fix(tika): corregge rilevamento JAR Tika con supporto versioni e env var TIKA_JAR_PATH

Author: strawberry-code

CHANGELOG.md

@@ -12,6 +12,8 @@ and the project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.
 - Ollama context size reduced from 8192 to 2048 to match nomic-embed-text model limit
 - Centralized logging configuration for consistent pipeline logs visibility
 - UI now shows actual error messages instead of generic failure messages
+- Tika JAR detection now supports versioned filenames and TIKA_JAR_PATH env var
+- Improved Tika pre-download in Dockerfile.tika with verification and copy to expected path
 
 ## [1.1.2] - 2025-12-02
 

Dockerfile.tika

@@ -41,6 +41,8 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
 RUN ln -s /usr/lib/jvm/java-21-openjdk-* /usr/lib/jvm/java-21
 ENV JAVA_HOME=/usr/lib/jvm/java-21
 ENV PATH="${JAVA_HOME}/bin:${PATH}"
+# Tika JAR path (set after pre-download, used by lib/tika_check.py)
+ENV TIKA_JAR_PATH=/tmp/tika-server.jar
 
 # Install Ollama
 RUN curl -fsSL https://ollama.ai/install.sh | sh
@@ -81,8 +83,23 @@ RUN ollama serve & \
     ollama pull nomic-embed-text && \
     pkill ollama || true
 
-# Pre-download Tika JAR (triggers auto-download to /tmp/tika-server.jar)
-RUN python -c "from tika import parser; parser.from_buffer('test', xmlContent=False)" || true
+# Pre-download Tika JAR and ensure it's in the expected location
+RUN python3 -c "\
+import os, shutil; \
+from tika import parser, tika; \
+# Trigger download \
+parser.from_buffer(b'test', xmlContent=False); \
+# Get the actual JAR path \
+jar_path = getattr(tika, 'TikaJarPath', None); \
+print(f'Tika JAR downloaded to: {jar_path}'); \
+# Copy to expected location if different \
+if jar_path and os.path.exists(jar_path) and jar_path != '/tmp/tika-server.jar': \
+    shutil.copy2(jar_path, '/tmp/tika-server.jar'); \
+    print('Copied to /tmp/tika-server.jar'); \
+# Verify \
+assert os.path.exists('/tmp/tika-server.jar'), 'Tika JAR not found at /tmp/tika-server.jar'; \
+print('Tika JAR verified at /tmp/tika-server.jar'); \
+" || echo "WARNING: Tika JAR pre-download failed, will download at runtime"
 
 # Create data directory for Qdrant
 RUN mkdir -p /data/qdrant

api/routes/upload.py

@@ -115,15 +115,19 @@ def run_indexing(job_id: str, collection_dir: Path, collection: str, filenames:
         # Import RagifyPipeline
         from ragify import RagifyPipeline
         from lib.config import RagifyConfig
-        from lib.tika_check import is_tika_available
+        from lib.tika_check import is_tika_available, check_tika_available
 
         # Configure
         config = RagifyConfig.default()
         config.qdrant.collection = collection
 
-        # Check Tika availability
-        use_tika = is_tika_available()
-        logger.info(f"[{job_id}] Tika available: {use_tika}")
+        # Check Tika availability with diagnostic info
+        tika_status = check_tika_available()
+        use_tika = tika_status['can_use_tika']
+        logger.info(f"[{job_id}] Tika status: java={tika_status['java_installed']}, "
+                    f"jar={tika_status['tika_jar_available']}, path={tika_status.get('tika_jar_path')}")
+        if tika_status['issues']:
+            logger.warning(f"[{job_id}] Tika issues: {tika_status['issues']}")
 
         jobs[job_id]["progress"] = 0.2
         jobs[job_id]["message"] = f"Processing with {'Tika' if use_tika else 'text-only'} mode"

lib/tika_check.py

@@ -49,23 +49,41 @@ def check_tika_jar_available() -> Tuple[bool, Optional[Path]]:
     Returns:
         Tuple of (is_available, jar_path)
     """
-    # Tika downloads to /tmp on Unix systems
-    possible_paths = [
+    # Check environment variable first (for Docker containers)
+    env_path = os.getenv('TIKA_JAR_PATH')
+    if env_path:
+        jar_path = Path(env_path)
+        if jar_path.exists():
+            logger.debug(f"Tika JAR found via TIKA_JAR_PATH: {jar_path}")
+            return True, jar_path
+
+    # Static paths to check
+    static_paths = [
         Path('/tmp/tika-server.jar'),
-        Path('/var/folders').rglob('tika-server.jar'),  # macOS temp
         Path.home() / '.tika' / 'tika-server.jar',
     ]
 
-    for path in possible_paths:
-        if isinstance(path, Path) and path.exists():
+    for path in static_paths:
+        if path.exists():
+            logger.debug(f"Tika JAR found at: {path}")
             return True, path
-        # For glob results
-        try:
-            for found_path in path:
-                if found_path.exists():
-                    return True, found_path
-        except (TypeError, AttributeError):
-            continue
+
+    # Glob patterns for versioned JARs (tika-server-X.Y.Z.jar)
+    glob_patterns = [
+        (Path('/tmp'), 'tika-server*.jar'),
+        (Path.home() / '.tika', 'tika-server*.jar'),
+        (Path('/var/folders'), '**/tika-server*.jar'),  # macOS temp
+    ]
+
+    for base_path, pattern in glob_patterns:
+        if base_path.exists():
+            try:
+                for found_path in base_path.glob(pattern):
+                    if found_path.exists() and found_path.is_file():
+                        logger.debug(f"Tika JAR found via glob: {found_path}")
+                        return True, found_path
+            except (PermissionError, OSError):
+                continue
 
     return False, None
 
@@ -239,13 +257,34 @@ def ensure_tika_ready(interactive: bool = True, auto_skip: bool = False) -> bool
 
 def is_tika_available() -> bool:
     """
-    Simple check if Tika is available for use.
+    Check if Tika is available for use.
+
+    First checks via check_tika_available(), then falls back to trying
+    tika-python directly in case the JAR is in an unexpected location.
 
     Returns:
-        True if Java and Tika JAR are available, False otherwise.
+        True if Tika can be used, False otherwise.
     """
+    # First try standard check
     status = check_tika_available()
-    return status['can_use_tika']
+    if status['can_use_tika']:
+        return True
+
+    # Fallback: try tika-python directly (it knows its own JAR location)
+    if status['java_installed']:
+        try:
+            from tika import tika
+            # Check if tika-python has a valid JAR path
+            jar_path = getattr(tika, 'TikaJarPath', None)
+            if jar_path and Path(jar_path).exists():
+                logger.info(f"Tika JAR found via tika-python: {jar_path}")
+                return True
+        except ImportError:
+            pass
+        except Exception as e:
+            logger.debug(f"Tika-python fallback check failed: {e}")
+
+    return False
 
 
 def print_tika_status():