super-document-processor 0.1.3__py3-none-any.whl

super_document_processor/document_processors/process_images.py

@@ -0,0 +1,39 @@
+import os
+import logging
+
+logger = logging.getLogger("super_document_processor.images")
+
+try:
+    from super_document_processor.utils.vlm import generate_image_description
+except ImportError:
+    from utils.vlm import generate_image_description
+
+def process_image_file(img_path: str, original_filename: str = None, process_images: bool = None, model: str = None) -> dict:
+    if not isinstance(img_path, str):
+        raise TypeError(f"Image input must be a file path string, got {type(img_path)}")
+    
+    display_name = os.path.basename(original_filename) if original_filename else os.path.basename(img_path)
+
+    if not process_images:
+        raise ValueError(
+            f"Processing image file '{display_name}' requires image processing to be enabled.\n"
+            "Please enable image processing via CLI (--process-images) or API ('process_images': True)."
+        )
+
+    if not model:
+        raise ValueError(
+            "Cannot process images: No vision model specified.\n"
+            "Please specify a vision-capable model via CLI (--model) or API ('model' parameter)."
+        )
+
+    with open(img_path, "rb") as f:
+        image_bytes = f.read()
+
+    description = generate_image_description(image_bytes, model=model)
+    logger.debug(f"Generated description for image '{display_name}'")
+    
+    return {
+        "content": description,
+        "chunkIndex": 1,
+        "filename": display_name
+    }

super_document_processor/document_processors/process_msg.py

@@ -0,0 +1,62 @@
+import os
+import logging
+import extract_msg
+
+logger = logging.getLogger("super_document_processor.msg")
+
+try:
+    from super_document_processor.utils.chunker import collector
+except ImportError:
+    from utils.chunker import collector
+
+def process_msg_file(filename, original_filename=None):
+    """
+    Extract text content from a .msg (Outlook email) file.
+    Returns chunks of content similar to other document processors.
+    """
+    display_name = os.path.basename(original_filename) if original_filename else os.path.basename(filename)
+    try:
+        # Open and extract the MSG file
+        msg = extract_msg.Message(filename)
+        
+        # Extract main content
+        content_parts = []
+        
+        # Subject
+        if msg.subject:
+            content_parts.append(f"Subject: {msg.subject}")
+        
+        # Sender and recipients
+        if msg.sender:
+            content_parts.append(f"From: {msg.sender}")
+        if msg.to:
+            content_parts.append(f"To: {msg.to}")
+        if msg.cc:
+            content_parts.append(f"CC: {msg.cc}")
+        
+        # Email body
+        if msg.body:
+            content_parts.append(f"\n{msg.body}")
+        
+        # Combine all content
+        full_content = "\n".join(content_parts)
+        
+        # Add to collector for chunking
+        collector.add(full_content)
+        
+        # Get chunked texts
+        all_chunks = []
+        chunked_texts = collector.get_chunked_texts()
+        for i, chunk in enumerate(chunked_texts, 1):
+            all_chunks.append({
+                "content": f"--- Chunk {i} ---\n{chunk}",
+                "chunkIndex": i,
+                "filename": display_name
+            })
+        
+        collector.clear()
+        return all_chunks
+        
+    except Exception as e:
+        logger.error(f"Error processing MSG file '{display_name}': {e}")
+        return []