == value)
|
||||
case "is not" | "≠":
|
||||
if isinstance(value, str):
|
||||
filters.append(Document.doc_metadata[metadata_name] != f'"{value}"')
|
||||
else:
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Integer) != value)
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Float) != value)
|
||||
case "empty":
|
||||
filters.append(Document.doc_metadata[metadata_name].is_(None))
|
||||
case "not empty":
|
||||
filters.append(Document.doc_metadata[metadata_name].isnot(None))
|
||||
case "before" | "<":
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Integer) < value)
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Float) < value)
|
||||
case "after" | ">":
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Integer) > value)
|
||||
case "≤" | ">=":
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Integer) <= value)
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Float) > value)
|
||||
case "≤" | "<=":
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Float) <= value)
|
||||
case "≥" | ">=":
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Integer) >= value)
|
||||
filters.append(sqlalchemy_cast(Document.doc_metadata[metadata_name].astext, Float) >= value)
|
||||
case _:
|
||||
pass
|
||||
return filters
|
||||
@ -618,7 +618,7 @@ class KnowledgeRetrievalNode(LLMNode):
|
||||
)
|
||||
prompt_messages.append(assistant_prompt_message_1)
|
||||
user_prompt_message_2 = LLMNodeChatModelMessage(
|
||||
role=PromptMessageRole.USER, text=QUESTION_CLASSIFIER_USER_PROMPT_2
|
||||
role=PromptMessageRole.USER, text=METADATA_FILTER_USER_PROMPT_2
|
||||
)
|
||||
prompt_messages.append(user_prompt_message_2)
|
||||
assistant_prompt_message_2 = LLMNodeChatModelMessage(
|
||||
|
||||
@ -2,7 +2,7 @@ METADATA_FILTER_SYSTEM_PROMPT = """
|
||||
### Job Description',
|
||||
You are a text metadata extract engine that extract text's metadata based on user input and set the metadata value
|
||||
### Task
|
||||
Your task is to ONLY extract the metadatas that exist in the input text from the provided metadata list and Use the following operators ["=", "!=", ">", "<", ">=", "<="] to express logical relationships, then return result in JSON format with the key "metadata_fields" and value "metadata_field_value" and comparison operator "comparison_operator".
|
||||
Your task is to ONLY extract the metadatas that exist in the input text from the provided metadata list and Use the following operators ["contains", "not contains", "start with", "end with", "is", "is not", "empty", "not empty", "=", "≠", ">", "<", "≥", "≤", "before", "after"] to express logical relationships, then return result in JSON format with the key "metadata_fields" and value "metadata_field_value" and comparison operator "comparison_operator".
|
||||
### Format
|
||||
The input text is in the variable input_text. Metadata are specified as a list in the variable metadata_fields.
|
||||
### Constraint
|
||||
@ -50,7 +50,7 @@ You are a text metadata extract engine that extract text's metadata based on use
|
||||
# Your task is to ONLY extract the metadatas that exist in the input text from the provided metadata list and Use the following operators ["=", "!=", ">", "<", ">=", "<="] to express logical relationships, then return result in JSON format with the key "metadata_fields" and value "metadata_field_value" and comparison operator "comparison_operator".
|
||||
### Format
|
||||
The input text is in the variable input_text. Metadata are specified as a list in the variable metadata_fields.
|
||||
### Constraint
|
||||
### Constraint
|
||||
DO NOT include anything other than the JSON array in your response.
|
||||
### Example
|
||||
Here is the chat example between human and assistant, inside <example></example> XML tags.
|
||||
@ -59,7 +59,7 @@ User:{{"input_text": ["I want to know which company’s email address test@examp
|
||||
Assistant:{{"metadata_map": [{{"metadata_field_name": "email", "metadata_field_value": "test@example.com", "comparison_operator": "="}}]}}
|
||||
User:{{"input_text": "What are the movies with a score of more than 9 in 2024?", "metadata_fields": ["name", "year", "rating", "country"]}}
|
||||
Assistant:{{"metadata_map": [{{"metadata_field_name": "year", "metadata_field_value": "2024", "comparison_operator": "="}, {{"metadata_field_name": "rating", "metadata_field_value": "9", "comparison_operator": ">"}}]}}
|
||||
</example>
|
||||
</example>
|
||||
### User Input
|
||||
{{"input_text" : "{input_text}", "metadata_fields" : {metadata_fields}}}
|
||||
### Assistant Output
|
||||
|
||||
@ -38,3 +38,8 @@ class MemoryRolePrefixRequiredError(LLMNodeError):
|
||||
class FileTypeNotSupportError(LLMNodeError):
|
||||
def __init__(self, *, type_name: str):
|
||||
super().__init__(f"{type_name} type is not supported by this model")
|
||||
|
||||
|
||||
class UnsupportedPromptContentTypeError(LLMNodeError):
|
||||
def __init__(self, *, type_name: str) -> None:
|
||||
super().__init__(f"Prompt content type {type_name} is not supported.")
