markitdown-pro 1.3.3__tar.gz → 1.3.5__tar.gz

{markitdown_pro-1.3.3 → markitdown_pro-1.3.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: markitdown-pro
-Version: 1.3.3
+Version: 1.3.5
 Summary: A package that converts almost any file format to Markdown.
 Author: Developer
 Classifier: Programming Language :: Python :: 3

markitdown_pro-1.3.5/markitdown_pro/common/schemas.py

@@ -0,0 +1,20 @@
+import enum
+from io import BytesIO
+
+
+class NamedBytesIO(BytesIO):
+    def __init__(self, data: bytes, name: str):
+        super().__init__(data)
+        self.name = name
+
+
+class ExtensionCategory(enum.Enum):
+    IMAGE = "image"
+    AUDIO = "audio"
+    VIDEO = "video"
+    PDF = "pdf"
+    OFFICE = "office"
+    TABULAR = "tabular"
+    LIGHTWEIGHT_TEXT = "lightweight_text"
+    FORMATTED_DOCS = "formatted_docs"
+    OTHER = "other"

{markitdown_pro-1.3.3 → markitdown_pro-1.3.5}/markitdown_pro/conversion_pipeline.py

@@ -38,7 +38,9 @@ from typing import Optional
 import httpx
 
 from .common.logger import logger
+from .common.schemas import ExtensionCategory
 from .common.utils import clean_markdown, detect_extension, ensure_minimum_content
+from .converters.azure_doc_intel_wrapper import DocIntelligenceWrapper
 from .handlers.audio_handler import AudioHandler
 from .handlers.base_handler import BaseHandler
 from .handlers.email_handler import EmailHandler
@@ -88,80 +90,82 @@ class ConversionPipeline:
         self.pst_handler = PSTHandler()
         self.ipynb_handler = IpynbHandler()
         self.markitdown_handler = MarkItDownHandler()
-
-        # Map extensions to handlers (normalized lowercase with dot)
-        self.handlers_mapping: dict[str, BaseHandler] = {
-            # region PDFs
-            ".pdf": self.pdf_handler,
-            # endregion
-            # region Audio
-            ".mp3": self.audio_handler,
-            ".wav": self.audio_handler,
-            # ".ogg": self.audio_handler,
-            # ".flac": self.audio_handler,
-            # ".m4a": self.audio_handler,
-            # ".aac": self.audio_handler,
-            # ".wma": self.audio_handler,
-            # ".webm": self.audio_handler,
-            # ".opus": self.audio_handler,
-            # endregion
-            # region Images
-            ".bmp": self.image_handler,
-            ".gif": self.image_handler,
-            ".heic": self.image_handler,
-            ".jpeg": self.image_handler,
-            ".jpg": self.image_handler,
-            ".png": self.image_handler,
-            # ".prn": self.image_handler,
-            ".svg": self.image_handler,
-            ".tiff": self.image_handler,
-            ".webp": self.image_handler,
-            ".heif": self.image_handler,
-            # endregion
-            # region Plain/code text
-            ".txt": self.text_handler,
-            ".md": self.text_handler,
-            ".py": self.text_handler,
-            ".go": self.text_handler,
-            # endregion
-            # region Tabular (via MarkItDown)
-            ".csv": self.tabular_handler,
-            ".tsv": self.tabular_handler,
-            # endregion
-            # region Office/Spreadsheet/Slides
-            ".xls": self.office_handler,
-            ".xlsx": self.office_handler,
-            ".docx": self.office_handler,
-            # ".odt": self.office_handler,
-            # ".rtf": self.office_handler,
-            # ".ppt": self.office_handler,
-            ".pptx": self.office_handler,
-            # endregion
-            # region Markup/serialization
-            ".html": self.markup_handler,
-            ".htm": self.markup_handler,
-            ".xml": self.markup_handler,
-            ".json": self.markup_handler,
-            ".ndjson": self.markup_handler,
-            ".yaml": self.markup_handler,
-            ".yml": self.markup_handler,
-            # endregion
-            # region Others
-            ".epub": self.epub_handler,
-            ".eml": self.email_handler,
-            ".p7s": self.email_handler,
-            ".msg": self.email_handler,
-            ".pst": self.pst_handler,
-            ".ipynb": self.ipynb_handler,
-            # endregion
+        self.doc_intel_wrapper = DocIntelligenceWrapper()
+
+        self.handlers_categories: dict[ExtensionCategory, dict[str, BaseHandler]] = {
+            ExtensionCategory.IMAGE: {
+                ".bmp": self.image_handler,
+                ".gif": self.image_handler,
+                # ".heic": self.image_handler,
+                ".jpeg": self.image_handler,
+                ".jpg": self.image_handler,
+                ".png": self.image_handler,
+                # ".prn": self.image_handler,
+                ".svg": self.image_handler,
+                ".tiff": self.image_handler,
+                ".webp": self.image_handler,
+                ".heif": self.image_handler,
+            },
+            ExtensionCategory.AUDIO: {
+                ".mp3": self.audio_handler,
+                ".wav": self.audio_handler,
+                # ".ogg": self.audio_handler,
+                # ".flac": self.audio_handler,
+                # ".m4a": self.audio_handler,
+                # ".aac": self.audio_handler,
+                # ".wma": self.audio_handler,
+                # ".webm": self.audio_handler,
+                # ".opus": self.audio_handler,
+            },
+            ExtensionCategory.LIGHTWEIGHT_TEXT: {
+                ".txt": self.text_handler,
+                ".md": self.text_handler,
+                ".py": self.text_handler,
+                ".go": self.text_handler,
+                ".csv": self.tabular_handler,
+                ".tsv": self.tabular_handler,
+                ".html": self.markup_handler,
+                ".htm": self.markup_handler,
+                ".xml": self.markup_handler,
+                ".json": self.markup_handler,
+                ".ndjson": self.markup_handler,
+                ".yaml": self.markup_handler,
+                ".yml": self.markup_handler,
+                ".epub": self.epub_handler,
+                ".eml": self.email_handler,
+                ".p7s": self.email_handler,
+                ".msg": self.email_handler,
+                ".pst": self.pst_handler,
+                ".ipynb": self.ipynb_handler,
+            },
+            ExtensionCategory.FORMATTED_DOCS: {
+                ".pdf": self.pdf_handler,
+                ".xls": self.tabular_handler,
+                ".xlsx": self.tabular_handler,
+                ".docx": self.office_handler,
+                # ".odt": self.office_handler,
+                # ".rtf": self.office_handler,
+                # ".ppt": self.office_handler,
+                ".pptx": self.office_handler,
+            },
         }
 
