Adding git and git-lfs to the HF spaces

Dockerfile

@@ -3,10 +3,12 @@ FROM python:3.10-slim
 # Set working directory
 WORKDIR /app
 
-# Install system dependencies
+# Install system dependencies including git for PyABSA
 RUN apt-get update && apt-get install -y \
     build-essential \
     curl \
+    git \
+    git-lfs \
     && rm -rf /var/lib/apt/lists/*
 
 # Copy all files first
@@ -15,6 +17,9 @@ COPY . .
 # Install Python dependencies
 RUN pip install --no-cache-dir -r requirements.txt
 
+# Pre-download PyABSA checkpoint during build to avoid runtime git issues
+RUN python -c "import pyabsa; from pyabsa import ATEPCCheckpointManager; ATEPCCheckpointManager.get_checkpoint('multilingual')" || echo "PyABSA checkpoint download failed, will retry at runtime"
+
 # Expose port
 EXPOSE 7860
 

PYABSA_FIX.md

@@ -0,0 +1,109 @@
+# PyABSA Loading Fix - Deployment Checklist
+
+## Changes Made
+
+### 1. Dockerfile Updates ✅
+- Added `git` and `git-lfs` to system dependencies
+- Added PyABSA checkpoint pre-download during Docker build
+- This ensures the model is ready before the app starts
+
+### 2. Data Processor Improvements ✅
+- Added git availability check with detailed logging
+- Better error handling with full traceback
+- Attempts local checkpoint first, then downloads if needed
+- Clear status messages (✅, ❌, ⚠️) for easy debugging
+
+### 3. Environment Configuration ✅
+- Set `GIT_PYTHON_REFRESH=quiet` to suppress git warnings
+- Suppressed deprecation warnings from PyABSA
+
+## Deployment Steps
+
+### Option A: Redeploy to HuggingFace Spaces
+
+1. **Commit changes:**
+   ```bash
+   cd c:\Users\Lenovo\insights\ABSA
+   git add Dockerfile src/utils/data_processor.py
+   git commit -m "Fix PyABSA loading with git installation"
+   git push
+   ```
+
+2. **HF Spaces will automatically rebuild** with the new Dockerfile
+3. **Check logs** to verify PyABSA loads:
+   - Look for: `✅ PyABSA model loaded successfully`
+   - Should NOT see: `⚠️ Using fallback method`
+
+### Option B: Test Locally First
+
+1. **Rebuild Docker image:**
+   ```bash
+   cd c:\Users\Lenovo\insights\ABSA
+   docker build -t absa-backend .
+   ```
+
+2. **Run container:**
+   ```bash
+   docker run -p 7860:7860 --env-file .env absa-backend
+   ```
+
+3. **Check logs for:**
+   ```
+   Git available: git version X.X.X
+   ✅ PyABSA model loaded successfully
+   ```
+
+## Expected Behavior
+
+### Before Fix ❌
+```
+WARNING:utils.data_processor:Failed to load PyABSA model: Bad git executable
+⚠️ Using fallback method
+```
+
+### After Fix ✅
+```
+INFO:utils.data_processor:Git available: git version 2.X.X
+INFO:utils.data_processor:This may take a few minutes on first run...
+✅ PyABSA model loaded successfully from downloaded checkpoint
+```
+
+## Troubleshooting
+
+### If PyABSA still fails to load:
+
+1. **Check Docker build logs:**
+   - Verify git was installed successfully
+   - Check if PyABSA checkpoint download during build succeeded
+
+2. **Check runtime logs:**
+   - Look for full traceback after "❌ Failed to load PyABSA model"
+   - Verify git command is available
+
+3. **Manual checkpoint download** (if needed):
+   ```python
+   # Run this once to cache the checkpoint
+   import pyabsa
+   from pyabsa import ATEPCCheckpointManager
+   checkpoint = ATEPCCheckpointManager.get_checkpoint('multilingual')
+   print(f"Checkpoint saved to: {checkpoint}")
+   ```
+
+## What This Fixes
+
+- ✅ PyABSA can now download checkpoints (git available)
+- ✅ Better error messages for debugging
+- ✅ Checkpoint pre-downloaded during build (faster startup)
+- ✅ Fallback still works if PyABSA fails
+- ✅ No more cryptic git errors in production
+
+## Next Steps
+
+After successful deployment, verify in HF Spaces logs:
+1. Git is detected
+2. PyABSA loads without errors
+3. Reviews are processed with actual aspect extraction (not fallback)
+
+---
+
+**Deploy now:** Push changes to trigger HF Spaces rebuild! 🚀

src/utils/data_processor.py

@@ -18,6 +18,9 @@ import requests
 import os
 import time
 
+# Suppress GitPython warnings for PyABSA
+os.environ['GIT_PYTHON_REFRESH'] = 'quiet'
+
 # Set up logging
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
@@ -207,16 +210,44 @@ class ABSAProcessor:
     def _load_pyabsa_model(self):
         """Load PyABSA multilingual model with caching."""
         try:
+            # Suppress additional git warnings
+            import warnings
+            warnings.filterwarnings('ignore', category=DeprecationWarning)
+            
             import pyabsa
             from pyabsa import ATEPCCheckpointManager
             
-            # Use multilingual checkpoint - works for English and some Hindi
-            checkpoint = ATEPCCheckpointManager.get_checkpoint('multilingual')
-            self.model = pyabsa.load_aspect_extractor(checkpoint=checkpoint)
-            logger.info("PyABSA model loaded successfully")
+            # Check if git is available
+            try:
+                import subprocess
+                git_check = subprocess.run(['git', '--version'], capture_output=True, timeout=5)
+                if git_check.returncode == 0:
+                    logger.info(f"Git available: {git_check.stdout.decode().strip()}")
+                else:
+                    logger.warning("Git command failed, PyABSA may have issues downloading checkpoints")
+            except Exception as git_err:
+                logger.warning(f"Git not found or not executable: {str(git_err)}")
+            
+            # Try local checkpoint first (if exists in checkpoints/ directory)
+            local_checkpoint = os.path.join(os.path.dirname(__file__), '..', '..', 'checkpoints', 'ATEPC_MULTILINGUAL_CHECKPOINT')
+            
+            if os.path.exists(local_checkpoint):
+                logger.info(f"Loading PyABSA from local checkpoint: {local_checkpoint}")
+                self.model = pyabsa.load_aspect_extractor(checkpoint=local_checkpoint)
+                logger.info("✅ PyABSA model loaded successfully from local checkpoint")
+            else:
+                # Fallback to downloading checkpoint (requires git)
+                logger.info("Local checkpoint not found, downloading multilingual checkpoint...")
+                logger.info("This may take a few minutes on first run...")
+                checkpoint = ATEPCCheckpointManager.get_checkpoint('multilingual')
+                self.model = pyabsa.load_aspect_extractor(checkpoint=checkpoint)
+                logger.info("✅ PyABSA model loaded successfully from downloaded checkpoint")
             
         except Exception as e:
-            logger.warning(f"Failed to load PyABSA model: {str(e)}. Using fallback.")
+            import traceback
+            logger.error(f"❌ Failed to load PyABSA model: {str(e)}")
+            logger.error(f"Traceback: {traceback.format_exc()}")
+            logger.warning("⚠️ Using fallback method for aspect extraction")
             self.model = None
     
     def set_task_manager(self, task_manager):