|
||||
|
||||
160
api/core/workflow/nodes/llm/file_saver.py
Normal file
160
api/core/workflow/nodes/llm/file_saver.py
Normal file
@ -0,0 +1,160 @@
|
||||
import mimetypes
|
||||
import typing as tp
|
||||
|
||||
from sqlalchemy import Engine
|
||||
|
||||
from constants.mimetypes import DEFAULT_EXTENSION, DEFAULT_MIME_TYPE
|
||||
from core.file import File, FileTransferMethod, FileType
|
||||
from core.helper import ssrf_proxy
|
||||
from core.tools.signature import sign_tool_file
|
||||
from core.tools.tool_file_manager import ToolFileManager
|
||||
from models import db as global_db
|
||||
|
||||
|
||||
class LLMFileSaver(tp.Protocol):
|
||||
"""LLMFileSaver is responsible for save multimodal output returned by
|
||||
LLM.
|
||||
"""
|
||||
|
||||
def save_binary_string(
|
||||
self,
|
||||
data: bytes,
|
||||
mime_type: str,
|
||||
file_type: FileType,
|
||||
extension_override: str | None = None,
|
||||
) -> File:
|
||||
"""save_binary_string saves the inline file data returned by LLM.
|
||||
|
||||
Currently (2025-04-30), only some of Google Gemini models will return
|
||||
multimodal output as inline data.
|
||||
|
||||
:param data: the contents of the file
|
||||
:param mime_type: the media type of the file, specified by rfc6838
|
||||
(https://datatracker.ietf.org/doc/html/rfc6838)
|
||||
:param file_type: The file type of the inline file.
|
||||
:param extension_override: Override the auto-detected file extension while saving this file.
|
||||
|
||||
The default value is `None`, which means do not override the file extension and guessing it
|
||||
from the `mime_type` attribute while saving the file.
|
||||
|
||||
Setting it to values other than `None` means override the file's extension, and
|
||||
will bypass the extension guessing saving the file.
|
||||
|
||||
Specially, setting it to empty string (`""`) will leave the file extension empty.
|
||||
|
||||
When it is not `None` or empty string (`""`), it should be a string beginning with a
|
||||
dot (`.`). For example, `.py` and `.tar.gz` are both valid values, while `py`
|
||||
and `tar.gz` are not.
|
||||
"""
|
||||
pass
|
||||
|
||||
def save_remote_url(self, url: str, file_type: FileType) -> File:
|
||||
"""save_remote_url saves the file from a remote url returned by LLM.
|
||||
|
||||
Currently (2025-04-30), no model returns multimodel output as a url.
|
||||
|
||||
:param url: the url of the file.
|
||||
:param file_type: the file type of the file, check `FileType` enum for reference.
|
||||
"""
|
||||
pass
|
||||
|
||||
|
||||
EngineFactory: tp.TypeAlias = tp.Callable[[], Engine]
|
||||
|
||||
|
||||
class FileSaverImpl(LLMFileSaver):
|
||||
_engine_factory: EngineFactory
|
||||
_tenant_id: str
|
||||
_user_id: str
|
||||
|
||||
def __init__(self, user_id: str, tenant_id: str, engine_factory: EngineFactory | None = None):
|
||||
if engine_factory is None:
|
||||
|
||||
def _factory():
|
||||
return global_db.engine
|
||||
|
||||
engine_factory = _factory
|
||||
self._engine_factory = engine_factory
|
||||
self._user_id = user_id
|
||||
self._tenant_id = tenant_id
|
||||
|
||||
def _get_tool_file_manager(self):
|
||||
return ToolFileManager(engine=self._engine_factory())
|
||||
|
||||
def save_remote_url(self, url: str, file_type: FileType) -> File:
|
||||
http_response = ssrf_proxy.get(url)
|
||||
http_response.raise_for_status()
|
||||
data = http_response.content
|
||||
mime_type_from_header = http_response.headers.get("Content-Type")
|
||||
mime_type, extension = _extract_content_type_and_extension(url, mime_type_from_header)
|
||||
return self.save_binary_string(data, mime_type, file_type, extension_override=extension)
|
||||
|
||||
def save_binary_string(
|
||||
self,
|
||||
data: bytes,
|
||||
mime_type: str,
|
||||
file_type: FileType,
|
||||
extension_override: str | None = None,
|
||||
) -> File:
|
||||
tool_file_manager = self._get_tool_file_manager()
