BabelDOC

Tue, 15 Apr 2025 15:30:04 +0800

funstory-ai/BabelDOC

PDF scientific paper translation and bilingual comparison library.

Online Service: Beta version launched Immersive Translate - BabelDOC 1000 free pages per month.
Self-deployment: PDFMathTranslate 1.9.3+ Experimental support for BabelDOC, available for self-deployment + WebUI with more translation services.
Provides a simple command line interface.
Provides a Python API.
Mainly designed to be embedded into other programs, but can also be used directly for simple translation tasks.

Preview

We are hiring

See details: EN | ZH

Getting Started

Install from PyPI

We recommend using the Tool feature of uv to install yadt.

First, you need to refer to uv installation to install uv and set up the PATH environment variable as prompted.
Use the following command to install yadt:

1
2
3

uv tool install --python 3.12 BabelDOC

babeldoc --help

Use the babeldoc command. For example:

babeldoc --bing  --files example.pdf

# multiple files
babeldoc --bing  --files example1.pdf --files example2.pdf

Install from Source

We still recommend using uv to manage virtual environments.

First, you need to refer to uv installation to install uv and set up the PATH environment variable as prompted.
Use the following command to install yadt:

# clone the project
git clone https://github.com/funstory-ai/BabelDOC

# enter the project directory
cd BabelDOC

# install dependencies and run babeldoc
uv run babeldoc --help

Use the uv run babeldoc command. For example:

uv run babeldoc --files example.pdf --openai --openai-model "gpt-4o-mini" --openai-base-url "https://api.openai.com/v1" --openai-api-key "your-api-key-here"

# multiple files
uv run babeldoc --files example.pdf --files example2.pdf --openai --openai-model "gpt-4o-mini" --openai-base-url "https://api.openai.com/v1" --openai-api-key "your-api-key-here"

[!TIP] The absolute path is recommended.

Advanced Options

[!NOTE] This CLI is mainly for debugging purposes. Although end users can use this CLI to translate files, we do not provide any technical support for this purpose.

End users should directly use Online Service: Beta version launched Immersive Translate - BabelDOC 1000 free pages per month.

End users who need self-deployment should use PDFMathTranslate

If you find that an option is not listed below, it means that this option is a debugging option for maintainers. Please do not use these options.

Language Options

--lang-in, -li: Source language code (default: en)
--lang-out, -lo: Target language code (default: zh)

[!TIP] Currently, this project mainly focuses on English-to-Chinese translation, and other scenarios have not been tested yet.

(2025.3.1 update): Basic English target language support has been added, primarily to minimize line breaks within words([0-9A-Za-z]+).

HELP WANTED: Collecting word regular expressions for more languages

PDF Processing Options

--files: One or more file paths to input PDF documents.
--pages, -p: Specify pages to translate (e.g., “1,2,1-,-3,3-5”). If not set, translate all pages
--split-short-lines: Force split short lines into different paragraphs (may cause poor typesetting & bugs)
--short-line-split-factor: Split threshold factor (default: 0.8). The actual threshold is the median length of all lines on the current page * this factor
--skip-clean: Skip PDF cleaning step
--dual-translate-first: Put translated pages first in dual PDF mode (default: original pages first)
--disable-rich-text-translate: Disable rich text translation (may help improve compatibility with some PDFs)
--enhance-compatibility: Enable all compatibility enhancement options (equivalent to –skip-clean –dual-translate-first –disable-rich-text-translate)
--use-alternating-pages-dual: Use alternating pages mode for dual PDF. When enabled, original and translated pages are arranged in alternate order. When disabled (default), original and translated pages are shown side by side on the same page.
--watermark-output-mode: Control watermark output mode: ‘watermarked’ (default) adds watermark to translated PDF, ’no_watermark’ doesn’t add watermark, ‘both’ outputs both versions.
--max-pages-per-part: Maximum number of pages per part for split translation. If not set, no splitting will be performed.
--no-watermark: [DEPRECATED] Use –watermark-output-mode=no_watermark instead.
--translate-table-text: Translate table text (experimental, default: False)
--skip-scanned-detection: Skip scanned document detection (default: False). When using split translation, only the first part performs detection if not skipped.

