Volver al ranking

google-research/arxiv-latex-cleaner

Python

arXiv LaTeX Cleaner: Easily clean the LaTeX code of your paper to submit to arXiv

arxivlatex
Crecimiento de estrellas
Estrellas
7k
Forks
412
Crecimiento semanal
Issues
30
2k4k6k
ene 2019jul 2021ene 2024jul 2026
ArtefactosPyPIpip install arxiv-latex-cleaner
README

arxiv_latex_cleaner

This tool allows you to easily clean the LaTeX code of your paper to submit to arXiv. From a folder containing all your code, e.g. /path/to/latex/, it creates a new folder /path/to/latex_arXiv/, that is ready to ZIP and upload to arXiv.

Example call:

arxiv_latex_cleaner /path/to/latex --resize_images --im_size 500 --images_allowlist='{"images/im.png":2000}'

Or simply from a config file

arxiv_latex_cleaner /path/to/latex --config cleaner_config.yaml

Installation:

pip install arxiv-latex-cleaner
:exclamation: arxiv_latex_cleaner is only compatible with Python >=3.9 :exclamation:

If using MacOS, you can install using Homebrew:

brew install arxiv_latex_cleaner

Alternatively, you can download the source code:

git clone https://github.com/google-research/arxiv-latex-cleaner
cd arxiv-latex-cleaner/
python -m arxiv_latex_cleaner --help

And install as a command-line program directly from the source code:

python setup.py install

Main features:

Privacy-oriented

  • Removes all auxiliary files (.aux, .log, .out, etc.).
  • Removes all comments from your code (yes, those are visible on arXiv and you do not want them to be). These also include \begin{comment}\end{comment}, \iffalse\fi, and \if0\fi environments.
  • Optionally removes user-defined commands entered with commands_to_delete (such as \todo{} that you redefine as the empty string at the end).
  • Optionally allows you to define custom regex replacement rules through a cleaner_config.yaml file.

Size-oriented

There is a 50MB limit on arXiv submissions, so to make it fit:

  • Removes all unused .tex files (those that are not in the root and not included in any other .tex file).
  • Removes all unused images that take up space (those that are not actually included in any used .tex file).
  • Optionally resizes all images to im_size pixels, to reduce the size of the submission. You can allowlist some images to skip the global size using images_allowlist.
  • Optionally compresses .pdf files using ghostscript (Linux and Mac only). You can allowlist some PDFs to skip the global size using images_allowlist.
  • Optionally converts PNG images to JPG format to reduce file size.

TikZ picture source code concealment

To prevent the upload of tikzpicture source code or raw simulation data, this feature:

  • Replaces the tikzpicture environment \begin{tikzpicture} ... \end{tikzpicture} with the respective \includegraphics{EXTERNAL_TIKZ_FOLDER/picture_name.pdf}.
  • Requires externally compiled TikZ pictures as .pdf files in folder EXTERNAL_TIKZ_FOLDER. See section 52 (Externalization Library) in the PGF/TikZ manual on TikZ picture externalization.
  • Only replaces environments with preceding \tikzsetnextfilename{picture_name} command (as in \tikzsetnextfilename{picture_name}\begin{tikzpicture} ... \end{tikzpicture}) where the externalized picture_name.pdf filename matches picture_name.

More sophisticated pattern replacement based on regex group captures

Sometimes it is useful to work with a set of custom LaTeX commands when writing a paper. To get rid of them upon arXiv submission, one can simply revert them to plain LaTeX with a regular expression insertion.

{
    "pattern" : '(?:\\figcomp{\s*)(?P<first>.*?)\s*}\s*{\s*(?P<second>.*?)\s*}\s*{\s*(?P<third>.*?)\s*}',
    "insertion" : '\parbox[c]{{ {second} \linewidth}} {{ \includegraphics[width= {third} \linewidth]{{figures/{first} }} }}',
    "description" : "Replace figcomp"
}

The pattern above will find all \figcomp{path}{w1}{w2} commands and replace them with \parbox[c]{w1\linewidth}{\includegraphics[width=w2\linewidth]{figures/path}}. Note that the insertion template is filled with the named groups captures from the pattern. Note that the replacement is processed before all \includegraphics commands are processed and corresponding file paths are copied, making sure all figure files are copied to the cleaned version. See also cleaner_config.yaml for details on how to specify the patterns.