+        # Normalize map of extensions to handlers
+        self.handlers_mapping: dict[str, BaseHandler] = {}
+        for handlers in self.handlers_categories.values():
+            for ext, handler in handlers.items():
+                self.handlers_mapping[ext] = handler
+
     # ---------------------------------------------------------------------
     # Public APIs
     # ---------------------------------------------------------------------
 
     async def convert_document_to_md(
-        self, file_path: str | Path, output_md: Optional[str | Path] = None
+        self,
+        file_path: str | Path,
+        output_md: Optional[str | Path] = None,
+        *args,
+        **kwargs,
     ) -> Optional[str]:
         """
         Convert a **local file** to Markdown.
@@ -193,7 +197,7 @@ class ConversionPipeline:
             raise RuntimeError(f"No handler for extension '{extension}'.")
 
         try:
-            md_content = await handler.handle(str(file_path))
+            md_content = await handler.handle(str(file_path), *args, **kwargs)
             if not md_content or not ensure_minimum_content(md_content):
                 return None
 
@@ -281,3 +285,30 @@ class ConversionPipeline:
             return await self.convert_document_to_md(local_path, output_md=output_md)
         finally:
             local_path.unlink(missing_ok=True)
+
+    async def get_page_count(self, file_path: str | Path) -> Optional[int]:
+        """
+        Get the page count of a document if supported by its handler.
+
+        Parameters
+        ----------
+        file_path : str | Path
+            Path to the local file.
+        """
+        file_path = Path(file_path)
+        if not file_path.is_file():
+            raise ValueError(f"The provided path '{file_path}' is not a valid file.")
+
+        # handle file extensions individually to avoid using the wrong handlers
+        extension = detect_extension(str(file_path)).lower()
+
+        if extension in self.image_handler.SUPPORTED_EXTENSIONS:
+            return await self.image_handler.get_page_count(file_path)
+        elif extension in self.pdf_handler.SUPPORTED_EXTENSIONS:
+            return await self.doc_intel_wrapper.get_page_count(file_path)
+        # elif extension in self.doc_intel_wrapper.SUPPORTED_EXTENSIONS:
+        #     return await self.doc_intel_wrapper.get_page_count(file_path)
+        else:
+            logger.warning(f"ConversionPipeline: No page count support for extension '{extension}'")
+            return None
+            return None