|
||||
tool_file = tool_file_manager.create_file_by_raw(
|
||||
user_id=self._user_id,
|
||||
tenant_id=self._tenant_id,
|
||||
# TODO(QuantumGhost): what is conversation id?
|
||||
conversation_id=None,
|
||||
file_binary=data,
|
||||
mimetype=mime_type,
|
||||
)
|
||||
extension_override = _validate_extension_override(extension_override)
|
||||
extension = _get_extension(mime_type, extension_override)
|
||||
url = sign_tool_file(tool_file.id, extension)
|
||||
|
||||
return File(
|
||||
tenant_id=self._tenant_id,
|
||||
type=file_type,
|
||||
transfer_method=FileTransferMethod.TOOL_FILE,
|
||||
filename=tool_file.name,
|
||||
extension=extension,
|
||||
mime_type=mime_type,
|
||||
size=len(data),
|
||||
related_id=tool_file.id,
|
||||
url=url,
|
||||
# TODO(QuantumGhost): how should I set the following key?
|
||||
# What's the difference between `remote_url` and `url`?
|
||||
# What's the purpose of `storage_key` and `dify_model_identity`?
|
||||
storage_key=tool_file.file_key,
|
||||
)
|
||||
|
||||
|
||||
def _get_extension(mime_type: str, extension_override: str | None = None) -> str:
|
||||
"""get_extension return the extension of file.
|
||||
|
||||
If the `extension_override` parameter is set, this function should honor it and
|
||||
return its value.
|
||||
"""
|
||||
if extension_override is not None:
|
||||
return extension_override
|
||||
return mimetypes.guess_extension(mime_type) or DEFAULT_EXTENSION
|
||||
|
||||
|
||||
def _extract_content_type_and_extension(url: str, content_type_header: str | None) -> tuple[str, str]:
|
||||
"""_extract_content_type_and_extension tries to
|
||||
guess content type of file from url and `Content-Type` header in response.
|
||||
"""
|
||||
if content_type_header:
|
||||
extension = mimetypes.guess_extension(content_type_header) or DEFAULT_EXTENSION
|
||||
return content_type_header, extension
|
||||
content_type = mimetypes.guess_type(url)[0] or DEFAULT_MIME_TYPE
|
||||
extension = mimetypes.guess_extension(content_type) or DEFAULT_EXTENSION
|
||||
return content_type, extension
|
||||
|
||||
|
||||
def _validate_extension_override(extension_override: str | None) -> str | None:
|
||||
# `extension_override` is allow to be `None or `""`.
|
||||
if extension_override is None:
|
||||
return None
|
||||
if extension_override == "":
|
||||
return ""
|
||||
if not extension_override.startswith("."):
|
||||
raise ValueError("extension_override should start with '.' if not None or empty.", extension_override)
|
||||
return extension_override
|
||||
@ -1,3 +1,5 @@
|
||||
import base64
|
||||
import io
|
||||
import json
|
||||
import logging
|
||||
from collections.abc import Generator, Mapping, Sequence
|
||||
@ -21,7 +23,7 @@ from core.model_runtime.entities import (
|
||||
PromptMessageContentType,
|
||||
TextPromptMessageContent,
|
||||
)
|
||||
from core.model_runtime.entities.llm_entities import LLMResult, LLMUsage
|
||||
from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunk, LLMUsage
|
||||
from core.model_runtime.entities.message_entities import (
|
||||
AssistantPromptMessage,
|
||||
PromptMessageContentUnionTypes,
|
||||
@ -38,7 +40,6 @@ from core.model_runtime.entities.model_entities import (
|
||||
)
|
||||
from core.model_runtime.model_providers.__base.large_language_model import LargeLanguageModel
|
||||
from core.model_runtime.utils.encoders import jsonable_encoder
|
||||
from core.model_runtime.utils.helper import convert_llm_result_chunk_to_str
|
||||
from core.plugin.entities.plugin import ModelProviderID
|
||||
from core.prompt.entities.advanced_prompt_entities import CompletionModelPromptTemplate, MemoryConfig
|
||||
from core.prompt.utils.prompt_message_util import PromptMessageUtil
|
||||
@ -95,9 +96,13 @@ from .exc import (
|
||||
TemplateTypeNotSupportError,
|
||||
VariableNotFoundError,
|
||||
)
|
||||
from .file_saver import FileSaverImpl, LLMFileSaver
|
||||
|
||||
if TYPE_CHECKING:
|
||||
from core.file.models import File
|
||||
from core.workflow.graph_engine.entities.graph import Graph
|
||||
from core.workflow.graph_engine.entities.graph_init_params import GraphInitParams
|
||||
from core.workflow.graph_engine.entities.graph_runtime_state import GraphRuntimeState
|
||||
|
||||
logger = logging.getLogger(__name__)
|
||||
|
||||
@ -106,6 +111,43 @@ class LLMNode(BaseNode[LLMNodeData]):
