Better CSV and DOCX support

app.py

@@ -416,10 +416,18 @@ def query_agent(question: str, files: Optional[List[FilePayload]] = None) -> str
             try:
                 # Import necessary types and libraries inside logic to avoid top-level failures if missing
                 import io
+                import csv
                 import pypdf
                 from strands.types.content import ImageContent
                 from strands.types.media import ImageSource
                 
+                # Try import python-docx
+                try:
+                    import docx
+                except ImportError:
+                    docx = None
+                    logger.warning("python-docx not installed. DOCX support disabled.")
+
                 image_formats = {'png', 'jpeg', 'gif', 'webp', 'jpg'}
                 
                 for file_obj in files:
@@ -461,8 +469,47 @@ def query_agent(question: str, files: Optional[List[FilePayload]] = None) -> str
                                 except Exception as pdf_err:
                                     logger.error(f"PDF extraction failed for {file_obj.name}: {pdf_err}")
                                     extracted_text = f"[Error extracting PDF text for {file_obj.name}: {str(pdf_err)}]"
+
+                            elif fmt == 'docx':
+                                if docx:
+                                    try:
+                                        doc = docx.Document(io.BytesIO(file_bytes))
+                                        full_text = []
+                                        for para in doc.paragraphs:
+                                            full_text.append(para.text)
+                                        extracted_text = '\n'.join(full_text)
+
+                                        if not extracted_text.strip():
+                                             extracted_text = "[No text found in this DOCX file.]"
+
+                                    except Exception as docx_err:
+                                        logger.error(f"DOCX extraction failed for {file_obj.name}: {docx_err}")
+                                        extracted_text = f"[Error extracting DOCX text for {file_obj.name}: {str(docx_err)}]"
+                                else:
+                                    extracted_text = "[DOCX support is not available. Please install python-docx.]"
+
+                            elif fmt == 'csv':
+                                try:
+                                    # Decode bytes to string
+                                    csv_text = file_bytes.decode('utf-8', errors='replace')
+                                    csv_file = io.StringIO(csv_text)
+                                    csv_reader = csv.reader(csv_file)
+
+                                    rows = []
+                                    for row in csv_reader:
+                                        rows.append(','.join(row))
+
+                                    extracted_text = '\n'.join(rows)
+
+                                    if not extracted_text.strip():
+                                         extracted_text = "[Empty CSV file.]"
+
+                                except Exception as csv_err:
+                                    logger.error(f"CSV extraction failed for {file_obj.name}: {csv_err}")
+                                    extracted_text = f"[Error extracting CSV text for {file_obj.name}: {str(csv_err)}]"
+
                             else:
-                                # Try decoding as plain text (csv, txt, md, html, etc)
+                                # Try decoding as plain text (txt, md, html, etc)
                                 try:
                                     extracted_text = file_bytes.decode('utf-8', errors='replace')
                                 except Exception as dec_err:
@@ -477,7 +524,7 @@ def query_agent(question: str, files: Optional[List[FilePayload]] = None) -> str
                 
             except ImportError as ie:
                 logger.error(f"Missing dependency for file processing: {ie}")
-                return "Error: Server missing dependencies (pypdf or strands types) for file processing."
+                return "Error: Server missing dependencies (pypdf, python-docx, or strands types) for file processing."
 
         # Construct final payload
         message_content = [{"text": combined_text}]