{markitdown_pro-1.3.3 → markitdown_pro-1.3.5}/markitdown_pro/converters/azure_doc_intel_wrapper.py

@@ -1,3 +1,4 @@
+from pathlib import Path
 from typing import Optional
 
 from ..services.azure_doc_intelligence import DocumentIntelligenceHandler
@@ -16,3 +17,6 @@ class DocIntelligenceWrapper(ConverterWrapper):
         markdown = await self.converter.convert_to_md(file_path)
         await self.converter.aclose()
         return markdown
+
+    async def get_page_count(self, file_path: str | Path) -> Optional[int]:
+        return await self.converter.get_page_count(file_path)

{markitdown_pro-1.3.3 → markitdown_pro-1.3.5}/markitdown_pro/converters/markitdown_wrapper.py

@@ -3,6 +3,7 @@ from typing import Optional
 from markitdown_pro.handlers.markitdown_handler import MarkItDownHandler
 
 from ..common.logger import logger
+from ..common.utils import detect_extension
 from .base import ConverterWrapper
 
 
@@ -14,7 +15,7 @@ class MarkItDownWrapper(ConverterWrapper):
         self.handler = MarkItDownHandler(*args, **kwargs)
 
     async def convert(self, file_path: str) -> Optional[str]:
-        file_extension = f".{file_path.split('.')[-1].lower()}"
+        file_extension = detect_extension(file_path)
         if file_extension not in self.SUPPORTED_EXTENSIONS:
             logger.warning(f"MarkItDownWrapper: Unsupported file format: {file_extension}")
             return None

{markitdown_pro-1.3.3 → markitdown_pro-1.3.5}/markitdown_pro/converters/pymupdf_wrapper.py

@@ -3,6 +3,7 @@ from typing import Optional
 import fitz  # PyMuPDF
 
 from ..common.logger import logger
+from ..common.utils import detect_extension
 from .base import ConverterWrapper
 
 
@@ -13,11 +14,7 @@ class PyMuPDFWrapper(ConverterWrapper):
         super().__init__("PyMuPDF")
 
     async def convert(self, file_path: str) -> Optional[str]:
-        file_extension = file_path.split(".")[-1].lower()
-        logger.info(
-            f"PyMuPDFWrapper: Processing file: {file_path} with extension: {file_extension}"
-        )
-
+        file_extension = detect_extension(file_path)
         if file_extension not in self.SUPPORTED_EXTENSIONS:
             logger.warning(f"PyMuPDFWrapper: Unsupported file format: {file_extension}")
             return None

{markitdown_pro-1.3.3 → markitdown_pro-1.3.5}/markitdown_pro/handlers/audio_handler.py

@@ -3,9 +3,9 @@ from __future__ import annotations
 from typing import Optional
 
 from ..common.logger import logger
+from ..common.utils import ensure_minimum_content
 from ..converters.azure_speech_wrapper import AzureSpeechWrapper
 from ..converters.base import ConverterWrapper
-from ..converters.markitdown_wrapper import MarkItDownWrapper
 from ..handlers.base_handler import BaseHandler
 
 
@@ -13,8 +13,7 @@ class AudioHandler(BaseHandler):
     """
     Convert audio files to Markdown by trying two converters in order:
 
-      1) **MarkItDown** — fast local extractor (works when the file format is supported).
-      2) **Azure Speech** — cloud speech-to-text fallback for actual audio transcription.
+      1) **Azure Speech** — cloud speech-to-text fallback for actual audio transcription.
 
     The first converter that yields non-trivial Markdown (as determined by
     `ensure_minimum_content`) is returned. If both fail, `None` is returned.
@@ -26,27 +25,13 @@ class AudioHandler(BaseHandler):
     - This handler runs converters **sequentially** to avoid unnecessary API calls.
     """
 
-    SUPPORTED_EXTENSIONS = frozenset(
-        {
-            ".mp3",
-            ".wav",
-            ".ogg",
-            ".flac",
-            ".m4a",
-            ".aac",
-            ".wma",
-            ".webm",
-            ".opus",
-        }
-    )
+    SUPPORTED_EXTENSIONS = AzureSpeechWrapper.SUPPORTED_EXTENSIONS
 
     def __init__(self, *args, **kwargs) -> None:
         super().__init__(*args, **kwargs)
-        self.markitdown = MarkItDownWrapper()
         self.azure_speech = AzureSpeechWrapper()
 
         self._pipeline: list[tuple[ConverterWrapper, str]] = [
-            (self.markitdown, "MarkItDown"),
             (self.azure_speech, "Azure Speech"),
         ]
 
