v0.2.36 png support

VERSION

@@ -1 +1 @@
-v0.2.35
+v0.2.36

impl/main.py

@@ -3,6 +3,7 @@
 import logging
 import os
 import sys
+import traceback
 from typing import Callable, Sequence, Union, Any
 
 import httpx
@@ -102,29 +103,45 @@ async def shutdown_event():
 
 class APIVersionMiddleware(BaseHTTPMiddleware):
     async def dispatch(self, request: Request, call_next):
-        version_header = request.headers.get("OpenAI-Beta")
-        if version_header is None or version_header == "assistants=v1":
-            response = await call_next(request)
-            return response
-        if version_header == "assistants=v2":
-            request.scope['path'] = request.scope['path'].replace("v1", "v2")
-            if 'raw_path' in request.scope:
-                request.scope['raw_path'] = request.scope['raw_path'].replace(b'v1', b'v2')
-            try:
-                response = await call_next(request)
-                return response
-            except Exception as e:
-                if hasattr(request.state, "dbid"):
-                    logger.error(f"Error: {e}, dbid: {request.state.dbid}")
-                else:
-                    logger.error(f"Error: {e}")
-                print(e)
-                raise e
+        # Retrieve and normalize the version header
+        version_header = request.headers.get("OpenAI-Beta", "").lower()
 
-        else:
+        try:
+            if version_header in [None, "", "assistants=v1"]:
+                # Default version or v1: Proceed as is
+                return await call_next(request)
+
+            if version_header == "assistants=v2":
+                # Modify path for v2 requests
+                request.scope['path'] = request.scope['path'].replace("v1", "v2")
+                if 'raw_path' in request.scope:
+                    request.scope['raw_path'] = request.scope['raw_path'].replace(b'v1', b'v2')
+
+                # Proceed with the modified request
+                return await call_next(request)
+
+            # Unsupported version
             return Response(
-                f"Unsupported version: {version_header})",
-                status_code=400)
+                content=f"Unsupported version: {version_header}",
+                status_code=400
+            )
+
+        except Exception as e:
+            # Structured logging for errors
+            trace = traceback.format_exc()
+            dbid = getattr(request.state, "dbid", None)
+            error_message = f"Error processing request: {e}\nTraceback: {trace}"
+            if dbid:
+                logger.error(f"{error_message}, dbid: {dbid}")
+            else:
+                logger.error(error_message)
+
+            # Return a generic internal server error response
+            return Response(
+                content="An internal server error occurred.",
+                status_code=500
+            )
+
 
 
 app.add_middleware(APIVersionMiddleware)

impl/services/file.py

@@ -1,6 +1,7 @@
 import csv
 import mimetypes
 import os
+import traceback
 from io import BufferedReader
 from typing import Optional, List
 
@@ -9,7 +10,7 @@
 from fastapi import UploadFile, HTTPException
 from loguru import logger
 from PyPDF2 import PdfReader
-from markitdown import MarkItDown
+from markitdown import MarkItDown, UnsupportedFormatException
 from openai import OpenAI
 
 from impl.astra_vector import HandledResponse
@@ -112,7 +113,7 @@ def extract_text_from_file(file: BufferedReader, mimetype: str, openai_api_key,
         # Extract text from pdf using PyPDF2
         reader = PdfReader(file)
         extracted_text = " ".join([page.extract_text() for page in reader.pages])
-    elif mimetype == "text/plain" or mimetype == "text/markdown" or mimetype == "application/sql":
+    elif mimetype == "text/plain" or mimetype == "text/markdown" or mimetype == "application/sql" or mimetype == "message/rfc822":
         # Read text from plain text file
         extracted_text = file.read().decode("utf-8")
     elif (
@@ -145,15 +146,27 @@ def extract_text_from_file(file: BufferedReader, mimetype: str, openai_api_key,
                             extracted_text += run.text + " "
                     extracted_text += "\n"
     else:
+        if openai_api_key is None:
+            raise HTTPException(
+                status_code=400,
+                detail="OpenAI API key is required for this file type: {}\nRemember to set your OPENAI_API_KEY env var".format(mimetype),
+            )
         client = OpenAI(api_key=openai_api_key)
-        md = MarkItDown(mlm_client=client, mlm_model="gpt-4o")
+        md = MarkItDown(llm_client=client, llm_model="gpt-4o")
         try:
             extracted_text = md.convert(source=filepath).text_content
+        except UnsupportedFormatException as e:
+            logger.error(f"markitdown unsupported exception for: {mimetype}\nerror: {e}")
+            raise HTTPException(
+                status_code=400,
+                detail="Unsupported file type: {}\nError: {}".format(mimetype, e)
+            )
         except Exception as e:
+            logger.error(f"Error: {e}\nTrace: {traceback.format_exc()}")
             # Unsupported file type
             raise HTTPException(
                 status_code=400,
-                detail="Unsupported file type: {}".format(mimetype),
+                detail="Unknown error parsing file type: {}\nError: {}".format(mimetype, e)
             )
     return extracted_text