Refactor Docker build process for Python bindings and enhance version extraction

Author: tae898

bindings/python/Dockerfile.build

@@ -1,53 +1,32 @@
 # Multi-stage Docker build for ArcadeDB Python bindings
 # Supports building all three distributions: headless, minimal, full
+# 
+# REQUIRED BUILD ARGS (passed from build-all.sh):
+#   DISTRIBUTION - one of: headless, minimal, full
+#   ARCADEDB_TAG - version tag from pom.xml (e.g., 25.10.1-SNAPSHOT)
 
-ARG DISTRIBUTION=headless
-
-# Stage 1: Build Java components
-FROM maven:3.9-eclipse-temurin-21 AS java-builder
-
-WORKDIR /arcadedb
-
-# Copy git for buildnumber-maven-plugin (needs git revision info)
-COPY .git .git/
-
-# Copy all necessary files for Java build (excluding Python docs/build artifacts)
-COPY pom.xml .
-COPY README.md .
-COPY LICENSE .
-COPY console console/
-COPY coverage coverage/
-COPY e2e e2e/
-COPY e2e-perf e2e-perf/
-COPY engine engine/
-COPY graphql graphql/
-COPY gremlin gremlin/
-COPY grpc grpc/
-COPY grpc-client grpc-client/
-COPY grpcw grpcw/
-COPY integration integration/
-COPY metrics metrics/
-COPY mongodbw mongodbw/
-COPY network network/
-COPY package package/
-COPY postgresw postgresw/
-COPY redisw redisw/
-COPY server server/
-COPY studio studio/
-COPY bindings/python/src bindings/python/src/
-COPY bindings/python/tests bindings/python/tests/
-COPY bindings/python/*.py bindings/python/
-COPY bindings/python/pyproject.toml bindings/python/
-
-# Build all Java modules (skip tests for faster build)
-RUN echo "🔨 Building Java components..." && \
-    mvn clean package -DskipTests -B -q && \
-    echo "✅ Java build complete!"
+ARG DISTRIBUTION
+ARG ARCADEDB_TAG
+
+# Stage 1: Use prebuilt ArcadeDB image to obtain compiled JARs
+# ArcadeDB publishes three separate images:
+#   - arcadedata/arcadedb-headless (minimal JARs, no Studio)
+#   - arcadedata/arcadedb-minimal (includes Studio, excludes Gremlin/GraphQL/MongoDB/Redis)
+#   - arcadedata/arcadedb (full distribution with all features)
+FROM arcadedata/arcadedb:${ARCADEDB_TAG} AS java-builder-full
+FROM arcadedata/arcadedb-minimal:${ARCADEDB_TAG} AS java-builder-minimal  
+FROM arcadedata/arcadedb-headless:${ARCADEDB_TAG} AS java-builder-headless
+
+# Select the right builder based on DISTRIBUTION arg (passed from build-all.sh)
+FROM java-builder-${DISTRIBUTION} AS java-builder
+
+# nothing to do here; jars will be copied from /home/arcadedb/lib in the python-builder stage
 
 # Stage 2: Build Python wheel
+
 FROM python:3.11-slim AS python-builder
 
-# Install system dependencies
+# Install system dependencies (JDK needed for JPype at build/test time)
 RUN apt-get update && apt-get install -y \
     openjdk-21-jdk \
     && rm -rf /var/lib/apt/lists/*
@@ -58,8 +37,10 @@ ENV PATH=$PATH:$JAVA_HOME/bin
 
 WORKDIR /build
 
-# Copy Java build artifacts
-COPY --from=java-builder /arcadedb /arcadedb
+# Copy compiled JARs from the prebuilt ArcadeDB image
+# The official image places JARs under /home/arcadedb/lib
+RUN mkdir -p /build/jars
+COPY --from=java-builder /home/arcadedb/lib /build/jars/
 
 # Copy Python binding source
 COPY bindings/python/src ./src
@@ -69,23 +50,29 @@ COPY bindings/python/extract_version.py .
 COPY bindings/python/write_version.py .
 COPY bindings/python/pyproject.toml ./
 
+# Also copy repository pom.xml so extract_version.py can read it
+COPY ../../pom.xml /arcadedb/pom.xml
+
 # Install Python build dependencies
 RUN pip install --no-cache-dir build wheel setuptools jpype1
 
