From 53f54fe05d0e0cdfd8942d0b02a55ad710503629 Mon Sep 17 00:00:00 2001
From: matevip <matevip@mate.vip>
Date: Sat, 25 Apr 2026 10:32:23 +0800
Subject: [PATCH] build(docker): pass MAVEN_FLAGS build-arg to support
 aliyun-first profile in CN builds

---
 .env.example               |  5 +++++
 docker-compose.yml         |  2 ++
 mateclaw-server/Dockerfile | 25 ++++++++++++++++++++++---
 3 files changed, 29 insertions(+), 3 deletions(-)

diff --git a/.env.example b/.env.example
index 9561c692..dbe2a616 100644
--- a/.env.example
+++ b/.env.example
@@ -49,3 +49,8 @@ SEARXNG_SECRET=
 MATECLAW_BROWSER_CDP_URL=
 MATECLAW_BROWSER_CHROME_PATH=
 MATECLAW_BROWSER_CHANNEL=
+
+# ── Maven 镜像（国内加速）─────────────────────────────────────────
+# 在中国大陆构建时取消注释，将 Aliyun 仓库优先级提前，大幅提速 mvn 拉包。
+# 空值（默认）使用 US Maven Central → Google CDN → Aliyun 的顺序。
+#MAVEN_FLAGS=-Paliyun-first
diff --git a/docker-compose.yml b/docker-compose.yml
index b9ce2daf..05fea015 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -61,6 +61,8 @@ services:
     build:
       context: .
       dockerfile: mateclaw-server/Dockerfile
+      args:
+        MAVEN_FLAGS: ${MAVEN_FLAGS:-}
     container_name: mateclaw-server
     restart: unless-stopped
     depends_on:
diff --git a/mateclaw-server/Dockerfile b/mateclaw-server/Dockerfile
index e3220acb..f927d68b 100644
--- a/mateclaw-server/Dockerfile
+++ b/mateclaw-server/Dockerfile
@@ -21,6 +21,11 @@ RUN pnpm exec vite build --outDir /static --emptyOutDir
 # Stage 2 — Backend (Maven)
 FROM maven:3.9-eclipse-temurin-21 AS builder
 
+# Optional Maven extra flags passed at build time.
+# Set MAVEN_FLAGS=-Paliyun-first in .env (or via --build-arg) to put Aliyun
+# repos first — speeds up builds dramatically inside mainland China.
+ARG MAVEN_FLAGS=""
+
 # Inject mirror settings to avoid Maven Central timeouts in restricted networks
 COPY mateclaw-server/settings.xml /root/.m2/settings.xml
 
@@ -28,18 +33,18 @@ COPY mateclaw-server/settings.xml /root/.m2/settings.xml
 WORKDIR /plugin-api
 COPY mateclaw-plugin-api/pom.xml ./pom.xml
 COPY mateclaw-plugin-api/src ./src
-RUN mvn install -DskipTests -q
+RUN mvn install -DskipTests -q ${MAVEN_FLAGS}
 
 # Pre-fetch mateclaw-server dependencies (uses mirror, so this won't hang)
 WORKDIR /build
 COPY mateclaw-server/pom.xml .
-RUN mvn dependency:go-offline -q
+RUN mvn dependency:go-offline -q ${MAVEN_FLAGS}
 
 # Copy backend source and inject pre-built frontend into the right classpath location
 COPY mateclaw-server/src ./src
 COPY --from=frontend-builder /static ./src/main/resources/static
 
-RUN mvn package -DskipTests -q
+RUN mvn package -DskipTests -q ${MAVEN_FLAGS}
 
 # Stage 3 — Runtime
 #
@@ -58,12 +63,26 @@ WORKDIR /app
 # JDK 21 is NOT part of the base image (it ships Node for the JS driver).
 # Install openjdk-21 explicitly and add CJK fonts so Chinese pages render
 # correctly in screenshots and snapshots.
+#
+# PDF extraction toolchain — DocumentExtractTool tries pdftotext first, then
+# Python pdfplumber/pypdf, then falls through to a naive Java parser that
+# reads bytes as ISO_8859_1 (mojibake for CJK). Without poppler-utils and
+# python3, Docker always hits the naive path and feeds garbled text to the
+# Wiki pipeline. Installing these puts the Docker runtime on the same
+# extraction path as local Mac (brew poppler + pyenv pdfplumber).
+# Tesseract + chi-sim is the final OCR fallback for scanned PDFs.
 RUN apt-get update \
     && apt-get install -y --no-install-recommends \
         openjdk-21-jre-headless \
         fonts-noto-cjk \
         fonts-noto-color-emoji \
+        poppler-utils \
+        python3 \
+        python3-pip \
+        tesseract-ocr \
+        tesseract-ocr-chi-sim \
         tzdata \
+    && pip3 install --no-cache-dir --break-system-packages pdfplumber pypdf \
     && rm -rf /var/lib/apt/lists/*
 
 # Tell Playwright Java where Microsoft's image stored the browsers.