|
||||
_node_data_cls = LLMNodeData
|
||||
_node_type = NodeType.LLM
|
||||
|
||||
# Instance attributes specific to LLMNode.
|
||||
# Output variable for file
|
||||
_file_outputs: list["File"]
|
||||
|
||||
_llm_file_saver: LLMFileSaver
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
id: str,
|
||||
config: Mapping[str, Any],
|
||||
graph_init_params: "GraphInitParams",
|
||||
graph: "Graph",
|
||||
graph_runtime_state: "GraphRuntimeState",
|
||||
previous_node_id: Optional[str] = None,
|
||||
thread_pool_id: Optional[str] = None,
|
||||
*,
|
||||
llm_file_saver: LLMFileSaver | None = None,
|
||||
) -> None:
|
||||
super().__init__(
|
||||
id=id,
|
||||
config=config,
|
||||
graph_init_params=graph_init_params,
|
||||
graph=graph,
|
||||
graph_runtime_state=graph_runtime_state,
|
||||
previous_node_id=previous_node_id,
|
||||
thread_pool_id=thread_pool_id,
|
||||
)
|
||||
# LLM file outputs, used for MultiModal outputs.
|
||||
self._file_outputs: list[File] = []
|
||||
|
||||
if llm_file_saver is None:
|
||||
llm_file_saver = FileSaverImpl(
|
||||
user_id=graph_init_params.user_id,
|
||||
tenant_id=graph_init_params.tenant_id,
|
||||
)
|
||||
self._llm_file_saver = llm_file_saver
|
||||
|
||||
def _run(self) -> Generator[NodeEvent | InNodeEvent, None, None]:
|
||||
def process_structured_output(text: str) -> Optional[dict[str, Any] | list[Any]]:
|
||||
"""Process structured output if enabled"""
|
||||
@ -215,6 +257,9 @@ class LLMNode(BaseNode[LLMNodeData]):
|
||||
structured_output = process_structured_output(result_text)
|
||||
if structured_output:
|
||||
outputs["structured_output"] = structured_output
|
||||
if self._file_outputs is not None:
|
||||
outputs["files"] = self._file_outputs
|
||||
|
||||
yield RunCompletedEvent(
|
||||
run_result=NodeRunResult(
|
||||
status=WorkflowNodeExecutionStatus.SUCCEEDED,
|
||||
@ -240,6 +285,7 @@ class LLMNode(BaseNode[LLMNodeData]):
|
||||
)
|
||||
)
|
||||
except Exception as e:
|
||||
logger.exception("error while executing llm node")
|
||||
yield RunCompletedEvent(
|
||||
run_result=NodeRunResult(
|
||||
status=WorkflowNodeExecutionStatus.FAILED,
|
||||
@ -268,44 +314,45 @@ class LLMNode(BaseNode[LLMNodeData]):
|
||||
|
||||
return self._handle_invoke_result(invoke_result=invoke_result)
|
||||
|
||||
def _handle_invoke_result(self, invoke_result: LLMResult | Generator) -> Generator[NodeEvent, None, None]:
|
||||
def _handle_invoke_result(
|
||||
self, invoke_result: LLMResult | Generator[LLMResultChunk, None, None]
|
||||
) -> Generator[NodeEvent, None, None]:
|
||||
# For blocking mode
|
||||
if isinstance(invoke_result, LLMResult):
|
||||
message_text = convert_llm_result_chunk_to_str(invoke_result.message.content)
|
||||
|
||||
yield ModelInvokeCompletedEvent(
|
||||
text=message_text,
|
||||
usage=invoke_result.usage,
|
||||
finish_reason=None,
|
||||
)
|
||||
event = self._handle_blocking_result(invoke_result=invoke_result)
|
||||
yield event
|
||||
return
|
||||
|
||||
model = None
|
||||
# For streaming mode
|
||||
model = ""
|
||||
prompt_messages: list[PromptMessage] = []
|
||||
full_text = ""
|
||||
usage = None
|
||||
|
||||
usage = LLMUsage.empty_usage()
|
||||
finish_reason = None
|
||||
full_text_buffer = io.StringIO()
|
||||
for result in invoke_result:
|
||||
text = convert_llm_result_chunk_to_str(result.delta.message.content)
|
||||
full_text += text
|
||||
contents = result.delta.message.content
|
||||
for text_part in self._save_multimodal_output_and_convert_result_to_markdown(contents):
|
||||
full_text_buffer.write(text_part)
|
||||
yield RunStreamChunkEvent(chunk_content=text_part, from_variable_selector=[self.node_id, "text"])
|
||||
|
||||
yield RunStreamChunkEvent(chunk_content=text, from_variable_selector=[self.node_id, "text"])
|
||||
|
||||
if not model:
|
||||
# Update the whole metadata
|
||||
if not model and result.model:
|
||||
model = result.model
|
||||
|
||||
if not prompt_messages:
|
||||
prompt_messages = result.prompt_messages
|
||||
|
||||
if not usage and result.delta.usage:
|
||||
if len(prompt_messages) == 0:
