pdf2md-llm


Namepdf2md-llm JSON
Version 0.1.2 PyPI version JSON
download
home_pageNone
SummaryA package to convert PDF files to Markdown using a local LLM.
upload_time2025-03-07 08:37:50
maintainerNone
docs_urlNone
authorLeon Eversberg
requires_python>=3.10
licenseMIT License Copyright (c) 2025 Leon Eversberg Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
keywords markdown pdf llm parser converter
VCS
bugtrack_url
requirements pdf2img vllm qwen-vl-utils accelerate transformers torch
Travis-CI No Travis.
coveralls test coverage No coveralls.
            # pdf2md_llm

`pdf2md_llm` is a Python package that converts PDF files to Markdown using a local Large Language Model (LLM). 

The package leverages the `pdf2image` library to convert PDF pages to images and a vision language model to generate Markdown text from these images.

## Features

- Convert PDF files to images.
- Generate Markdown text from images using a local LLM.
- Keep your data private. No third-party file uploads. 

## Installation

You need a CUDA compatible GPU to run local LLMs with vLLM.

You can use `pip` to install the package:

```bash
pip install pdf2md-llm
```
## Usage

### CLI

You can use the `pdf2md_llm` package via the **command line interface (CLI)**.

To convert a PDF file to Markdown, run the following command:

```bash
pdf2md_llm <pdf_file> [options]
```

#### Options

* `pdf_file`: Path to the PDF file to convert.
* `--model`: Name of the model to use (default: `Qwen/Qwen2.5-VL-3B-Instruct-AWQ`).
* `--dtype`: Data type for the model weights and activations (default: `None`).
* `--max_model_len`: Max model context length (default: `7000`).
* `--size`: Image size as a tuple (default: `(700, None)`).
* `--dpi`: DPI of the images (default: `200`).
* `--fmt`: Image format (default: `jpeg`).
* `--output_folder`: Folder to save the output Markdown file (default: `./out`).

#### Example

```bash
pdf2md_llm example.pdf --model "Qwen/Qwen2.5-VL-3B-Instruct-AWQ" --output_folder "./output"
```

##### Model Support:
Currently the following Qwen2.5-VL models are supported: 

* `Qwen/Qwen2.5-VL-3B-Instruct`
* `Qwen/Qwen2.5-VL-3B-Instruct-AWQ`
* `Qwen/Qwen2.5-VL-7B-Instruct`
* `Qwen/Qwen2.5-VL-7B-Instruct-AWQ`
* `Qwen/Qwen2.5-VL-72B-Instruct`
* `Qwen/Qwen2.5-VL-72B-Instruct-AWQ`

If you want to use a different model, feel free to add a vLLM compatible model to the factory function `llm_model()` in `llm.py`

### Python API

You can use the `pdf2md_llm` package via the **Python API**.

Basic usage:

```python
from vllm import SamplingParams

from pdf2md_llm.llm import llm_model
from pdf2md_llm.pdf2img import PdfToImg

pdf2img = PdfToImg(size=(700, None), output_folder="./out")
img_files = pdf2img.convert("example.pdf")

llm = llm_model(
    model="Qwen/Qwen2.5-VL-3B-Instruct-AWQ",
    dtype="half",
    max_num_seqs=1,
    max_model_len=7000,
)

sampling_params = SamplingParams(
    temperature=0.1,
    min_p=0.1,
    max_tokens=8192,
    stop_token_ids=[],
)

# Append all pages to one output Markdown file
for img_file in img_files:
    markdown_text = llm.generate(
        img_file, sampling_params=sampling_params
    )  # convert image to Markdown with LLM
    with open("example.md", "a", encoding="utf-8") as myfile:
        myfile.write(markdown_text)
```

For a full example, see [example_api.py](./pdf2md_llm/example_api.py)


## License

This project is licensed under the MIT License. See the LICENSE file for details.

## Acknowledgements

* [pdf2image](https://github.com/Belval/pdf2image) for converting PDF files to images.

