markitdown-pro 1.0.2__tar.gz → 1.0.4__tar.gz

{markitdown_pro-1.0.2 → markitdown_pro-1.0.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: markitdown-pro
-Version: 1.0.2
+Version: 1.0.4
 Summary: A package that converts almost any file format to Markdown.
 Author: Developer
 Classifier: Programming Language :: Python :: 3

markitdown_pro-1.0.4/markitdown_pro/common/logger.py

@@ -0,0 +1,24 @@
+import logging
+import os
+
+log_level = int(os.getenv("LOG_LEVEL", "20"))
+logger = logging.getLogger(__name__)
+logging.basicConfig(level=log_level, format="[%(asctime)s] [%(levelname)s] %(message)s")
+
+NOISY_LIST = [
+    "azure.core.pipeline.policies.http_logging_policy",
+    "azure.core.pipeline.policies._universal",
+    "Transmission succeeded...",
+    "azure.monitor.opentelemetry.exporter",
+    "azure.monitor.opentelemetry.exporter._transmission",
+    "azure",
+    "opentelemetry",
+    "urllib3",
+    "urllib3.connectionpool",
+    "httpx",
+    "httpcore",
+    "openai",
+]
+
+for name in NOISY_LIST:
+    logging.getLogger(name).setLevel(logging.WARNING)

{markitdown_pro-1.0.2 → markitdown_pro-1.0.4}/markitdown_pro/conversion_pipeline.py

@@ -186,14 +186,12 @@ class ConversionPipeline:
             raise RuntimeError(f"No handler for extension '{extension}'.")
 
         try:
-            logger.debug(f"convert_document_to_md: Using handler for extension '{extension}'")
             md_content = await handler.handle(str(file_path))
             if not md_content or not ensure_minimum_content(md_content):
                 return None
 
             logger.debug(
-                "convert_document_to_md: Handler returned %d characters",
-                len(md_content),
+                f"convert_document_to_md: {file_path} returned {len(md_content)} characters"
             )
             return _write_md(output_md, md_content) if output_md else md_content
 

{markitdown_pro-1.0.2 → markitdown_pro-1.0.4}/markitdown_pro/handlers/pdf_handler.py

@@ -164,8 +164,12 @@ class PDFHandler(BaseHandler):
                         if page.get_images(full=True):
                             pages_with_images += 1
 
-                    logger.debug(f"Pages with text: {pages_with_text}/{total_pages}")
-                    logger.debug(f"Pages with images: {pages_with_images}/{total_pages}")
+                    logger.debug(
+                        f"PDFHandler: Pages with text for {file_path}: {pages_with_text}/{total_pages}"
+                    )
+                    logger.debug(
+                        f"PDFHandler: Pages with images for {file_path}: {pages_with_images}/{total_pages}"
+                    )
 
                     is_text_only = pages_with_text == total_pages and pages_with_images == 0
                     is_all_images = pages_with_images == total_pages and pages_with_text == 0

{markitdown_pro-1.0.2 → markitdown_pro-1.0.4}/markitdown_pro/services/openai_services.py