[!TIP]

Both --skip-clean and --dual-translate-first may help improve compatibility with some PDF readers

--disable-rich-text-translate can also help with compatibility by simplifying translation input

However, using --skip-clean will result in larger file sizes

If you encounter any compatibility issues, try using --enhance-compatibility first

Use --max-pages-per-part for large documents to split them into smaller parts for translation and automatically merge them back.

Use --skip-scanned-detection to speed up processing when you know your document is not a scanned PDF.

Translation Service Options

--qps: QPS (Queries Per Second) limit for translation service (default: 4)
--ignore-cache: Ignore translation cache and force retranslation
--no-dual: Do not output bilingual PDF files
--no-mono: Do not output monolingual PDF files
--min-text-length: Minimum text length to translate (default: 5)
--openai: Use OpenAI for translation (default: False)

[!TIP]

Currently, only OpenAI-compatible LLM is supported. For more translator support, please use PDFMathTranslate.

It is recommended to use models with strong compatibility with OpenAI, such as: glm-4-flash, deepseek-chat, etc.

Currently, it has not been optimized for traditional translation engines like Bing/Google, it is recommended to use LLMs.

You can use litellm to access multiple models.

OpenAI Specific Options

--openai-model: OpenAI model to use (default: gpt-4o-mini)
--openai-base-url: Base URL for OpenAI API
--openai-api-key: API key for OpenAI service

[!TIP]

This tool supports any OpenAI-compatible API endpoints. Just set the correct base URL and API key. (e.g. https://xxx.custom.xxx/v1)

For local models like Ollama, you can use any value as the API key (e.g. --openai-api-key a).

Output Control

--output, -o: Output directory for translated files. If not set, use current working directory.
--debug, -d: Enable debug logging level and export detailed intermediate results in ~/.cache/yadt/working.
--report-interval: Progress report interval in seconds (default: 0.1).

Offline Assets Management

--generate-offline-assets: Generate an offline assets package in the specified directory. This creates a zip file containing all required models and fonts.
--restore-offline-assets: Restore an offline assets package from the specified file. This extracts models and fonts from a previously generated package.

[!TIP]

Offline assets packages are useful for environments without internet access or to speed up installation on multiple machines.

Generate a package once with babeldoc --generate-offline-assets /path/to/output/dir and then distribute it.

Restore the package on target machines with babeldoc --restore-offline-assets /path/to/offline_assets_*.zip.

The offline assets package name cannot be modified because the file list hash is encoded in the name.

If you provide a directory path to --restore-offline-assets, the tool will automatically look for the correct offline assets package file in that directory.

The package contains all necessary fonts and models required for document processing, ensuring consistent results across different environments.

The integrity of all assets is verified using SHA3-256 hashes during both packaging and restoration.

If you’re deploying in an air-gapped environment, make sure to generate the package on a machine with internet access first.

Configuration File

--config, -c: Configuration file path. Use the TOML format.

Example Configuration:

[babeldoc]
# Basic settings
debug = true
lang-in = "en-US"
lang-out = "zh-CN"
qps = 10
output = "/path/to/output/dir"

# PDF processing options
split-short-lines = false
short-line-split-factor = 0.8
skip-clean = false
dual-translate-first = false
disable-rich-text-translate = false
use-alternating-pages-dual = false
watermark-output-mode = "watermarked"  # Choices: "watermarked", "no_watermark", "both"
max-pages-per-part = 50  # Automatically split the document for translation and merge it back.
# no-watermark = false  # DEPRECATED: Use watermark-output-mode instead
skip-scanned-detection = false  # Skip scanned document detection for faster processing