|
||||
# TODO(QuantumGhost): it seems that this update has no visable effect.
|
||||
# What's the purpose of the line below?
|
||||
prompt_messages = list(result.prompt_messages)
|
||||
if usage.prompt_tokens == 0 and result.delta.usage:
|
||||
usage = result.delta.usage
|
||||
|
||||
if not finish_reason and result.delta.finish_reason:
|
||||
if finish_reason is None and result.delta.finish_reason:
|
||||
finish_reason = result.delta.finish_reason
|
||||
|
||||
if not usage:
|
||||
usage = LLMUsage.empty_usage()
|
||||
yield ModelInvokeCompletedEvent(text=full_text_buffer.getvalue(), usage=usage, finish_reason=finish_reason)
|
||||
|
||||
yield ModelInvokeCompletedEvent(text=full_text, usage=usage, finish_reason=finish_reason)
|
||||
def _image_file_to_markdown(self, file: "File", /):
|
||||
text_chunk = f"})"
|
||||
return text_chunk
|
||||
|
||||
def _transform_chat_messages(
|
||||
self, messages: Sequence[LLMNodeChatModelMessage] | LLMNodeCompletionModelPromptTemplate, /
|
||||
@ -963,6 +1010,42 @@ class LLMNode(BaseNode[LLMNodeData]):
|
||||
|
||||
return prompt_messages
|
||||
|
||||
def _handle_blocking_result(self, *, invoke_result: LLMResult) -> ModelInvokeCompletedEvent:
|
||||
buffer = io.StringIO()
|
||||
for text_part in self._save_multimodal_output_and_convert_result_to_markdown(invoke_result.message.content):
|
||||
buffer.write(text_part)
|
||||
|
||||
return ModelInvokeCompletedEvent(
|
||||
text=buffer.getvalue(),
|
||||
usage=invoke_result.usage,
|
||||
finish_reason=None,
|
||||
)
|
||||
|
||||
def _save_multimodal_image_output(self, content: ImagePromptMessageContent) -> "File":
|
||||
"""_save_multimodal_output saves multi-modal contents generated by LLM plugins.
|
||||
|
||||
There are two kinds of multimodal outputs:
|
||||
|
||||
- Inlined data encoded in base64, which would be saved to storage directly.
|
||||
- Remote files referenced by an url, which would be downloaded and then saved to storage.
|
||||
|
||||
Currently, only image files are supported.
|
||||
"""
|
||||
# Inject the saver somehow...
|
||||
_saver = self._llm_file_saver
|
||||
|
||||
# If this
|
||||
if content.url != "":
|
||||
saved_file = _saver.save_remote_url(content.url, FileType.IMAGE)
|
||||
else:
|
||||
saved_file = _saver.save_binary_string(
|
||||
data=base64.b64decode(content.base64_data),
|
||||
mime_type=content.mime_type,
|
||||
file_type=FileType.IMAGE,
|
||||
)
|
||||
self._file_outputs.append(saved_file)
|
||||
return saved_file
|
||||
|
||||
def _handle_native_json_schema(self, model_parameters: dict, rules: list[ParameterRule]) -> dict:
|
||||
"""
|
||||
Handle structured output for models with native JSON schema support.
|
||||
@ -1123,6 +1206,41 @@ class LLMNode(BaseNode[LLMNodeData]):
|
||||
else SupportStructuredOutputStatus.UNSUPPORTED
|
||||
)
|
||||
|
||||
def _save_multimodal_output_and_convert_result_to_markdown(
|
||||
self,
|
||||
contents: str | list[PromptMessageContentUnionTypes] | None,
|
||||
) -> Generator[str, None, None]:
|
||||
"""Convert intermediate prompt messages into strings and yield them to the caller.
|
||||
|
||||
If the messages contain non-textual content (e.g., multimedia like images or videos),
|
||||
it will be saved separately, and the corresponding Markdown representation will
|
||||
be yielded to the caller.
|
||||
"""
|
||||
|
||||
# NOTE(QuantumGhost): This function should yield results to the caller immediately
|
||||
# whenever new content or partial content is available. Avoid any intermediate buffering
|
||||
# of results. Additionally, do not yield empty strings; instead, yield from an empty list
|
||||
# if necessary.
|
||||
if contents is None:
|
||||
yield from []
|
||||
return
|
||||
if isinstance(contents, str):
|
||||
yield contents
|
||||
elif isinstance(contents, list):
|
||||
for item in contents:
|
||||
if isinstance(item, TextPromptMessageContent):
|
||||
yield item.data
|
||||
elif isinstance(item, ImagePromptMessageContent):
|
||||
file = self._save_multimodal_image_output(item)
|
||||
self._file_outputs.append(file)
|
||||
yield self._image_file_to_markdown(file)
|
||||
else:
|
||||
logger.warning("unknown item type encountered, type=%s", type(item))
|
||||
yield str(item)
|
||||
else:
|
||||
logger.warning("unknown contents type encountered, type=%s", type(contents))
|
||||
yield str(contents)
|
||||
|
||||
|
||||
def _combine_message_content_with_role(
|
||||
*, contents: Optional[str | list[PromptMessageContentUnionTypes]] = None, role: PromptMessageRole
|
||||
|
||||
@ -17,7 +17,7 @@ Some additional information is provided below. Always adhere to these instructio
|
||||
</instruction>
|
||||
Steps:
|
||||
1. Review the chat history provided within the <histories> tags.
|
||||
2. Extract the relevant information based on the criteria given, output multiple values if there is multiple relevant information that match the criteria in the given text.