* [Qwen2.5-VL](https://github.com/QwenLM/Qwen2.5-VL) LLM model

* [vLLM](https://github.com/vllm-project/vllm) for efficient LLM model inference

            

Raw data

            {
    "_id": null,
    "home_page": null,
    "name": "pdf2md-llm",
    "maintainer": null,
    "docs_url": null,
    "requires_python": ">=3.10",
    "maintainer_email": null,
    "keywords": "markdown, pdf, llm, parser, converter",
    "author": "Leon Eversberg",
    "author_email": null,
    "download_url": "https://files.pythonhosted.org/packages/ae/8c/d76733999e4b1968e4a7d15daa11580ec6f1c25b6b5e3033e5173b7b171c/pdf2md_llm-0.1.2.tar.gz",
    "platform": null,
    "description": "# pdf2md_llm\n\n`pdf2md_llm` is a Python package that converts PDF files to Markdown using a local Large Language Model (LLM). \n\nThe package leverages the `pdf2image` library to convert PDF pages to images and a vision language model to generate Markdown text from these images.\n\n## Features\n\n- Convert PDF files to images.\n- Generate Markdown text from images using a local LLM.\n- Keep your data private. No third-party file uploads. \n\n## Installation\n\nYou need a CUDA compatible GPU to run local LLMs with vLLM.\n\nYou can use `pip` to install the package:\n\n```bash\npip install pdf2md-llm\n```\n## Usage\n\n### CLI\n\nYou can use the `pdf2md_llm` package via the **command line interface (CLI)**.\n\nTo convert a PDF file to Markdown, run the following command:\n\n```bash\npdf2md_llm <pdf_file> [options]\n```\n\n#### Options\n\n* `pdf_file`: Path to the PDF file to convert.\n* `--model`: Name of the model to use (default: `Qwen/Qwen2.5-VL-3B-Instruct-AWQ`).\n* `--dtype`: Data type for the model weights and activations (default: `None`).\n* `--max_model_len`: Max model context length (default: `7000`).\n* `--size`: Image size as a tuple (default: `(700, None)`).\n* `--dpi`: DPI of the images (default: `200`).\n* `--fmt`: Image format (default: `jpeg`).\n* `--output_folder`: Folder to save the output Markdown file (default: `./out`).\n\n#### Example\n\n```bash\npdf2md_llm example.pdf --model \"Qwen/Qwen2.5-VL-3B-Instruct-AWQ\" --output_folder \"./output\"\n```\n\n##### Model Support:\nCurrently the following Qwen2.5-VL models are supported: \n\n* `Qwen/Qwen2.5-VL-3B-Instruct`\n* `Qwen/Qwen2.5-VL-3B-Instruct-AWQ`\n* `Qwen/Qwen2.5-VL-7B-Instruct`\n* `Qwen/Qwen2.5-VL-7B-Instruct-AWQ`\n* `Qwen/Qwen2.5-VL-72B-Instruct`\n* `Qwen/Qwen2.5-VL-72B-Instruct-AWQ`\n\nIf you want to use a different model, feel free to add a vLLM compatible model to the factory function `llm_model()` in `llm.py`\n\n### Python API\n\nYou can use the `pdf2md_llm` package via the **Python API**.\n\nBasic usage:\n\n```python\nfrom vllm import SamplingParams\n\nfrom pdf2md_llm.llm import llm_model\nfrom pdf2md_llm.pdf2img import PdfToImg\n\npdf2img = PdfToImg(size=(700, None), output_folder=\"./out\")\nimg_files = pdf2img.convert(\"example.pdf\")\n\nllm = llm_model(\n    model=\"Qwen/Qwen2.5-VL-3B-Instruct-AWQ\",\n    dtype=\"half\",\n    max_num_seqs=1,\n    max_model_len=7000,\n)\n\nsampling_params = SamplingParams(\n    temperature=0.1,\n    min_p=0.1,\n    max_tokens=8192,\n    stop_token_ids=[],\n)\n\n# Append all pages to one output Markdown file\nfor img_file in img_files:\n    markdown_text = llm.generate(\n        img_file, sampling_params=sampling_params\n    )  # convert image to Markdown with LLM\n    with open(\"example.md\", \"a\", encoding=\"utf-8\") as myfile:\n        myfile.write(markdown_text)\n```\n\nFor a full example, see [example_api.py](./pdf2md_llm/example_api.py)\n\n\n## License\n\nThis project is licensed under the MIT License. See the LICENSE file for details.\n\n## Acknowledgements\n\n* [pdf2image](https://github.com/Belval/pdf2image) for converting PDF files to images.\n\n* [Qwen2.5-VL](https://github.com/QwenLM/Qwen2.5-VL) LLM model\n\n* [vLLM](https://github.com/vllm-project/vllm) for efficient LLM model inference\n",
    "bugtrack_url": null,
