Skip to main content

A traceable LaTeX flattener

Project description

flachtex

Tools (e.g. cktex, YaLafi, TeXtidote) for analyzing LaTeX-documents often only work on single files, making them tedious to use for complex documents. The purpose of flachtex is to preprocess even complicated LaTeX-documents such that they can be easily analyzed as a single document. The important part is that it also provides a data structure to reverse that process and get the origin of a specific part (allowing to trace issues back to their source). While there are other tools to flatten LaTeX, they all are neither capable of dealing with complex imports nor do they allow you to trace back to the origins.

Notable features of flachtex are:

  • Flattening of LaTeX-documents with various rules (\include, \input, \subimport ,%%FLACHTEX-EXPLICIT-IMPORT[path/to/file]...).
  • Any character in the output can be traced back to its origin.
  • Diff-friendly formatter for version control (one sentence per line, environment indentation).
  • Remove comments.
  • Remove \todo{...}.
  • Remove highlights of \usepackage{changes}. (This substitution is actually more robust than the one supplied with the package.)
  • Substitute commands defined by \newcommand.
  • A modular design that allows to add additional rules.

Installation

flachtex is available via pip: pip install flachtex.

Example

Let us look on a quick example that shows the power of the tool. We have a LaTeX-document consisting of three files.

main.tex

\documentclass{article}
\usepackage[utf8]{inputenc}
\usepackage{amsmath,amssymb,amsfonts,amsthm}
\usepackage{todonotes}
\usepackage{xspace}

\newcommand{\importantterm}{\emph{ImportantTerm}\xspace}

%%FLACHTEX-SKIP-START
Technicalities (e.g., configuration of Journal-template) that we want to skip.
%%FLACHTEX-SKIP-STOP

\begin{document}

\section{Introduction}

\todo[inline]{This TODO will not be shown because we don't want to analyze it.}

Let us use \importantterm here.

% including part_a with 'input' and without extension
\input{./part_a}

% including part_b with 'include' and with extension
\include{./part_b.tex}

\end{document}

part_a.tex

\subsection{Part A}

This is Part A. We can also use \importantterm here.

part_b.tex

\subsection{Part B}
And Part B.

flachtex can create the following output for us that is much easier to analyze.

\documentclass{article}
\usepackage[utf8]{inputenc}
\usepackage{amsmath,amssymb,amsfonts,amsthm}
\usepackage{todonotes}
\usepackage{xspace}

\newcommand{\importantterm}{\emph{ImportantTerm}\xspace}

\begin{document}

\section{Introduction}

Let us use \emph{ImportantTerm}\xspace here.

\subsection{Part A}

This is Part A. We can also use \emph{ImportantTerm}\xspace here.

\subsection{Part B}
And Part B.

\end{document}

(currently, flachtex will actually add some redundant empty lines, but those usually do no harm and could be easily eliminated by some simple postprocessing.)

Usage

CLI

flachtex comes with a simple CLI, if you don't want to use it via Python.

usage: flachtex [-h] [--to_json] [--comments] [--attach] [--changes]
              [--changes_prefix] [--todos] [--newcommand] [--format]
              [--indent N] [--no-expand]
              path

flachtex: Traceable LaTeX flattening.

positional arguments:
  path              Path to main.tex

options:
  -h, --help        show this help message and exit
  --to_json         Return a json.
  --comments        Remove comments.
  --attach          Attach sources to json.
  --changes         Replace the commands of the changes package.
  --changes_prefix  Use the prefix option in changes.
  --todos           Remove todo-notes.
  --newcommand      Automatically substitute custom commands.
  --format          Format output with one sentence per line for diff-friendly results.
  --indent N        Indent environments with N spaces (default: 0, disabled). Use with --format.
  --no-expand       Don't expand \input and \include commands (only format the main file).

Quick Examples:

# Flatten and format for version control
flachtex --format --indent 2 main.tex > output.tex

# Format without flattening (keep \input commands)
flachtex --no-expand --format --indent 2 main.tex > output.tex

# Full pipeline: flatten, remove comments, format
flachtex --comments --format --indent 2 main.tex > output.tex

See docs/formatter.md for complete formatter documentation.

Python

from flachtex import Preprocessor, remove_comments
from flachtex.rules import TodonotesRule

# basic usage
preprocessor = Preprocessor("/path/to/latex_document/")
preprocessor.skip_rules.append(TodonotesRule())  # remove todos
doc = preprocessor.expand_file("main.tex")

# remove the comments (optional)
doc = remove_comments(doc)

# The document can be read as a string (but contains also further information)
print(f"The process LaTeX-document is {doc}")

# Get the used files
for f, data in preprocessor.structure.items():
    print(
        f"Used file {f} which contains the content '{data['content']}' and includes"
        f" the files {data['includes']}."
    )

# query origin
origin_file, pos = doc.get_origin_of_line(line=3, col=6)
print(
    f"The seventh character of the fourth line origins from file {origin_file}:{pos}."
)
origin_file, pos = doc.get_origin(5)
print(f"The sixth character  origins from file {origin_file}:{pos}.")

