diff --git a/README.md b/README.md index c7dd5f3..b807a6a 100644 --- a/README.md +++ b/README.md @@ -4,6 +4,9 @@ ![PyPI - Downloads](https://img.shields.io/pypi/dd/markitdown) [![Built by AutoGen Team](https://img.shields.io/badge/Built%20by-AutoGen%20Team-blue)](https://github.com/microsoft/autogen) +> [!TIP] +> MarkItDown now offers an MCP (Model Context Protocol) server for integration with LLM applications like Claude Desktop. See [markitdown-mcp](https://github.com/microsoft/markitdown/tree/main/packages/markitdown-mcp) for more information. + > [!IMPORTANT] > Breaking changes between 0.0.1 to 0.1.0: > * Dependencies are now organized into optional feature-groups (further details below). Use `pip install 'markitdown[all]'` to have backward-compatible behavior. diff --git a/packages/markitdown-mcp/Dockerfile b/packages/markitdown-mcp/Dockerfile new file mode 100644 index 0000000..fe52a4b --- /dev/null +++ b/packages/markitdown-mcp/Dockerfile @@ -0,0 +1,26 @@ +FROM python:3.13-slim-bullseye + +ENV DEBIAN_FRONTEND=noninteractive +ENV EXIFTOOL_PATH=/usr/bin/exiftool +ENV FFMPEG_PATH=/usr/bin/ffmpeg + +# Runtime dependency +RUN apt-get update && apt-get install -y --no-install-recommends \ + ffmpeg \ + exiftool + +# Cleanup +RUN rm -rf /var/lib/apt/lists/* + +COPY . /app +RUN pip --no-cache-dir install /app + +WORKDIR /workdir + +# Default USERID and GROUPID +ARG USERID=nobody +ARG GROUPID=nogroup + +USER $USERID:$GROUPID + +ENTRYPOINT [ "markitdown-mcp" ] diff --git a/packages/markitdown-mcp/README.md b/packages/markitdown-mcp/README.md new file mode 100644 index 0000000..923a42f --- /dev/null +++ b/packages/markitdown-mcp/README.md @@ -0,0 +1,134 @@ +# MarkItDown-MCP + +[![PyPI](https://img.shields.io/pypi/v/markitdown-mcp.svg)](https://pypi.org/project/markitdown-mcp/) +![PyPI - Downloads](https://img.shields.io/pypi/dd/markitdown-mcp) +[![Built by AutoGen Team](https://img.shields.io/badge/Built%20by-AutoGen%20Team-blue)](https://github.com/microsoft/autogen) + +The `markitdown-mcp` package provides a lightweight STDIO and SSE MCP server for calling MarkItDown. + +It exposes one tool: `convert_to_markdown(uri)`, where uri can be any `http:`, `https:`, `file:`, or `data:` URI. + +## Installation + +To install the package, use pip: + +```bash +pip install markitdown-mcp +``` + +## Usage + +To run the MCP server, ussing STDIO (default) use the following command: + + +```bash +markitdown-mcp +``` + +To run the MCP server, ussing SSE use the following command: + +```bash +markitdown-mcp --sse --host 127.0.0.1 --port 3001 +``` + +## Running in Docker + +To run `markitdown-mcp` in Docker, build the Docker image using the provided Dockerfile: +```bash +docker build -t markitdown-mcp:latest . +``` + +And run it using: +```bash +docker run -it --rm markitdown-mcp:latest +``` +This will be sufficient for remote URIs. To access local files, you need to mount the local directory into the container. For example, if you want to access files in `/home/user/data`, you can run: + +```bash +docker run -it --rm -v /home/user/data:/workdir markitdown-mcp:latest +``` + +Once mounted, all files under data will be accessible under `/workdir` in the container. For example, if you have a file `example.txt` in `/home/user/data`, it will be accessible in the container at `/workdir/example.txt`. + +## Accessing from Claude Desktop + +It is recommended to use the Docker image when running the MCP server for Claude Desktop. + +Follow [these instrutions](https://modelcontextprotocol.io/quickstart/user#for-claude-desktop-users) to access Claude's `claude_desktop_config.json` file. + +Edit it to include the following JSON entry: + +```json +{ + "mcpServers": { + "markitdown": { + "command": "docker", + "args": [ + "run", + "--rm", + "-i", + "markitdown-mcp:latest" + ] + } + } +} +``` + +If you want to mount a directory, adjust it accordingly: + +```json +{ + "mcpServers": { + "markitdown": { + "command": "docker", + "args": [ + "run", + "--rm", + "-i", + "-v", + "/home/user/data:/workdir", + "markitdown-mcp:latest" + ] + } + } +} +``` + +## Debugging + +To debug the MCP server you can use the `mcpinspector` tool. + +```bash +npx @modelcontextprotocol/inspector +``` + +You can then connect to the insepctor through the specified host and port (e.g., `http://localhost:5173/`). + +If using STDIO: +* select `STDIO` as the transport type, +* input `markitdown-mcp` as the command, and +* click `Connect` + +If using SSE: +* select `SSE` as the transport type, +* input `http://127.0.0.1:3001/sse` as the URL, and +* click `Connect` + +Finally: +* click the `Tools` tab, +* click `List Tools`, +* click `convert_to_markdown`, and +* run the tool on any valid URI. + +## Security Considerations + +The server does not support authentication, and runs with the privileges if the user running it. For this reason, when running in SSE mode, it is recommended to run the server bound to `localhost` (default). + + +## Trademarks + +This project may contain trademarks or logos for projects, products, or services. Authorized use of Microsoft +trademarks or logos is subject to and must follow +[Microsoft's Trademark & Brand Guidelines](https://www.microsoft.com/en-us/legal/intellectualproperty/trademarks/usage/general). +Use of Microsoft trademarks or logos in modified versions of this project must not cause confusion or imply Microsoft sponsorship. +Any use of third-party trademarks or logos are subject to those third-party's policies. diff --git a/packages/markitdown-mcp/pyproject.toml b/packages/markitdown-mcp/pyproject.toml new file mode 100644 index 0000000..6cbc0e5 --- /dev/null +++ b/packages/markitdown-mcp/pyproject.toml @@ -0,0 +1,69 @@ +[build-system] +requires = ["hatchling"] +build-backend = "hatchling.build" + +[project] +name = "markitdown-mcp" +dynamic = ["version"] +description = 'An MCP server for the "markitdown" library.' +readme = "README.md" +requires-python = ">=3.10" +license = "MIT" +keywords = [] +authors = [ + { name = "Adam Fourney", email = "adamfo@microsoft.com" }, +] +classifiers = [ + "Development Status :: 4 - Beta", + "Programming Language :: Python", + "Programming Language :: Python :: 3.10", + "Programming Language :: Python :: 3.11", + "Programming Language :: Python :: 3.12", + "Programming Language :: Python :: 3.13", + "Programming Language :: Python :: Implementation :: CPython", + "Programming Language :: Python :: Implementation :: PyPy", +] +dependencies = [ + "mcp~=1.5.0", + "markitdown[all]>=0.1.1,<0.2.0", +] + +[project.urls] +Documentation = "https://github.com/microsoft/markitdown#readme" +Issues = "https://github.com/microsoft/markitdown/issues" +Source = "https://github.com/microsoft/markitdown" + +[tool.hatch.version] +path = "src/markitdown_mcp/__about__.py" + +[project.scripts] +markitdown-mcp = "markitdown_mcp.__main__:main" + +[tool.hatch.envs.types] +extra-dependencies = [ + "mypy>=1.0.0", +] +[tool.hatch.envs.types.scripts] +check = "mypy --install-types --non-interactive {args:src/markitdown_mcp tests}" + +[tool.coverage.run] +source_pkgs = ["markitdown-mcp", "tests"] +branch = true +parallel = true +omit = [ + "src/markitdown_mcp/__about__.py", +] + +[tool.coverage.paths] +markitdown-mcp = ["src/markitdown_mcp", "*/markitdown-mcp/src/markitdown_mcp"] +tests = ["tests", "*/markitdown-mcp/tests"] + +[tool.coverage.report] +exclude_lines = [ + "no cov", + "if __name__ == .__main__.:", + "if TYPE_CHECKING:", +] + +[tool.hatch.build.targets.sdist] +only-include = ["src/markitdown_mcp"] diff --git a/packages/markitdown-mcp/src/markitdown_mcp/__about__.py b/packages/markitdown-mcp/src/markitdown_mcp/__about__.py new file mode 100644 index 0000000..a365900 --- /dev/null +++ b/packages/markitdown-mcp/src/markitdown_mcp/__about__.py @@ -0,0 +1,4 @@ +# SPDX-FileCopyrightText: 2024-present Adam Fourney +# +# SPDX-License-Identifier: MIT +__version__ = "0.0.1a3" diff --git a/packages/markitdown-mcp/src/markitdown_mcp/__init__.py b/packages/markitdown-mcp/src/markitdown_mcp/__init__.py new file mode 100644 index 0000000..8464d59 --- /dev/null +++ b/packages/markitdown-mcp/src/markitdown_mcp/__init__.py @@ -0,0 +1,9 @@ +# SPDX-FileCopyrightText: 2024-present Adam Fourney +# +# SPDX-License-Identifier: MIT + +from .