@@ -71,7 +56,8 @@ class AudioHandler(BaseHandler):
             logger.info(f"AudioHandler: Trying {name} for '{file_path}'")
             try:
                 md = await converter.convert(file_path)
-                return md
+                if md and ensure_minimum_content(md):
+                    return md
             except Exception as e:
                 logger.error(f"AudioHandler: {name} failed for '{file_path}': {e}")
 

{markitdown_pro-1.3.3 → markitdown_pro-1.3.5}/markitdown_pro/handlers/image_handler.py

@@ -1,7 +1,9 @@
 import contextlib
+from pathlib import Path
 from typing import Optional
 
 from ..common.logger import logger
+from ..common.utils import detect_extension
 from ..converters.gpt_vision_wrapper import GPTVisionWrapper
 from .base_handler import BaseHandler
 
@@ -37,6 +39,27 @@ class ImageHandler(BaseHandler):
                 if hasattr(self.gpt_vision, "gpt_vision"):
                     await self.gpt_vision.gpt_vision.aclose()
 
+    async def get_page_count(self, file_path: str | Path) -> Optional[int]:
+        """
+        Determines the number of pages in the given image file.
+        """
+        p = Path(file_path)
+        ext = detect_extension(str(p.absolute()))
+        if ext not in self.SUPPORTED_EXTENSIONS:
+            logger.warning(f"ImageHandler: Unsupported file format: {ext}")
+            return None
+
+        if ext in {".tif", ".tiff"}:
+            try:
+                from PIL import Image, ImageSequence  # optional dependency
+
+                with Image.open(str(p)) as im:
+                    return sum(1 for _ in ImageSequence.Iterator(im)) or 1
+            except Exception as e_tiff:
+                logger.warning(f"ImageHandler: local TIFF page count failed for '{p}': {e_tiff}")
+        else:
+            return 1
+
     async def aclose(self) -> None:
         if hasattr(self, "gpt_vision") and self.gpt_vision:
             await self.gpt_vision.aclose()

{markitdown_pro-1.3.3 → markitdown_pro-1.3.5}/markitdown_pro/handlers/office_handler.py

@@ -4,7 +4,6 @@ from ..common.logger import logger
 from ..converters.azure_doc_intel_wrapper import DocIntelligenceWrapper
 from ..converters.base import ConverterWrapper
 from ..converters.markitdown_wrapper import MarkItDownWrapper
-from ..converters.tabular_wrapper import TabularWrapper
 from ..converters.unstructured_wrapper import UnstructuredWrapper
 from .base_handler import BaseHandler
 
@@ -13,10 +12,9 @@ class OfficeHandler(BaseHandler):
     """
     Convert Office documents to Markdown by trying a series of converters in order:
 
-      1) TabularHandler (specialized handler for tabular data)
-      2) MarkItDown (fast, local parsing for Office formats)
-      3) Azure Document Intelligence (cloud extraction + OCR for embedded images)
-      4) Unstructured (robust fallback parser)
+      1) MarkItDown (fast, local parsing for Office formats)
+      2) Azure Document Intelligence (cloud extraction + OCR for embedded images)
+      3) Unstructured (robust fallback parser)
 
     The first converter that returns non-trivial Markdown (as judged by
     `ensure_minimum_content`) wins. If all fail, `None` is returned.
@@ -28,21 +26,18 @@ class OfficeHandler(BaseHandler):
       avoid unnecessary API calls when earlier options succeed.
     """
 
-    # Supported Office-like extensions (lowercase with dot)
     SUPPORTED_EXTENSIONS = frozenset({".doc", ".docx", ".odt", ".rtf", ".ppt", ".pptx"})
 
     def __init__(self, *args, **kwargs) -> None:
         super().__init__(*args, **kwargs)
 
         # Ordered, fastest → most comprehensive
-        self.tabular = TabularWrapper()
         self.markitdown = MarkItDownWrapper()
         self.doc_intelligence = DocIntelligenceWrapper()
         self.unstructured = UnstructuredWrapper()
 
         # (converter, human_readable_name) in the exact order to try
         self._pipeline: list[tuple[ConverterWrapper, str]] = [
-            (self.tabular, "TabularHandler"),
             (self.markitdown, "MarkItDown"),
             (self.doc_intelligence, "Azure Document Intelligence"),
             (self.unstructured, "Unstructured"),