|
||||
2. Extract the relevant information based on the criteria given, output multiple values if there is multiple relevant information that match the criteria in the given text.
|
||||
3. Generate a well-formatted output using the defined functions and arguments.
|
||||
4. Use the `extract_parameter` function to create structured outputs with appropriate parameters.
|
||||
5. Do not include any XML tags in your output.
|
||||
@ -89,13 +89,13 @@ Some extra information are provided below, I should always follow the instructio
|
||||
</instructions>
|
||||
|
||||
### Extract parameter Workflow
|
||||
I need to extract the following information from the input text. The <information to be extracted> tag specifies the 'type', 'description' and 'required' of the information to be extracted.
|
||||
I need to extract the following information from the input text. The <information to be extracted> tag specifies the 'type', 'description' and 'required' of the information to be extracted.
|
||||
<information to be extracted>
|
||||
{{ structure }}
|
||||
</information to be extracted>
|
||||
|
||||
Step 1: Carefully read the input and understand the structure of the expected output.
|
||||
Step 2: Extract relevant parameters from the provided text based on the name and description of object.
|
||||
Step 2: Extract relevant parameters from the provided text based on the name and description of object.
|
||||
Step 3: Structure the extracted parameters to JSON object as specified in <structure>.
|
||||
Step 4: Ensure that the JSON object is properly formatted and valid. The output should not contain any XML tags. Only the JSON object should be outputted.
|
||||
|
||||
@ -106,10 +106,10 @@ Here are the chat histories between human and assistant, inside <histories></his
|
||||
</histories>
|
||||
|
||||
### Structure
|
||||
Here is the structure of the expected output, I should always follow the output structure.
|
||||
Here is the structure of the expected output, I should always follow the output structure.
|
||||
{{γγγ
|
||||
'properties1': 'relevant text extracted from input',
|
||||
'properties2': 'relevant text extracted from input',
|
||||
'properties1': 'relevant text extracted from input',
|
||||
'properties2': 'relevant text extracted from input',
|
||||
}}γγγ
|
||||
|
||||
### Input Text
|
||||
@ -119,7 +119,7 @@ Inside <text></text> XML tags, there is a text that I should extract parameters
|
||||
</text>
|
||||
|
||||
### Answer
|
||||
I should always output a valid JSON object. Output nothing other than the JSON object.
|
||||
I should always output a valid JSON object. Output nothing other than the JSON object.
|
||||
```JSON
|
||||
""" # noqa: E501
|
||||
|
||||
|
||||
@ -55,7 +55,7 @@ You are a text classification engine that analyzes text data and assigns categor
|
||||
Your task is to assign one categories ONLY to the input text and only one category may be assigned returned in the output. Additionally, you need to extract the key words from the text that are related to the classification.
|
||||
### Format
|
||||
The input text is in the variable input_text. Categories are specified as a category list with two filed category_id and category_name in the variable categories. Classification instructions may be included to improve the classification accuracy.
|
||||
### Constraint
|
||||
### Constraint
|
||||
DO NOT include anything other than the JSON array in your response.
|
||||
### Example
|
||||
Here is the chat example between human and assistant, inside <example></example> XML tags.
|
||||
@ -64,7 +64,7 @@ User:{{"input_text": ["I recently had a great experience with your company. The
|
||||
Assistant:{{"keywords": ["recently", "great experience", "company", "service", "prompt", "staff", "friendly"],"category_id": "f5660049-284f-41a7-b301-fd24176a711c","category_name": "Customer Service"}}
|
||||
User:{{"input_text": ["bad service, slow to bring the food"], "categories": [{{"category_id":"80fb86a0-4454-4bf5-924c-f253fdd83c02","category_name":"Food Quality"}},{{"category_id":"f6ff5bc3-aca0-4e4a-8627-e760d0aca78f","category_name":"Experience"}},{{"category_id":"cc771f63-74e7-4c61-882e-3eda9d8ba5d7","category_name":"Price"}}], "classification_instructions": []}}
|
||||
Assistant:{{"keywords": ["bad service", "slow", "food", "tip", "terrible", "waitresses"],"category_id": "f6ff5bc3-aca0-4e4a-8627-e760d0aca78f","category_name": "Experience"}}
|
||||
</example>
|
||||
</example>
|
||||
### Memory
|
||||
Here are the chat histories between human and assistant, inside <histories></histories> XML tags.