@@ -224,44 +224,37 @@ class GPTVision:
         if not self.lang_client:
             logger.error("GPT-vision client is not initialized.")
             raise ValueError("GPT-vision client is not initialized.")
-
         attempt = 0
-        payload = list(messages)  # ensure materialized list
-
+        payload = list(messages)
         while True:
             try:
-                task = asyncio.create_task(self.lang_client.ainvoke(payload))
-                try:
-                    response = await asyncio.wait_for(task, timeout=self.request_timeout_s)
-                except asyncio.TimeoutError as te:
-                    # Cancel in-flight task cleanly
-                    task.cancel()
-                    with suppress(asyncio.CancelledError):
-                        await task
-                    raise te
-
+                async with asyncio.timeout(self.request_timeout_s):
+                    response = await self.lang_client.ainvoke(payload)
                 if not response:
                     logger.error("No response from GPT-vision client.")
                     return None
                 return response
-
+            except asyncio.CancelledError:
+                # If caller cancels (shutdown), propagate immediately
+                logger.error(f"ainvoke cancelled for {file_or_url}")
+                raise
+            except asyncio.TimeoutError as e:
+                # Treat like transient below
+                last_err = e
             except Exception as e:
-                # Classify timeout + common transient failures as retryable
-                is_transient = self._is_transient_error(e, retry_http_statuses)
-                if not is_transient:
+                if not self._is_transient_error(e, retry_http_statuses):
                     raise
-
-                attempt += 1
-                if attempt > max_retries:
-                    raise
-
-                cap = min(max_delay, base_delay * (2 ** (attempt - 1)))
-                sleep_for = random.uniform(0, cap)
-                logger.warning(
-                    f"ainvoke transient error {file_or_url}: (attempt {attempt}/{max_retries}): {type(e).__name__}: {e}. "
-                    f"Retrying in {sleep_for:.2f}s"
-                )
-                await asyncio.sleep(sleep_for)
+                last_err = e
+            attempt += 1
+            if attempt > max_retries:
+                raise last_err
+            cap = min(max_delay, base_delay * (2 ** (attempt - 1)))
+            sleep_for = random.uniform(0, cap)
+            logger.warning(
+                f"ainvoke transient error {file_or_url}: attempt {attempt}/{max_retries}: "
+                f"{type(last_err).__name__}: {last_err}. Retrying in {sleep_for:.2f}s"
+            )
+            await asyncio.sleep(sleep_for)
 
     @staticmethod
     def _is_transient_error(e: Exception, retry_http_statuses: Tuple[int, ...]) -> bool:
@@ -443,7 +436,7 @@ class GPTVision:
                                 )
                             except asyncio.TimeoutError:
                                 logger.error(
-                                    f"ocr_page: Page {page_index} timed out after {self.page_timeout_s}s"
+                                    f"ocr_page: {file_stem} - page {page_index} timed out after {self.page_timeout_s}s"
                                 )
                                 partial_md = "[Timed out]"
 
@@ -455,7 +448,7 @@ class GPTVision:
                                 logger.warning(
                                     f"ocr_page: {file_stem} - page {page_index}: no OCR result"
                                 )
-                                partial_md = "[No OCR result]"
+                                partial_md = ""
                             return (page_index, f"## Page {page_index + 1}\n\n{partial_md}")
                         finally:
                             await self._safe_unlink(png_path)

{markitdown_pro-1.0.2 → markitdown_pro-1.0.4}/markitdown_pro.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: markitdown-pro
-Version: 1.0.2
+Version: 1.0.4
 Summary: A package that converts almost any file format to Markdown.
 Author: Developer
 Classifier: Programming Language :: Python :: 3

{markitdown_pro-1.0.2 → markitdown_pro-1.0.4}/setup.py

@@ -2,7 +2,7 @@ from setuptools import find_packages, setup
 
 setup(
     name="markitdown-pro",
-    version="1.0.2",
+    version="1.0.4",
     author="Developer",
     description="A package that converts almost any file format to Markdown.",
     long_description=open("README.md").read(),

markitdown_pro-1.0.2/markitdown_pro/common/logger.py

@@ -1,20 +0,0 @@
-import logging
-import os
-
-log_level = int(os.getenv("LOG_LEVEL", "20"))
-logger = logging.getLogger(__name__)
-logging.basicConfig(level=log_level, format="[%(asctime)s] [%(levelname)s] %(message)s")
-
-# Quiet the Azure SDK HTTP logging
-logging.getLogger("azure.core.pipeline.policies.http_logging_policy").setLevel(logging.WARNING)
-logging.getLogger("azure.core.pipeline.policies._universal").setLevel(logging.WARNING)
-
-# 2) Azure Monitor exporter "Transmission succeeded..." spam
-logging.getLogger("azure.monitor.opentelemetry.exporter").setLevel(logging.WARNING)
-logging.getLogger("azure.monitor.opentelemetry.exporter._transmission").setLevel(logging.WARNING)
-
-# 3) Quiet the rest of the Azure SDK + OTel + urllib3 connection noise
-logging.getLogger("azure").setLevel(logging.WARNING)
-logging.getLogger("opentelemetry").setLevel(logging.WARNING)
-logging.getLogger("urllib3").setLevel(logging.WARNING)
-logging.getLogger("urllib3.connectionpool").setLevel(logging.WARNING)