-# Build argument for distribution type
-ARG DISTRIBUTION=headless
+# Re-declare build args for this stage (required after FROM)
+ARG DISTRIBUTION
+ARG ARCADEDB_TAG
 ENV ARCADEDB_DISTRIBUTION=${DISTRIBUTION}
 
 # Extract version and copy JARs
+# Version is passed from build-all.sh via ARCADEDB_TAG build arg
 RUN echo "📌 Building distribution: ${DISTRIBUTION}" && \
-    export ARCADEDB_VERSION=$(python3 extract_version.py) && \
-    echo "📌 Version: ${ARCADEDB_VERSION}" && \
+    echo "📌 Docker tag used: ${ARCADEDB_TAG}" && \
     python3 write_version.py /arcadedb/pom.xml && \
     python3 setup_jars.py && \
     echo "📦 JAR files copied"
 
 # Build the wheel
-RUN export ARCADEDB_VERSION=$(python3 extract_version.py /arcadedb/pom.xml) && \
+# Extract PEP 440 version from pom.xml
+RUN export ARCADEDB_VERSION=$(python3 extract_version.py --format=pep440 /arcadedb/pom.xml) && \
+    echo "📦 Python package version: ${ARCADEDB_VERSION}" && \
     case ${DISTRIBUTION} in \
         headless) \
             PACKAGE_NAME="arcadedb-embedded-headless" && \

bindings/python/build-all.sh

@@ -55,6 +55,13 @@ if [[ ! "$DISTRIBUTION" =~ ^(all|headless|minimal|full)$ ]]; then
     exit 1
 fi
 
+# Auto-detect Docker tag from pom.xml
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+echo -e "${CYAN}🔍 Detecting version from pom.xml...${NC}"
+DOCKER_TAG=$(python3 "$SCRIPT_DIR/extract_version.py" --format=docker)
+echo -e "${CYAN}📌 Docker tag: ${YELLOW}${DOCKER_TAG}${NC}"
+echo ""
+
 # Check if Docker is available
 if ! command -v docker &> /dev/null; then
     echo -e "${RED}❌ Docker is not installed or not in PATH${NC}"
@@ -95,6 +102,7 @@ build_distribution() {
     echo -e "${CYAN}📦 Building Docker image for $dist distribution...${NC}"
     docker build \
         --build-arg DISTRIBUTION=$dist \
+        --build-arg ARCADEDB_TAG=$DOCKER_TAG \
         --target export \
         -t arcadedb-python-bindings-$dist-export \
         -f Dockerfile.build \
@@ -104,6 +112,7 @@ build_distribution() {
     echo -e "${CYAN}🧪 Running tests in Docker...${NC}"
     docker build \
         --build-arg DISTRIBUTION=$dist \
+        --build-arg ARCADEDB_TAG=$DOCKER_TAG \
         --target tester \
         -t arcadedb-python-bindings-$dist \
         -f Dockerfile.build \

bindings/python/extract_version.py

@@ -1,61 +1,95 @@
 #!/usr/bin/env python3
-"""Extract ArcadeDB version from parent pom.xml"""
+"""Extract ArcadeDB version from parent pom.xml
+
+Usage:
+    python extract_version.py [pom_path] [--format=pep440|docker]
+    
+    --format=pep440  : Convert to PEP 440 for Python packaging (default)
+                       25.10.1-SNAPSHOT -> 25.10.1.dev0
+    --format=docker  : Raw Maven version for Docker tags
+                       25.10.1-SNAPSHOT -> 25.10.1-SNAPSHOT
+"""
 import re
 import sys
 from pathlib import Path
 
 # Python-specific patch version (increment for Python-only releases)
 PYTHON_PATCH = 0  # Change this to 2, 3, etc. for subsequent Python patches
 
-def extract_version_from_pom(pom_path):
-    """Extract version from Maven pom.xml and convert to PEP 440 format"""
-    with open(pom_path, 'r') as f:
+
+def extract_raw_version_from_pom(pom_file):
+    """Extract the raw version string from pom.xml (no conversion)"""
+    with open(pom_file, 'r', encoding='utf-8') as f:
         content = f.read()
 