Usage:

usage: arxiv_latex_cleaner@v1.0.11 [-h] [--resize_images] [--im_size IM_SIZE]
                                   [--compress_pdf]
                                   [--pdf_im_resolution PDF_IM_RESOLUTION]
                                   [--images_allowlist IMAGES_ALLOWLIST]
                                   [--keep_bib]
                                   [--commands_to_delete COMMANDS_TO_DELETE [COMMANDS_TO_DELETE ...]]
                                   [--commands_only_to_delete COMMANDS_ONLY_TO_DELETE [COMMANDS_ONLY_TO_DELETE ...]]
                                   [--environments_to_delete ENVIRONMENTS_TO_DELETE [ENVIRONMENTS_TO_DELETE ...]]
                                   [--if_exceptions IF_EXCEPTIONS [IF_EXCEPTIONS ...]]
                                   [--use_external_tikz USE_EXTERNAL_TIKZ]
                                   [--svg_inkscape [SVG_INKSCAPE]]
                                   [--convert_png_to_jpg]
                                   [--png_quality PNG_QUALITY]
                                   [--png_size_threshold PNG_SIZE_THRESHOLD]
                                   [--config CONFIG] [--verbose]
                                   input_folder

Clean the LaTeX code of your paper to submit to arXiv. Check the README for
more information on the use.

positional arguments:
  input_folder          Input folder containing the LaTeX code.

optional arguments:
  -h, --help            show this help message and exit
  --resize_images       Resize images.
  --im_size IM_SIZE     Size of the output images (in pixels, longest side).
                        Fine tune this to get as close to 10MB as possible.
  --compress_pdf        Compress PDF images using ghostscript (Linux and Mac
                        only).
  --pdf_im_resolution PDF_IM_RESOLUTION
                        Resolution (in dpi) to which the tool resamples the
                        PDF images.
  --images_allowlist IMAGES_ALLOWLIST
                        Images (and PDFs) that won't be resized to the default
                        resolution, but the one provided here. Value is pixel
                        for images, and dpi forPDFs, as in --im_size and
                        --pdf_im_resolution, respectively. Format is a
                        dictionary as: '{"path/to/im.jpg": 1000}'
  --keep_bib            Avoid deleting the *.bib files.
  --commands_to_delete COMMANDS_TO_DELETE [COMMANDS_TO_DELETE ...]
                        LaTeX commands that will be deleted. Useful for e.g.
                        user-defined \todo commands. For example, to delete
                        all occurrences of \todo1{} and \todo2{}, run the tool
                        with `--commands_to_delete todo1 todo2`.Please note
                        that the positional argument `input_folder` cannot
                        come immediately after `commands_to_delete`, as the
                        parser does not have any way to know if it's another
                        command to delete.
  --commands_only_to_delete COMMANDS_ONLY_TO_DELETE [COMMANDS_ONLY_TO_DELETE ...]
                        LaTeX commands that will be deleted but the text 
                        wrapped in the commands will be retained. Useful for
                        commands that change text formats and colors, which
                        you may want to remove but keep the text within. Usages
                        are exactly the same as commands_to_delete. Note that if
                        the commands listed here duplicate that after
                        commands_to_delete, the default action will be retaining
                        the wrapped text.
  --environments_to_delete ENVIRONMENTS_TO_DELETE [ENVIRONMENTS_TO_DELETE ...]
                        LaTeX environments that will be deleted. Useful for e.g. 
                        user-defined comment environments. For example, to 
                        delete all occurrences of \begin{note} ... \end{note},
                        run the tool with `--environments_to_delete note`. 
                        Please note that the positional argument `input_folder`
                        cannot come immediately after
                        `environments_to_delete`, as the parser does not have
                        any way to know if it's another environment to delete.
  --if_exceptions IF_EXCEPTIONS [IF_EXCEPTIONS ...]
                        Constant TeX primitive conditionals (\iffalse, \iftrue,
                        etc.) are simplified, i.e., true branches are kept, false
                        branches deleted. To parse the conditional constructs
                        correctly, all commands starting with `\if` are assumed to
                        be TeX primitive conditionals (e.g., declared by
                        \newif\ifvar). Some known exceptions to this rule are
                        already included (e.g., \iff, \ifthenelse, etc.), but you
                        can add custom exceptions using `--if_exceptions iffalt`.
  --use_external_tikz USE_EXTERNAL_TIKZ
                        Folder (relative to input folder) containing
                        externalized tikz figures in PDF format.
  --svg_inkscape [SVG_INKSCAPE]
                        Include PDF files generated by Inkscape via the
                        `\includesvg` command from the `svg` package. This is
                        done by replacing the `\includesvg` calls with
                        `\includeinkscape` calls pointing to the generated
                        `.pdf_tex` files. By default, these files and the
                        generated PDFs are located under `./svg-inkscape`
                        (relative to the input folder), but a different path
                        (relative to the input folder) can be provided in case a
                        different `inkscapepath` was set when loading the `svg`
                        package.
  --convert_png_to_jpg  Convert PNG images to JPG format to reduce file size
  --png_quality PNG_QUALITY
                        JPG quality for PNG conversion (0-100, default: 50)
  --png_size_threshold PNG_SIZE_THRESHOLD
                        Minimum PNG file size in MB to apply quality reduction (default: 0.5)
  --config CONFIG       Read settings from `.yaml` config file. If command
                        line arguments are provided additionally, the config
                        file parameters are updated with the command line
                        parameters.
  --verbose             Enable detailed output.