# Translation service
openai = true
openai-model = "gpt-4o-mini"
openai-base-url = "https://api.openai.com/v1"
openai-api-key = "your-api-key-here"

# Output control
no-dual = false
no-mono = false
min-text-length = 5
report-interval = 0.5

# Offline assets management
# Uncomment one of these options as needed:
# generate-offline-assets = "/path/to/output/dir"
# restore-offline-assets = "/path/to/offline_assets_package.zip"

Python API

[!TIP]

Before pdf2zh 2.0 is released, you can temporarily use BabelDOC’s Python API. However, after pdf2zh 2.0 is released, please directly use pdf2zh’s Python API.

This project’s Python API does not guarantee any compatibility. However, the Python API from pdf2zh will guarantee a certain level of compatibility.

You can refer to the example in main.py to use BabelDOC’s Python API.

Please note:

Make sure call babeldoc.high_level.init() before using the API
The current TranslationConfig does not fully validate input parameters, so you need to ensure the validity of input parameters

For offline assets management, you can use the following functions:

# Generate an offline assets package
from pathlib import Path
import babeldoc.assets.assets

# Generate package to a specific directory
# path is optional, default is ~/.cache/babeldoc/assets/offline_assets_{hash}.zip
babeldoc.assets.assets.generate_offline_assets_package(Path("/path/to/output/dir"))

# Restore from a package file
# path is optional, default is ~/.cache/babeldoc/assets/offline_assets_{hash}.zip
babeldoc.assets.assets.restore_offline_assets_package(Path("/path/to/offline_assets_package.zip"))

# You can also restore from a directory containing the offline assets package
# The tool will automatically find the correct package file based on the hash
babeldoc.assets.assets.restore_offline_assets_package(Path("/path/to/directory"))

[!TIP]

The offline assets package name cannot be modified because the file list hash is encoded in the name.

When using in production environments, it’s recommended to pre-generate the assets package and include it with your application distribution.

The package verification ensures that all required assets are intact and match their expected checksums.

Background

There are a lot projects and teams working on to make document editing and translating easier like:

There are also some solutions to solve specific parts of the problem like:

layoutreader: the read order of the text block in a pdf
Surya: the structure of the pdf

This project hopes to promote a standard pipeline and interface to solve the problem.

In fact, there are two main stages of a PDF parser or translator:

Parsing: A stage of parsing means to get the structure of the pdf such as text blocks, images, tables, etc.
Rendering: A stage of rendering means to render the structure into a new pdf or other format.

For a service like mathpix, it will parse the pdf into a structure may be in a XML format, and then render them using a single column reader order as layoutreader does. The bad news is that the original structure lost.

Some people will use Adobe PDF Parser because it will generate a Word document and it keeps the original structure. But it is somewhat expensive. And you know, a pdf or word document is not a good format for reading in mobile devices.

We offer an intermediate representation of the results from parser and can be rendered into a new pdf or other format. The pipeline is also a plugin-based system which everybody can add their new model, ocr, renderer, etc.

Roadmap

Add line support
Add table support
Add cross-page/cross-column paragraph support
More advanced typesetting features
Outline support
…

Our first 1.0 version goal is to finish a translation from PDF Reference, Version 1.7 to the following language version:

Simplified Chinese
Traditional Chinese
Japanese
Spanish

And meet the following requirements:

layout error less than 1%
content loss less than 1%

Known Issues

Parsing errors in the author and reference sections; they get merged into one paragraph after translation.
Lines are not supported.
Does not support drop caps.
Large pages will be skipped.

How to Contribute

We encourage you to contribute to YADT! Please check out the CONTRIBUTING guide.

Everyone interacting in YADT and its sub-projects’ codebases, issue trackers, chat rooms, and mailing lists is expected to follow the YADT Code of Conduct.

Immersive Translation sponsors monthly Pro membership redemption codes for active contributors to this project, see details at: CONTRIBUTOR_REWARD.md

PDFMathTranslate on Producthunt daily