GitLab

# What's Changed

### New Features

- Add table content recognition, we use weights of [StructEqTable](https://github.com/UniModal4Reasoning/StructEqTable-Deploy) to convert table image to LaTex.

### Instruction

- pip install pypandoc struct-eqtable==0.1.0
- Download [StructEqTable weights](https://huggingface.co/wanderkid/PDF-Extract-Kit/tree/main/models/TabRec) and put it under models/ directory.
- Edit 'table-mode' value to turn on table recognition function which is turned off by default.
- If you did not download any models before, refer to [how to download models](docs/how_to_download_models_zh_cn.md)。

Update the PyTorch installation command in the README files for both English and Chinese
versions to reflect the required version compatibility with CUDA 11.8. Include explicit
instructions to specify the PyTorch version to avoid automatic installation of higher,
unsupported versions. Additionally, clarify the importance of modifying the "device-mode"
parameter in the magic-pdf.json configuration file for proper CUDA device selection.

Update the PyTorch installation guide in both English and Chinese READMEs to explicitly
recommend using torch==2.3.1 and torchvision==0.18.1 for CUDA 11.8. Emphasize the
importance of specifying these versions to avoid compatibility issues with higher,
unsupported versions.

A pre-release version0.6.2b1 of magic-pdf is now available. This version includes
many fixes addressed in our logs but has not undergone full QA testing. Users are
advised to report any issues encountered or revert to version 0.6.1. The installationguides in both Japanese and Chinese READMEs have been updated to reflect the availability
of this pre-release version and the previous stable version.

BREAKING CHANGE: Installation commands now point to version 0.6.2b1 by default.
Users who wish to install the stable version 0.6.1 should follow the provided
command instead.

- Change the pip install command in README_zh-CN.md to reflect the new version 0.6.2b1.
- Include a notice about the pre-release of version 0.6.2beta, cautioning users about its未经完整QA测试的状态，并提供回退到0.6.1版本的指导。
- Verify the installed version with `magic-pdf --version` after installation to ensure
 the correct version is installed, addressing feedback about incorrect versions due to
 mirror source and dependency conflicts.

## Changelog
31/07/20204
- Support table recognition. Table images will be converted into LaTex.

### how to use the new feature:
set the attribute 'table-mode' to 'true' in magic-pdf.json

### caution:
it takes 200s to 500s to convert a single table image using cpu

We have pre-released the 0.6.2 beta version, which addresses numerous issues
reported in our logs. This commit updates the installation guide in the README to
include information on how to install this beta version. Users are advised that
this build has not undergone full QA testing and may contain issues. A
revert instruction to version 0.6.1 is also provided for users who encounter
problems.

BREAKING CHANGE: Installation instructions now include beta version
information. Users should be aware of potential issues with the0.6.2 beta
version and consider reverting to 0.6.1 if necessary.

## Changelog
31/07/20204
- Support table recognition. Table images will be converted into html.

### how to use the new feature:
set the attribute 'table-mode' to 'true' in magic-pdf.json

### caution:
it takes 200s to 500s to convert a single table image using cpu

Add detailed instructions for installing git lfs on Linux, macOS, and Windows
to facilitate users in downloading models from ModelScope repository. The guide
is included in the `how_to_download_models_zh_cn.md` document.

fix(readme): add China mirror for pip install in README_zh-CNAdd China mirror links for `pip install` commands in the Chinese README to improve
accessibility and speed up package installation for users in China.

- Emphasize the importance of mainline support for stability and reliability.
- Define the supported operating systems, CPU architectures, memory requirements,
  Python versions, Nvidia drivers, and CUDA environments.
- Provide guidance on GPU hardware support and minimum/recommended configurations.
- Include a note on compatibility issues and the importance of using the specified
  versions of dependencies.
- Update installation command and verify version to prevent incorrect package installations.
- Add missing GPU quick start guides for Ubuntu and Windows.
- Refactor usage example code by initializing an empty list for `model_list` and
  adding a call to `pipe.pipe_analyze()`.

Redesign the structure of the readme.md to make the document hierarchy clearer, making it easier for users to understand.

Redesign the structure of the readme.md to make the document hierarchy clearer, making it easier for users to understand.

Redesign the structure of the readme.md to make the document hierarchy clearer, making it easier for users to understand.

docs: add Japanese README file