Testing:

python -m unittest arxiv_latex_cleaner.tests.arxiv_latex_cleaner_test

Note

This is not an officially supported Google product.

Repositorios relacionados
kaixindelele/ChatPaper

Use ChatGPT to summarize the arXiv papers. 全流程加速科研,利用chatgpt进行论文全文总结+专业翻译+润色+审稿+审稿回复

PythonPyPIOtherarxivpaper
academic.chatpaper.top
19.7k1.9k
iperov/DeepFaceLab

DeepFaceLab is the leading software for creating deepfakes.

PythonPyPIGNU General Public License v3.0faceswapface-swap
19.3k919
LearningCircuit/local-deep-research

~95% on SimpleQA (e.g. Qwen3.6-27B on a 3090). Supports all local and cloud LLMs (llama.cpp, Ollama, Google, ...). 10+ search engines - arXiv, PubMed, your private documents. Everything Local & Encrypted.

PythonPyPIMIT Licenseacademiaarxiv
8.8k769
TideDra/zotero-arxiv-daily

Recommend new arxiv papers of your interest daily according to your Zotero libarary.

PythonPyPIGNU Affero General Public License v3.0arxivautomation
5.7k5k
sbrugman/deep-learning-papers

Papers about deep learning ordered by task, date. Current state-of-the-art papers are labelled.

deep-learningneural-networks
3.2k406
DeepGraphLearning/LiteratureDL4Graph

A comprehensive collection of recent papers on graph deep learning

MIT Licensemachine-learningdeep-learning
3.1k560
blazickjp/arxiv-mcp-server

A Model Context Protocol server for searching and analyzing arXiv papers

PythonPyPIApache License 2.0aiclaude-ai
3k238
dw-dengwei/daily-arXiv-ai-enhanced

Automatically crawl arXiv papers daily and summarize them using AI. Illustrating them using GitHub Pages.

JavaScriptnpmOtherai-toolsarxiv
arxiv.dw-dengwei.cn
2.9k1.1k
wenbihan/reproducible-image-denoising-state-of-the-art

Collection of popular and reproducible image denoising works.

image-denoisingbenchmarking
2.5k506
llmsresearch/paperbanana

Open source implementation and extension of Google Research’s PaperBanana for automated academic figures, diagrams, and research visuals, expanded to new domains like slide generation.

PythonPyPIMIT Licenseacademic-researcharxiv
2.2k321
jimmc414/onefilellm

Specify a github or local repo, github pull request, arXiv or Sci-Hub paper, Youtube transcript or documentation URL on the web and scrape into a text file and clipboard for easier LLM ingestion

PythonPyPIMIT Licensearxivgithub
2k176
papis/papis

Powerful and highly extensible command-line based document and bibliography manager.

HTMLGNU General Public License v3.0bibtexpublications
papis.readthedocs.io/en/latest/
1.7k123