    "license": "MIT License  Copyright (c) 2025 Leon Eversberg  Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the \"Software\"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:  The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.  THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. ",
    "summary": "A package to convert PDF files to Markdown using a local LLM.",
    "version": "0.1.2",
    "project_urls": {
        "Homepage": "https://github.com/leoneversberg/pdf2md_llm",
        "Repository": "https://github.com/leoneversberg/pdf2md_llm.git"
    },
    "split_keywords": [
        "markdown",
        " pdf",
        " llm",
        " parser",
        " converter"
    ],
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "blake2b_256": "59b802aade81da024cb40de2b5283702cda30469e6380f49cf3e5651ed2f8dff",
                "md5": "3fca9d6a78ad831271a2b0e032b41b92",
                "sha256": "467131d7c1e70778cfed29c45dbc09b2e2ed3b2a477cc49c67b2a13783c5f486"
            },
            "downloads": -1,
            "filename": "pdf2md_llm-0.1.2-py3-none-any.whl",
            "has_sig": false,
            "md5_digest": "3fca9d6a78ad831271a2b0e032b41b92",
            "packagetype": "bdist_wheel",
            "python_version": "py3",
            "requires_python": ">=3.10",
            "size": 10413,
            "upload_time": "2025-03-07T08:37:49",
            "upload_time_iso_8601": "2025-03-07T08:37:49.156974Z",
            "url": "https://files.pythonhosted.org/packages/59/b8/02aade81da024cb40de2b5283702cda30469e6380f49cf3e5651ed2f8dff/pdf2md_llm-0.1.2-py3-none-any.whl",
            "yanked": false,
            "yanked_reason": null
        },
        {
            "comment_text": "",
            "digests": {
                "blake2b_256": "ae8cd76733999e4b1968e4a7d15daa11580ec6f1c25b6b5e3033e5173b7b171c",
                "md5": "debc7ca9dfbc28a66bed326d2130a988",
                "sha256": "4caec22862361c577584d6fc37c426193b7e68bbce8861cdfb1652e43a99dc61"
            },
            "downloads": -1,
            "filename": "pdf2md_llm-0.1.2.tar.gz",
            "has_sig": false,
            "md5_digest": "debc7ca9dfbc28a66bed326d2130a988",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": ">=3.10",
            "size": 9315,
            "upload_time": "2025-03-07T08:37:50",
            "upload_time_iso_8601": "2025-03-07T08:37:50.598731Z",
            "url": "https://files.pythonhosted.org/packages/ae/8c/d76733999e4b1968e4a7d15daa11580ec6f1c25b6b5e3033e5173b7b171c/pdf2md_llm-0.1.2.tar.gz",
            "yanked": false,
            "yanked_reason": null
        }
    ],
    "upload_time": "2025-03-07 08:37:50",
    "github": true,
    "gitlab": false,
    "bitbucket": false,
    "codeberg": false,
    "github_user": "leoneversberg",
    "github_project": "pdf2md_llm",
    "travis_ci": false,
    "coveralls": false,
    "github_actions": true,
    "requirements": [
        {
            "name": "pdf2img",
            "specs": [
                [
                    "~=",
                    "0.1.2"
                ]
            ]
        },
        {
            "name": "vllm",
            "specs": [
                [
                    "==",
                    "0.7.3"
                ]
            ]
        },
        {
            "name": "qwen-vl-utils",
            "specs": [
                [
                    "~=",
                    "0.0.10"
                ]
            ]
        },
        {
            "name": "accelerate",
            "specs": [
                [
                    "~=",
                    "1.4.0"
                ]
            ]
        },
        {
            "name": "transformers",
            "specs": [
                [
                    "~=",
                    "4.49.0"
                ]
            ]
        },
        {
            "name": "torch",
            "specs": [
                [
                    "==",
                    "2.5.1"
                ]
            ]
        }
    ],
    "lcname": "pdf2md-llm"
}
        
Elapsed time: 1.86730s