     # Find the first <version> tag after <artifactId>arcadedb-parent</artifactId>
     pattern = r'<artifactId>arcadedb-parent</artifactId>.*?<version>(.*?)</version>'
     match = re.search(pattern, content, re.DOTALL)
 
     if match:
-        version = match.group(1).strip()
-        # Convert Maven version to PEP 440 compatible version
-        # Examples:
-        #   25.10.1-SNAPSHOT -> 25.10.1.dev0
-        #   25.10.1-RC1 -> 25.10.1rc1
-        #   25.10.1 -> 25.10.1
-        
-        if '-SNAPSHOT' in version:
-            version = version.replace('-SNAPSHOT', '.dev0')
-        elif '-RC' in version:
-            version = version.replace('-RC', 'rc')
-        elif '-' in version:
-            # Generic replacement for other pre-release identifiers
-            version = version.replace('-', '')
-        
-        # Add Python patch version for Python-specific releases
-        # PYTHON_PATCH = 0 → 25.9.1 (matches Java version)
-        # PYTHON_PATCH = 1 → 25.9.1.1 (first Python-only patch)
-        # PYTHON_PATCH = 2 → 25.9.1.2 (second Python-only patch)
-        # Not applied to dev versions (e.g., 25.9.1.dev0)
-        if PYTHON_PATCH > 0 and '.dev' not in version and 'rc' not in version.lower():
-            version = f"{version}.{PYTHON_PATCH}"
-        
-        return version
+        return match.group(1).strip()
 
     raise ValueError("Could not find version in pom.xml")
 
 
+def extract_version_from_pom(pom_file, fmt='pep440'):
+    """Extract version from Maven pom.xml
+    
+    Args:
+        pom_file: Path to pom.xml file
+        fmt: 'pep440' for Python packaging, 'docker' for Docker tags
+    """
+    ver = extract_raw_version_from_pom(pom_file)
+    
+    if fmt == 'docker':
+        # Return raw Maven version for Docker tags
+        return ver
+    
+    # Convert Maven version to PEP 440 compatible version
+    # Examples:
+    #   25.10.1-SNAPSHOT -> 25.10.1.dev0
+    #   25.10.1-RC1 -> 25.10.1rc1
+    #   25.10.1 -> 25.10.1
+    
+    if '-SNAPSHOT' in ver:
+        ver = ver.replace('-SNAPSHOT', '.dev0')
+    elif '-RC' in ver:
+        ver = ver.replace('-RC', 'rc')
+    elif '-' in ver:
+        # Generic replacement for other pre-release identifiers
+        ver = ver.replace('-', '')
+    
+    # Add Python patch version for Python-specific releases
+    # PYTHON_PATCH = 0 → 25.9.1 (matches Java version)
+    # PYTHON_PATCH = 1 → 25.9.1.1 (first Python-only patch)
+    # PYTHON_PATCH = 2 → 25.9.1.2 (second Python-only patch)
+    # Not applied to dev versions (e.g., 25.9.1.dev0)
+    if PYTHON_PATCH > 0 and '.dev' not in ver and 'rc' not in ver.lower():
+        ver = f"{ver}.{PYTHON_PATCH}"
+    
+    return ver
+
+
 if __name__ == "__main__":
     # Default to parent pom.xml (two levels up from bindings/python)
     script_dir = Path(__file__).parent
     pom_path = script_dir / "../../pom.xml"
+    output_format = 'pep440'
 
-    if len(sys.argv) > 1:
-        pom_path = Path(sys.argv[1])
+    # Parse arguments
+    for arg in sys.argv[1:]:
+        if arg.startswith('--format='):
+            output_format = arg.split('=', 1)[1]
+            if output_format not in ['pep440', 'docker']:
+                print("Error: Invalid format. Use 'pep440' or 'docker'",
+                      file=sys.stderr)
+                sys.exit(1)
+        elif not arg.startswith('--'):
+            pom_path = Path(arg)
 
     try:
-        version = extract_version_from_pom(pom_path)
-        print(version)
-    except Exception as e:
+        ver = extract_version_from_pom(pom_path, output_format)
+        print(ver)
+    except (ValueError, FileNotFoundError, OSError) as e:
         print(f"Error: {e}", file=sys.stderr)
         sys.exit(1)