__about__ import __version__ + +__all__ = [ + "__version__", +] diff --git a/packages/markitdown-mcp/src/markitdown_mcp/__main__.py b/packages/markitdown-mcp/src/markitdown_mcp/__main__.py new file mode 100644 index 0000000..32b7527 --- /dev/null +++ b/packages/markitdown-mcp/src/markitdown_mcp/__main__.py @@ -0,0 +1,83 @@ +import sys +from typing import Any +from mcp.server.fastmcp import FastMCP +from starlette.applications import Starlette +from mcp.server.sse import SseServerTransport +from starlette.requests import Request +from starlette.routing import Mount, Route +from mcp.server import Server +from markitdown import MarkItDown +import uvicorn + +# Initialize FastMCP server for MarkItDown (SSE) +mcp = FastMCP("markitdown") + + +@mcp.tool() +async def convert_to_markdown(uri: str) -> str: + """Convert a resource described by an http:, https:, file: or data: URI to markdown""" + return MarkItDown().convert_uri(uri).markdown + + +def create_starlette_app(mcp_server: Server, *, debug: bool = False) -> Starlette: + sse = SseServerTransport("/messages/") + + async def handle_sse(request: Request) -> None: + async with sse.connect_sse( + request.scope, + request.receive, + request._send, + ) as (read_stream, write_stream): + await mcp_server.run( + read_stream, + write_stream, + mcp_server.create_initialization_options(), + ) + + return Starlette( + debug=debug, + routes=[ + Route("/sse", endpoint=handle_sse), + Mount("/messages/", app=sse.handle_post_message), + ], + ) + + +# Main entry point +def main(): + import argparse + + mcp_server = mcp._mcp_server + + parser = argparse.ArgumentParser(description="Run MCP SSE-based MarkItDown server") + + parser.add_argument( + "--sse", + action="store_true", + help="Run the server with SSE transport rather than STDIO (default: False)", + ) + parser.add_argument( + "--host", default=None, help="Host to bind to (default: 127.0.0.1)" + ) + parser.add_argument( + "--port", type=int, default=None, help="Port to listen on (default: 3001)" + ) + args = parser.parse_args() + + if not args.sse and (args.host or args.port): + parser.error("Host and port arguments are only valid when using SSE transport.") + sys.exit(1) + + if args.sse: + starlette_app = create_starlette_app(mcp_server, debug=True) + uvicorn.run( + starlette_app, + host=args.host if args.host else "127.0.0.1", + port=args.port if args.port else 3001, + ) + else: + mcp.run() + + +if __name__ == "__main__": + main() diff --git a/packages/markitdown-mcp/src/markitdown_mcp/py.typed b/packages/markitdown-mcp/src/markitdown_mcp/py.typed new file mode 100644 index 0000000..e69de29 diff --git a/packages/markitdown-mcp/tests/__init__.py b/packages/markitdown-mcp/tests/__init__.py new file mode 100644 index 0000000..44af7d7 --- /dev/null +++ b/packages/markitdown-mcp/tests/__init__.py @@ -0,0 +1,3 @@ +# SPDX-FileCopyrightText: 2024-present Adam Fourney +# +# SPDX-License-Identifier: MIT diff --git a/packages/markitdown-sample-plugin/README.md b/packages/markitdown-sample-plugin/README.md index fd7115f..adf1d9e 100644 --- a/packages/markitdown-sample-plugin/README.md +++ b/packages/markitdown-sample-plugin/README.md @@ -1,7 +1,7 @@ # MarkItDown Sample Plugin -[![PyPI](https://img.shields.io/pypi/v/markitdown.svg)](https://pypi.org/project/markitdown/) -![PyPI - Downloads](https://img.shields.io/pypi/dd/markitdown) +[![PyPI](https://img.shields.io/pypi/v/markitdown-sample-plugin.svg)](https://pypi.org/project/markitdown-sample-plugin/) +![PyPI - Downloads](https://img.shields.io/pypi/dd/markitdown-sample-plugin) [![Built by AutoGen Team](https://img.shields.io/badge/Built%20by-AutoGen%20Team-blue)](https://github.com/microsoft/autogen) diff --git a/packages/markitdown/src/markitdown/__about__.py b/packages/markitdown/src/markitdown/__about__.py index e8d7ba6..159c6da 100644 --- a/packages/markitdown/src/markitdown/__about__.py +++ b/packages/markitdown/src/markitdown/__about__.py @@ -1,4 +1,4 @@ # SPDX-FileCopyrightText: 2024-present Adam Fourney # # SPDX-License-Identifier: MIT -__version__ = "0.1.0" +__version__ = "0.1.1" diff --git a/packages/markitdown/src/markitdown/_markitdown.py b/packages/markitdown/src/markitdown/_markitdown.py index a8f7c9e..54a0dc8 100644 --- a/packages/markitdown/src/markitdown/_markitdown.py +++ b/packages/markitdown/src/markitdown/_markitdown.py @@ -10,7 +10,7 @@ import traceback import io from dataclasses import dataclass from importlib.metadata import entry_points -from typing import Any, List, Optional, Union, BinaryIO +from typing import Any, List, Dict, Optional, Union, BinaryIO from pathlib import Path from urllib.parse import urlparse from warnings import warn @@ -20,6 +20,7 @@ import charset_normalizer import codecs from ._stream_info import StreamInfo +from ._uri_utils import parse_data_uri, file_uri_to_path from .converters import ( PlainTextConverter, @@ -197,8 +198,19 @@ class MarkItDown: # Register Document Intelligence converter at the top of the stack if endpoint is provided docintel_endpoint = kwargs.get("docintel_endpoint") if docintel_endpoint is not None: + docintel_args: Dict[str, Any] = {} + docintel_args["endpoint"] = docintel_endpoint + + docintel_credential = kwargs.get("docintel_credential") + if docintel_credential is not None: + docintel_args["credential"] = docintel_credential + + docintel_types = kwargs.get("docintel_file_types") + if docintel_types is not None: + docintel_args["file_types"] = docintel_types + self.register_converter( - DocumentIntelligenceConverter(endpoint=docintel_endpoint) + DocumentIntelligenceConverter(**docintel_args), ) self._builtins_enabled = True @@ -242,9 +254,10 @@ class MarkItDown: # Local path or url if isinstance(source, str): if ( - source.startswith("http://") - or source.startswith("https://") - or source.startswith("file://") + source.startswith("http:") + or source.startswith("https:") + or source.startswith("file:") + or source.startswith("data:") ): # Rename the url argument to mock_url # (Deprecated -- use stream_info) @@ -253,7 +266,7 @@ class MarkItDown: _kwargs["mock_url"] = _kwargs["url"] del _kwargs["url"] - return self.convert_url(source, stream_info=stream_info, **_kwargs) + return self.convert_uri(source, stream_info=stream_info, **_kwargs) else: return self.convert_local(source, stream_info=stream_info, **kwargs) # Path object @@ -363,22 +376,80 @@ class MarkItDown: url: str, *, stream_info: Optional[StreamInfo] = None, + file_extension: Optional[str] = None, + mock_url: Optional[str] = None, + **kwargs: Any, + ) -> DocumentConverterResult: + """Alias for convert_uri()""" + # convert_url will likely be deprecated in the future in favor of convert_uri + return self.convert_uri( + url, + stream_info=stream_info, + file_extension=file_extension, + mock_url=mock_url, + **kwargs, + ) + + def convert_uri( + self, + uri: str, + *, + stream_info: Optional[StreamInfo] = None, file_extension: Optional[str] = None, # Deprecated -- use stream_info mock_url: Optional[ str ] = None, # Mock the request as if it came from a different URL **kwargs: Any, - ) -> DocumentConverterResult: # TODO: fix kwargs type - # Send a HTTP request to the URL - response = self._requests_session.get(url, stream=True) - response.raise_for_status() - return self.convert_response( - response, - stream_info=stream_info, - file_extension=file_extension, - url=mock_url, - **kwargs, - ) + ) -> DocumentConverterResult: + uri = uri.strip() + + # File URIs + if uri.startswith("file:"): + netloc, path = file_uri_to_path(uri) + if netloc and netloc != "localhost": + raise ValueError( + f"Unsupported file URI: {uri}. Netloc must be empty or localhost." + ) + return self.convert_local( + path, + stream_info=stream_info, + file_extension=file_extension, + url=mock_url, + **kwargs, + ) + # Data URIs + elif uri.startswith("data:"): + mimetype, attributes, data = parse_data_uri(uri) + + base_guess = StreamInfo( + mimetype=mimetype, + charset=attributes.get("charset"), + ) + if stream_info is not None: + base_guess = base_guess.copy_and_update(stream_info) + + return self.convert_stream( + io.BytesIO(data), + stream_info=base_guess, + file_extension=file_extension, + url=mock_url, + **kwargs, + ) + # HTTP/HTTPS URIs + elif uri.startswith("http:") or uri.startswith("https:"): + response = self._requests_session.get(uri, stream=True) + response.raise_for_status() + return self.convert_response( + response, + stream_info=stream_info, + file_extension=file_extension, + url=mock_url, + **kwargs, + ) + else: + raise ValueError( + f"Unsupported URI scheme: {uri.split(':')[0]}. Supported schemes are: file:, data:, http:, https:" + ) def convert_response( self, diff --git a/packages/markitdown/src/markitdown/_uri_utils.py b/packages/markitdown/src/markitdown/_uri_utils.py new file mode 100644 index 0000000..603da63 --- /dev/null +++ b/packages/markitdown/src/markitdown/_uri_utils.py @@ -0,0 +1,52 @@ +import base64 +import os +from typing import Tuple, Dict +from urllib.request import url2pathname +from urllib.parse import urlparse, unquote_to_bytes + + +def file_uri_to_path(file_uri: str) -> Tuple[str | None, str]: + """Convert a file URI to a local file path""" + parsed = urlparse(file_uri) + if parsed.scheme != "file": + raise ValueError(f"Not a file URL: {file_uri}") + + netloc = parsed.netloc if parsed.netloc else None + path = os.path.abspath(url2pathname(parsed.path)) + return netloc, path + + +def parse_data_uri(uri: str) -> Tuple[str | None, Dict[str, str], bytes]: + if not uri.startswith("data:"): + raise ValueError("Not a data URI") + + header, _, data = uri.partition(",") + if not _: + raise ValueError("Malformed data URI, missing ',' separator") + + meta = header[5:] # Strip 'data:' + parts = meta.split(";") + + is_base64 = False + # Ends with base64? + if parts[-1] == "base64": + parts.pop() + is_base64 = True + + mime_type = None # Normally this would default to text/plain but we won't assume + if len(parts) and len(parts[0]) > 0: + # First part is the mime type + mime_type = parts.pop(0) + + attributes: Dict[str, str] = {} + for part in parts: + # Handle key=value pairs in the middle + if "=" in part: + key, value = part.split("=", 1) + attributes[key] = value + elif len(part) > 0: + attributes[part] = "" + + content = base64.b64decode(data) if is_base64 else unquote_to_bytes(data) + + return mime_type, attributes, content diff --git a/packages/markitdown/src/markitdown/converters/__init__.py b/packages/markitdown/src/markitdown/converters/__init__.py index 09e3cb1..c68d0c3 100644 --- a/packages/markitdown/src/markitdown/converters/__init__.py +++ b/packages/markitdown/src/markitdown/converters/__init__.py @@ -17,7 +17,10 @@ from ._image_converter import ImageConverter from ._audio_converter import AudioConverter from ._outlook_msg_converter import OutlookMsgConverter from ._zip_converter import ZipConverter -from ._doc_intel_converter import DocumentIntelligenceConverter +from ._doc_intel_converter import ( + DocumentIntelligenceConverter, + DocumentIntelligenceFileType, +) from ._epub_converter import EpubConverter __all__ = [ @@ -38,5 +41,6 @@ __all__ = [ "OutlookMsgConverter", "ZipConverter", "DocumentIntelligenceConverter", + "DocumentIntelligenceFileType", "EpubConverter", ] diff --git a/packages/markitdown/src/markitdown/converters/_doc_intel_converter.py b/packages/markitdown/src/markitdown/converters/_doc_intel_converter.py index 2f116d0..5f4069b 100644 --- a/packages/markitdown/src/markitdown/converters/_doc_intel_converter.py +++ b/packages/markitdown/src/markitdown/converters/_doc_intel_converter.py @@ -1,7 +1,9 @@ import sys import re +import os from typing import BinaryIO, Any, List +from enum import Enum from ._html_converter import HtmlConverter from .._base_converter import DocumentConverter, DocumentConverterResult @@ -18,6 +20,7 @@ try: AnalyzeResult, DocumentAnalysisFeature, ) + from azure.core.credentials import AzureKeyCredential, TokenCredential from azure.identity import DefaultAzureCredential except ImportError: # Preserve the error and stack trace for later @@ -29,38 +32,74 @@ except ImportError: CONTENT_FORMAT = "markdown" -OFFICE_MIME_TYPE_PREFIXES = [ - "application/vnd.openxmlformats-officedocument.wordprocessingml.document", - "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet", - "application/vnd.openxmlformats-officedocument.presentationml", - "application/xhtml", - "text/html", -] +class DocumentIntelligenceFileType(str, Enum): + """Enum of file types supported by the Document Intelligence Converter.""" -OTHER_MIME_TYPE_PREFIXES = [ - "application/pdf", - "application/x-pdf", - "text/html", - "image/", -] + # No OCR + DOCX = "docx" + PPTX = "pptx" + XLSX = "xlsx" + HTML = "html" + # OCR + PDF = "pdf" + JPEG = "jpeg" + PNG = "png" + BMP = "bmp" + TIFF = "tiff" -OFFICE_FILE_EXTENSIONS = [ - ".docx", - ".xlsx", - ".pptx", - ".html", - ".htm", -] -OTHER_FILE_EXTENSIONS = [ - ".pdf", - ".jpeg", - ".jpg", - ".png", - ".bmp", - ".tiff", - ".heif", -] +def _get_mime_type_prefixes(types: List[DocumentIntelligenceFileType]) -> List[str]: + """Get the MIME type prefixes for the given file types.""" + prefixes: List[str] = [] + for type_ in types: + if type_ == DocumentIntelligenceFileType.DOCX: + prefixes.append( + "application/vnd.openxmlformats-officedocument.wordprocessingml.document" + ) + elif type_ == DocumentIntelligenceFileType.PPTX: + prefixes.append( + "application/vnd.openxmlformats-officedocument.presentationml" + ) + elif type_ == DocumentIntelligenceFileType.XLSX: + prefixes.append( + "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet" + ) + elif type_ == DocumentIntelligenceFileType.PDF: + prefixes.append("application/pdf") + prefixes.append("application/x-pdf") + elif type_ == DocumentIntelligenceFileType.JPEG: + prefixes.append("image/jpeg") + elif type_ == DocumentIntelligenceFileType.PNG: + prefixes.append("image/png") + elif type_ == DocumentIntelligenceFileType.BMP: + prefixes.append("image/bmp") + elif type_ == DocumentIntelligenceFileType.TIFF: + prefixes.append("image/tiff") + return prefixes + + +def _get_file_extensions(types: List[DocumentIntelligenceFileType]) -> List[str]: + """Get the file extensions for the given file types.""" + extensions: List[str] = [] + for type_ in types: + if type_ == DocumentIntelligenceFileType.DOCX: + extensions.append(".docx") + elif type_ == DocumentIntelligenceFileType.PPTX: + extensions.append(".pptx") + elif type_ == DocumentIntelligenceFileType.XLSX: + extensions.append(".xlsx") + elif type_ == DocumentIntelligenceFileType.PDF: + extensions.append(".pdf") + elif type_ == DocumentIntelligenceFileType.JPEG: + extensions.append(".jpg") + extensions.append(".jpeg") + elif type_ == DocumentIntelligenceFileType.PNG: + extensions.append(".png") + elif type_ == DocumentIntelligenceFileType.BMP: + extensions.append(".bmp") + elif type_ == DocumentIntelligenceFileType.TIFF: + extensions.append(".tiff") + return extensions class DocumentIntelligenceConverter(DocumentConverter): @@ -71,8 +110,30 @@ class DocumentIntelligenceConverter(DocumentConverter): *, endpoint: str, api_version: str = "2024-07-31-preview", + credential: AzureKeyCredential | TokenCredential | None = None, + file_types: List[DocumentIntelligenceFileType] = [ + DocumentIntelligenceFileType.DOCX, + DocumentIntelligenceFileType.PPTX, + DocumentIntelligenceFileType.XLSX, + DocumentIntelligenceFileType.PDF, + DocumentIntelligenceFileType.JPEG, + DocumentIntelligenceFileType.PNG, + DocumentIntelligenceFileType.BMP, + DocumentIntelligenceFileType.TIFF, + ], ): + """ + Initialize the DocumentIntelligenceConverter. + + Args: + endpoint (str): The endpoint for the Document Intelligence service. + api_version (str): The API version to use. Defaults to "2024-07-31-preview". + credential (AzureKeyCredential | TokenCredential | None): The credential to use for authentication. + file_types (List[DocumentIntelligenceFileType]): The file types to accept. Defaults to all supported file types. + """ + super().