|
||||
<histories>
|
||||
|
||||
@ -11,6 +11,8 @@ class Operation(StrEnum):
|
||||
SUBTRACT = "-="
|
||||
MULTIPLY = "*="
|
||||
DIVIDE = "/="
|
||||
REMOVE_FIRST = "remove-first"
|
||||
REMOVE_LAST = "remove-last"
|
||||
|
||||
|
||||
class InputType(StrEnum):
|
||||
|
||||
@ -23,6 +23,15 @@ def is_operation_supported(*, variable_type: SegmentType, operation: Operation):
|
||||
SegmentType.ARRAY_NUMBER,
|
||||
SegmentType.ARRAY_FILE,
|
||||
}
|
||||
case Operation.REMOVE_FIRST | Operation.REMOVE_LAST:
|
||||
# Only array variable can have elements removed
|
||||
return variable_type in {
|
||||
SegmentType.ARRAY_ANY,
|
||||
SegmentType.ARRAY_OBJECT,
|
||||
SegmentType.ARRAY_STRING,
|
||||
SegmentType.ARRAY_NUMBER,
|
||||
SegmentType.ARRAY_FILE,
|
||||
}
|
||||
case _:
|
||||
return False
|
||||
|
||||
@ -51,7 +60,7 @@ def is_constant_input_supported(*, variable_type: SegmentType, operation: Operat
|
||||
|
||||
|
||||
def is_input_value_valid(*, variable_type: SegmentType, operation: Operation, value: Any):
|
||||
if operation == Operation.CLEAR:
|
||||
if operation in {Operation.CLEAR, Operation.REMOVE_FIRST, Operation.REMOVE_LAST}:
|
||||
return True
|
||||
match variable_type:
|
||||
case SegmentType.STRING:
|
||||
|
||||
@ -64,7 +64,7 @@ class VariableAssignerNode(BaseNode[VariableAssignerNodeData]):
|
||||
# Get value from variable pool
|
||||
if (
|
||||
item.input_type == InputType.VARIABLE
|
||||
and item.operation != Operation.CLEAR
|
||||
and item.operation not in {Operation.CLEAR, Operation.REMOVE_FIRST, Operation.REMOVE_LAST}
|
||||
and item.value is not None
|
||||
):
|
||||
value = self.graph_runtime_state.variable_pool.get(item.value)
|
||||
@ -165,5 +165,15 @@ class VariableAssignerNode(BaseNode[VariableAssignerNodeData]):
|
||||
return variable.value * value
|
||||
case Operation.DIVIDE:
|
||||
return variable.value / value
|
||||
case Operation.REMOVE_FIRST:
|
||||
# If array is empty, do nothing
|
||||
if not variable.value:
|
||||
return variable.value
|
||||
return variable.value[1:]
|
||||
case Operation.REMOVE_LAST:
|
||||
# If array is empty, do nothing
|
||||
if not variable.value:
|
||||
return variable.value
|
||||
return variable.value[:-1]
|
||||
case _:
|
||||
raise OperationNotSupportedError(operation=operation, variable_type=variable.value_type)
|
||||
|
||||
@ -9,6 +9,7 @@ from core.app.apps.base_app_queue_manager import GenerateTaskStoppedError
|
||||
from core.app.entities.app_invoke_entities import InvokeFrom
|
||||
from core.file.models import File
|
||||
from core.workflow.callbacks import WorkflowCallback
|
||||
from core.workflow.constants import ENVIRONMENT_VARIABLE_NODE_ID
|
||||
from core.workflow.entities.variable_pool import VariablePool
|
||||
from core.workflow.errors import WorkflowNodeRunFailedError
|
||||
from core.workflow.graph_engine.entities.event import GraphEngineEvent, GraphRunFailedEvent, InNodeEvent
|
||||
@ -364,4 +365,5 @@ class WorkflowEntry:
|
||||
input_value = file_factory.build_from_mappings(mappings=input_value, tenant_id=tenant_id)
|
||||
|
||||
# append variable and value to variable pool
|
||||
variable_pool.add([variable_node_id] + variable_key_list, input_value)
|
||||
if variable_node_id != ENVIRONMENT_VARIABLE_NODE_ID:
|
||||
variable_pool.add([variable_node_id] + variable_key_list, input_value)
|
||||
|
||||
@ -21,14 +21,14 @@ def upgrade():
|
||||
# ### commands auto generated by Alembic - please adjust! ###
|
||||
# Get the database connection
|
||||
conn = op.get_bind()
|
||||
|
||||
|
||||
# Use SQLAlchemy inspector to get the columns of the 'tool_files' table
|
||||
inspector = sa.inspect(conn)
|
||||
columns = [col['name'] for col in inspector.get_columns('tool_files')]
|
||||
|
||||
# If 'name' or 'size' columns already exist, exit the upgrade function
|
||||
