opensource
/
dify
mirror of https://github.com/langgenius/dify.git
2
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Fix: rename module flags to avoid pyright constant redefinition errors (_langdetect_available, _persian_chars_re); update tests

This commit is contained in:
nourzakhama2003 2025-12-22 15:12:52 +01:00
parent 3c9907ea81
commit 8dbae53c78
2 changed files with 9 additions and 25 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
api/core/llm_generator/llm_generator.py
View File

@ -44,19 +44,19 @@ _PERSIAN_HEURISTIC = re.compile(
)
# Precompiled regex for Persian-specific characters (including Persian ye U+06CC)
_PERSIAN_CHARS_RE = re.compile(r"[پچژگک\u06CC]")
_persian_chars_re = re.compile(r"[پچژگک\u06CC]")
# Optional langdetect import — import once at module import time to avoid repeated lookups
_LANGDETECT_AVAILABLE = False
_langdetect_available = False
try:
from langdetect import DetectorFactory, detect # type: ignore
DetectorFactory.seed = 0
_LANGDETECT_AVAILABLE = True
_langdetect_available = True
except Exception:
detect = None
DetectorFactory = None
_LANGDETECT_AVAILABLE = False
_langdetect_available = False
def _contains_persian(text: str) -> bool:
@ -68,7 +68,7 @@ def _contains_persian(text: str) -> bool:
text = text or ""
# 1) Quick check: Persian-specific letters
if _PERSIAN_CHARS_RE.search(text):
if _persian_chars_re.search(text):
return True
# 2) Heuristic check for common Persian words (fast, precompiled)
@ -76,7 +76,7 @@ def _contains_persian(text: str) -> bool:
return True
# 3) Fallback: language detection (more expensive) — only run if langdetect is available
if _LANGDETECT_AVAILABLE and detect is not None:
if _langdetect_available and detect is not None:
try:
return detect(text) == "fa"
except Exception as exc:
@ -86,22 +86,6 @@ def _contains_persian(text: str) -> bool:
return False
# Precompiled regex for Persian-specific characters (including Persian ye U+06CC)
_PERSIAN_CHARS_RE = re.compile(r"[پچژگک\u06CC]")
# Optional langdetect import — import once at module import time to avoid repeated lookups
_LANGDETECT_AVAILABLE = False
try:
from langdetect import DetectorFactory, detect # type: ignore
DetectorFactory.seed = 0
_LANGDETECT_AVAILABLE = True
except Exception:
detect = None
DetectorFactory = None
_LANGDETECT_AVAILABLE = False
class WorkflowServiceInterface(Protocol):
def get_draft_workflow(self, app_model: App, workflow_id: str | None = None) -> Workflow | None:
pass

6
tests/unit_tests/core/llm_generator/test_llm_generator_persian.py
View File

@ -233,7 +233,7 @@ def test_generate_conversation_name_persian(monkeypatch):
def test_contains_persian_character_and_heuristics(monkeypatch):
from core.llm_generator.llm_generator import _contains_persian, _PERSIAN_CHARS_RE, _PERSIAN_HEURISTIC
from core.llm_generator.llm_generator import _contains_persian, _persian_chars_re, _PERSIAN_HEURISTIC
# By single Persian-specific character
assert _contains_persian("این یک تست پ") is True
@ -246,11 +246,11 @@ def test_contains_persian_langdetect_fallback(monkeypatch):
import core.llm_generator.llm_generator as lg
# Simulate langdetect being available and detecting Persian
monkeypatch.setattr(lg, "_LANGDETECT_AVAILABLE", True)
monkeypatch.setattr(lg, "_langdetect_available", True)
monkeypatch.setattr(lg, "detect", lambda text: "fa")
assert lg._contains_persian("short ambiguous text") is True
# Reset monkeypatch
monkeypatch.setattr(lg, "_LANGDETECT_AVAILABLE", False)
monkeypatch.setattr(lg, "_langdetect_available", False)
monkeypatch.setattr(lg, "detect", None)