fix(docs): add poppler-utils dependency for PDF document processing

- Add poppler-utils to Dockerfile runtime dependencies
- Document poppler-utils installation in README.md prerequisites
- Add early validation check for poppler-utils availability
- Improve error messages with platform-specific installation instructions
- Fix logger initialization order in nemo_retriever.py

Resolves PDF processing failures when poppler-utils is not installed.
Required by pdf2image package for PDF to image conversion.

Author: T-DevH

Dockerfile

@@ -60,6 +60,7 @@ WORKDIR /app
 RUN apt-get update && apt-get install -y \
     curl \
     git \
+    poppler-utils \
     && rm -rf /var/lib/apt/lists/*
 
 # Copy Python dependencies from backend-deps stage

README.md

@@ -218,6 +218,11 @@ See the [Local Development Setup](#local-development-setup) section below for ma
   - **macOS**: `brew install postgresql` or `brew install libpq`
   - **Windows**: Install from [PostgreSQL downloads](https://www.postgresql.org/download/windows/)
   - **Alternative**: Use Docker (see [DEPLOYMENT.md](DEPLOYMENT.md))
+- **Poppler utilities** (`poppler-utils`) - Required for PDF document processing
+  - **Ubuntu/Debian**: `sudo apt-get install poppler-utils`
+  - **macOS**: `brew install poppler`
+  - **Windows**: Install from [Poppler for Windows](http://blog.alivate.com.au/poppler-windows/) or use Chocolatey: `choco install poppler`
+  - **Note**: Required by `pdf2image` package for converting PDF pages to images
 - **CUDA (for GPU acceleration)** - Optional but recommended for RAPIDS GPU-accelerated forecasting
   - **Recommended**: CUDA 12.x (default for RAPIDS packages)
   - **Supported**: CUDA 11.x (via `install_rapids.sh` auto-detection)

src/api/agents/document/preprocessing/nemo_retriever.py

@@ -29,6 +29,8 @@
 from PIL import Image
 import io
 
+logger = logging.getLogger(__name__)
+
 # Try to import pdf2image, fallback to None if not available
 try:
     from pdf2image import convert_from_path
@@ -37,7 +39,16 @@
     PDF2IMAGE_AVAILABLE = False
     logger.warning("pdf2image not available. PDF processing will be limited. Install with: pip install pdf2image")
 
-logger = logging.getLogger(__name__)
+
+def _check_poppler_available() -> bool:
+    """
+    Check if poppler-utils is installed and available in PATH.
+    
+    Returns:
+        True if poppler-utils is available, False otherwise
+    """
+    import shutil
+    return shutil.which("pdfinfo") is not None
 
 
 class NeMoRetrieverPreprocessor:
@@ -205,6 +216,15 @@ async def _extract_pdf_images(self, file_path: str) -> List[Image.Image]:
                     "Also requires poppler-utils system package: sudo apt-get install poppler-utils"
                 )
 
+            # Check if poppler-utils is available before attempting conversion
+            if not _check_poppler_available():
+                raise RuntimeError(
+                    "poppler-utils is not installed or not in PATH. "
+                    "Install it with: sudo apt-get install poppler-utils (Ubuntu/Debian) "
+                    "or brew install poppler (macOS). "
+                    "This is required for PDF to image conversion."
+                )
+            
             logger.info(f"Converting PDF to images: {file_path}")
 
             # Limit pages for faster processing