if 'name' in columns or 'size' in columns:
|
||||
return
|
||||
return
|
||||
|
||||
with op.batch_alter_table('tool_files', schema=None) as batch_op:
|
||||
batch_op.add_column(sa.Column('name', sa.String(), nullable=True))
|
||||
|
||||
@ -35,4 +35,4 @@ def downgrade():
|
||||
# batch_op.drop_column('retry_index')
|
||||
pass
|
||||
|
||||
# ### end Alembic commands ###
|
||||
# ### end Alembic commands ###
|
||||
|
||||
@ -23,7 +23,7 @@ def upgrade():
|
||||
conn = op.get_bind()
|
||||
inspector = inspect(conn)
|
||||
has_column = 'retry_index' in [col['name'] for col in inspector.get_columns('workflow_node_executions')]
|
||||
|
||||
|
||||
if has_column:
|
||||
with op.batch_alter_table('workflow_node_executions', schema=None) as batch_op:
|
||||
batch_op.drop_column('retry_index')
|
||||
|
||||
@ -1,7 +1,7 @@
|
||||
"""init
|
||||
|
||||
Revision ID: 64b051264f32
|
||||
Revises:
|
||||
Revises:
|
||||
Create Date: 2023-05-13 14:26:59.085018
|
||||
|
||||
"""
|
||||
|
||||
@ -99,12 +99,12 @@ def upgrade():
|
||||
id=id,
|
||||
tenant_id=tenant_id,
|
||||
user_id=user_id,
|
||||
provider='google',
|
||||
provider='google',
|
||||
encrypted_credentials=encrypted_credentials,
|
||||
created_at=created_at,
|
||||
updated_at=updated_at
|
||||
)
|
||||
|
||||
|
||||
# ### end Alembic commands ###
|
||||
|
||||
|
||||
|
||||
@ -10,4 +10,16 @@ POSTGRES_INDEXES_NAMING_CONVENTION = {
|
||||
}
|
||||
|
||||
metadata = MetaData(naming_convention=POSTGRES_INDEXES_NAMING_CONVENTION)
|
||||
|
||||
# ****** IMPORTANT NOTICE ******
|
||||
#
|
||||
# NOTE(QuantumGhost): Avoid directly importing and using `db` in modules outside of the
|
||||
# `controllers` package.
|
||||
#
|
||||
# Instead, import `db` within the `controllers` package and pass it as an argument to
|
||||
# functions or class constructors.
|
||||
#
|
||||
# Directly importing `db` in other modules can make the code more difficult to read, test, and maintain.
|
||||
#
|
||||
# Whenever possible, avoid this pattern in new code.
|
||||
db = SQLAlchemy(metadata=metadata)
|
||||
|
||||
@ -8,6 +8,7 @@ from typing import TYPE_CHECKING, Any, Literal, Optional, cast
|
||||
|
||||
from core.plugin.entities.plugin import GenericProviderID
|
||||
from core.tools.entities.tool_entities import ToolProviderType
|
||||
from core.tools.signature import sign_tool_file
|
||||
from services.plugin.plugin_service import PluginService
|
||||
|
||||
if TYPE_CHECKING:
|
||||
@ -23,7 +24,6 @@ from configs import dify_config
|
||||
from constants import DEFAULT_FILE_NUMBER_LIMITS
|
||||
from core.file import FILE_MODEL_IDENTITY, File, FileTransferMethod, FileType
|
||||
from core.file import helpers as file_helpers
|
||||
from core.file.tool_file_parser import ToolFileParser
|
||||
from libs.helper import generate_string
|
||||
from models.base import Base
|
||||
from models.enums import CreatedByRole
|
||||
@ -986,9 +986,7 @@ class Message(db.Model): # type: ignore[name-defined]
|
||||
if not tool_file_id:
|
||||
continue
|
||||
|
||||
sign_url = ToolFileParser.get_tool_file_manager().sign_file(
|
||||
tool_file_id=tool_file_id, extension=extension
|
||||
)
|
||||
sign_url = sign_tool_file(tool_file_id=tool_file_id, extension=extension)
|
||||
elif "file-preview" in url:
|
||||
# get upload file id
|
||||
upload_file_id_pattern = r"\/files\/([\w-]+)\/file-preview?\?timestamp="
|
||||
|
||||
@ -263,8 +263,8 @@ class ToolConversationVariables(Base):
|
||||
|
||||
|
||||
class ToolFile(Base):
|
||||
"""
|
||||
store the file created by agent
|
||||
"""This table stores file metadata generated in workflows,
|
||||
not only files created by agent.
|
||||
"""
|
||||
|
||||
__tablename__ = "tool_files"
|
||||
|
||||
Some files were not shown because too many files have changed in this diff Show More
Loading…
Reference in New Issue
Block a user