__init__() + self._file_types = file_types # Raise an error if the dependencies are not available. # This is different than other converters since this one isn't even instantiated @@ -86,12 +147,18 @@ class DocumentIntelligenceConverter(DocumentConverter): _dependency_exc_info[2] ) + if credential is None: + if os.environ.get("AZURE_API_KEY") is None: + credential = DefaultAzureCredential() + else: + credential = AzureKeyCredential(os.environ["AZURE_API_KEY"]) + self.endpoint = endpoint self.api_version = api_version self.doc_intel_client = DocumentIntelligenceClient( endpoint=self.endpoint, api_version=self.api_version, - credential=DefaultAzureCredential(), + credential=credential, ) def accepts( @@ -103,10 +170,10 @@ class DocumentIntelligenceConverter(DocumentConverter): mimetype = (stream_info.mimetype or "").lower() extension = (stream_info.extension or "").lower() - if extension in OFFICE_FILE_EXTENSIONS + OTHER_FILE_EXTENSIONS: + if extension in _get_file_extensions(self._file_types): return True - for prefix in OFFICE_MIME_TYPE_PREFIXES + OTHER_MIME_TYPE_PREFIXES: + for prefix in _get_mime_type_prefixes(self._file_types): if mimetype.startswith(prefix): return True @@ -121,10 +188,18 @@ class DocumentIntelligenceConverter(DocumentConverter): mimetype = (stream_info.mimetype or "").lower() extension = (stream_info.extension or "").lower() - if extension in OFFICE_FILE_EXTENSIONS: + # Types that don't support ocr + no_ocr_types = [ + DocumentIntelligenceFileType.DOCX, + DocumentIntelligenceFileType.PPTX, + DocumentIntelligenceFileType.XLSX, + DocumentIntelligenceFileType.HTML, + ] + + if extension in _get_file_extensions(no_ocr_types): return [] - for prefix in OFFICE_MIME_TYPE_PREFIXES: + for prefix in _get_mime_type_prefixes(no_ocr_types): if mimetype.startswith(prefix): return [] diff --git a/packages/markitdown/tests/test_module_misc.py b/packages/markitdown/tests/test_module_misc.py index ad0367b..5e9ced5 100644 --- a/packages/markitdown/tests/test_module_misc.py +++ b/packages/markitdown/tests/test_module_misc.py @@ -6,6 +6,8 @@ import shutil import openai import pytest +from markitdown._uri_utils import parse_data_uri, file_uri_to_path + from markitdown import ( MarkItDown, UnsupportedFormatException, @@ -177,6 +179,79 @@ def test_stream_info_operations() -> None: assert updated_stream_info.url == "url.1" +def test_data_uris() -> None: + # Test basic parsing of data URIs + data_uri = "data:text/plain;base64,SGVsbG8sIFdvcmxkIQ==" + mime_type, attributes, data = parse_data_uri(data_uri) + assert mime_type == "text/plain" + assert len(attributes) == 0 + assert data == b"Hello, World!" + + data_uri = "data:base64,SGVsbG8sIFdvcmxkIQ==" + mime_type, attributes, data = parse_data_uri(data_uri) + assert mime_type is None + assert len(attributes) == 0 + assert data == b"Hello, World!" + + data_uri = "data:text/plain;charset=utf-8;base64,SGVsbG8sIFdvcmxkIQ==" + mime_type, attributes, data = parse_data_uri(data_uri) + assert mime_type == "text/plain" + assert len(attributes) == 1 + assert attributes["charset"] == "utf-8" + assert data == b"Hello, World!" + + data_uri = "data:,Hello%2C%20World%21" + mime_type, attributes, data = parse_data_uri(data_uri) + assert mime_type is None + assert len(attributes) == 0 + assert data == b"Hello, World!" + + data_uri = "data:text/plain,Hello%2C%20World%21" + mime_type, attributes, data = parse_data_uri(data_uri) + assert mime_type == "text/plain" + assert len(attributes) == 0 + assert data == b"Hello, World!" + + data_uri = "data:text/plain;charset=utf-8,Hello%2C%20World%21" + mime_type, attributes, data = parse_data_uri(data_uri) + assert mime_type == "text/plain" + assert len(attributes) == 1 + assert attributes["charset"] == "utf-8" + assert data == b"Hello, World!" + + +def test_file_uris() -> None: + # Test file URI with an empty host + file_uri = "file:///path/to/file.txt" + netloc, path = file_uri_to_path(file_uri) + assert netloc is None + assert path == "/path/to/file.txt" + + # Test file URI with no host + file_uri = "file:/path/to/file.txt" + netloc, path = file_uri_to_path(file_uri) + assert netloc is None + assert path == "/path/to/file.txt" + + # Test file URI with localhost + file_uri = "file://localhost/path/to/file.txt" + netloc, path = file_uri_to_path(file_uri) + assert netloc == "localhost" + assert path == "/path/to/file.txt" + + # Test file URI with query parameters + file_uri = "file:///path/to/file.txt?param=value" + netloc, path = file_uri_to_path(file_uri) + assert netloc is None + assert path == "/path/to/file.txt" + + # Test file URI with fragment + file_uri = "file:///path/to/file.txt#fragment" + netloc, path = file_uri_to_path(file_uri) + assert netloc is None + assert path == "/path/to/file.txt" + + def test_docx_comments() -> None: markitdown = MarkItDown() @@ -327,6 +402,8 @@ if __name__ == "__main__": """Runs this file's tests from the command line.""" for test in [ test_stream_info_operations, + test_data_uris, + test_file_uris, test_docx_comments, test_input_as_strings, test_markitdown_remote, diff --git a/packages/markitdown/tests/test_module_vectors.py b/packages/markitdown/tests/test_module_vectors.py index 09e4a2b..98fd0c7 100644 --- a/packages/markitdown/tests/test_module_vectors.py +++ b/packages/markitdown/tests/test_module_vectors.py @@ -3,7 +3,9 @@ import os import time import pytest import codecs +import base64 +from pathlib import Path if __name__ == "__main__": from _test_vectors import GENERAL_TEST_VECTORS, DATA_URI_TEST_VECTORS @@ -108,8 +110,8 @@ def test_convert_stream_without_hints(test_vector): reason="do not run tests that query external urls", ) @pytest.mark.parametrize("test_vector", GENERAL_TEST_VECTORS) -def test_convert_url(test_vector): - """Test the conversion of a stream with no stream info.""" +def test_convert_http_uri(test_vector): + """Test the conversion of an HTTP:// or HTTPS:// URI.""" markitdown = MarkItDown() time.sleep(1) # Ensure we don't hit rate limits @@ -124,8 +126,44 @@ def test_convert_url(test_vector): assert string not in result.markdown +@pytest.mark.parametrize("test_vector", GENERAL_TEST_VECTORS) +def test_convert_file_uri(test_vector): + """Test the conversion of a file:// URI.""" + markitdown = MarkItDown() + + result = markitdown.convert( + Path(os.path.join(TEST_FILES_DIR, test_vector.filename)).as_uri(), + url=test_vector.url, + ) + for string in test_vector.must_include: + assert string in result.markdown + for string in test_vector.must_not_include: + assert string not in result.markdown + + +@pytest.mark.parametrize("test_vector", GENERAL_TEST_VECTORS) +def test_convert_data_uri(test_vector): + """Test the conversion of a data URI.""" + markitdown = MarkItDown() + + data = "" + with open(os.path.join(TEST_FILES_DIR, test_vector.filename), "rb") as stream: + data = base64.b64encode(stream.read()).decode("utf-8") + mimetype = test_vector.mimetype + data_uri = f"data:{mimetype};base64,{data}" + + result = markitdown.convert( + data_uri, + url=test_vector.url, + ) + for string in test_vector.must_include: + assert string in result.markdown + for string in test_vector.must_not_include: + assert string not in result.markdown + + @pytest.mark.parametrize("test_vector", DATA_URI_TEST_VECTORS) -def test_convert_with_data_uris(test_vector): +def test_convert_keep_data_uris(test_vector): """Test API functionality when keep_data_uris is enabled""" markitdown = MarkItDown() @@ -143,7 +181,7 @@ def test_convert_with_data_uris(test_vector): @pytest.mark.parametrize("test_vector", DATA_URI_TEST_VECTORS) -def test_convert_stream_with_data_uris(test_vector): +def test_convert_stream_keep_data_uris(test_vector): """Test the conversion of a stream with no stream info.""" markitdown = MarkItDown() @@ -175,7 +213,9 @@ if __name__ == "__main__": test_convert_local, test_convert_stream_with_hints, test_convert_stream_without_hints, - test_convert_url, + test_convert_http_uri, + test_convert_file_uri, + test_convert_data_uri, ]: for test_vector in GENERAL_TEST_VECTORS: print( @@ -186,8 +226,8 @@ if __name__ == "__main__": # Data URI tests for test_function in [ - test_convert_with_data_uris, - test_convert_stream_with_data_uris, + test_convert_keep_data_uris, + test_convert_stream_keep_data_uris, ]: for test_vector in DATA_URI_TEST_VECTORS: print(