Features

Diff-Friendly Formatter

flachtex includes an optional formatter that makes LaTeX documents more suitable for version control:

One sentence per line:

  • Splits text at sentence boundaries (periods, question marks, exclamation marks)
  • Intelligently handles abbreviations (Dr., et al., i.e., etc.)
  • Preserves decimal numbers (3.14)
  • Keeps comments with their sentences

Environment indentation:

  • Configurable indentation (default: 2 spaces)
  • Progressive indentation for nested environments
  • Excludes verbatim-like environments (verbatim, lstlisting, minted)
  • Document-level environments (document, abstract) don't cause indentation

Blank line normalization:

  • Reduces excessive blank lines (3+) to one blank line
  • Removes leading/trailing blank lines
  • Preserves paragraph structure

Two main use cases:

  1. Format only (without flattening):

    flachtex --no-expand --format --indent 2 main.tex
    

    Keeps \input commands intact, formats a single file.

  2. Full pipeline (flattening + formatting):

    flachtex --format --indent 2 main.tex
    

    Expands all includes, then formats the result.

See docs/formatter.md for detailed documentation.

Flatten LaTeX-documents

Currently, flachtex supports file inclusions of the following form:

% native includes/inputs
\include{path/file.tex}
\input{path/file.tex}

% subimport
\subimport{path}{file}
\subimport*{path}{file}

% manual import
%%FLACHTEX-EXPLICIT-IMPORT[path/to/file]
%%FLACHTEX-SKIP-START
Complex import logic that cannot be parsed by flachtex.
%%FLACHTEX-SKIP-STOP

Path Resolution

flachtex will first try to resolve the inclusion relative to the calling file. If no file is found (also trying with additional ".tex"), it tries the document folder (cwd) and the folder of the root tex-file. Afterwards, it tries the parent directories.

If this is not sufficient, try to use the %%FLACHTEX-EXPLICIT-IMPORT[path/file.tex] option.

Extending the tool

flachtex has a modular structure that allows it to receive additional rules or replace existing ones. You can find the current rules in ./flachtex/rules.

It is important that the matches do not overlap for SkipRules and ImportRules. For efficiency, flachtex will first find the matches and only then includes the files. Overlapping matches would need a complex resolution and my result in unexpected output. (It would not be too difficult to add some simple resolution rules instead of simply throwing an exception).

Usage for cleaning 'changes' of '\usepackage{changes}'

The changes-package is helpful for highlighting the changes, which is a good practice, e.g., when writing journal papers (which usually have to go through one or two reviewing iterations). These can of course disturb automatic language checkers and they have to be removed in the end. The script that is attached to the original package unfortunately is not compatible with some usages (e.g., comments can lead it astray). flachtex is capable of removing the highlights done with changes in a robust way. There are some nasty ways to trick it, but if you use brackets, it should work fine and independent of escaped symbols, comments, or line breaks.

Substitution of \newcommand

It is reasonably common to create your own commands with `\newcommand', e.g., for some terms which you may want to change later. If you want to analyze the tex-document, this can become cumbersome. Thus, flachtex gives you the option to automatically substitute such commands.

The primary reason I added this functionality to this tool (and not some higher level tool) is that I also saw that some people define their own \input/\include commands, which could not be imported easily without this feature.

Changelog

  • 0.7.0 Adding formatter
  • 0.6.0 Significant refactoring.
  • 0.5.0 Now will only replace \input and \include commands for which the file exists. Otherwise, it will leave the command as is. This allows you to use \input and \include commands for files that are not part of the document, e.g., for some automatically generated files.
  • 0.4.0 Support for the comments package.
  • 0.3.15 Fixes Issue #8
  • 0.3.14 Bugfix by Nutron2112
  • 0.3.13 improves robustness of command parsing (of potentially faulty LaTeX code)
  • 0.3.12 Made parsing of non utf-8 encodings more robust. Some templates you get have very strange file encodings. You don't always convert them manually to utf-8.
  • 0.3.11 newcommand should work reliably with multiple arguments now (hopefully).
  • 0.3.10 Support for newcommand* substitution
  • 0.3.9: PEP compliance which may have created problems in environments without setuptools
  • 0.3.8: Substituting newcommands is no longer enabled by default.
  • 0.3.7: Versions got slightly mixed up. Should be fixed now.
  • 0.3.6 bugfix: Using findall instead of finditer.
  • 0.3.4 Dealing with \xspace in command substitution.
  • 0.3.3
    • FileFinder now has a default and allows to set a new root.
    • Command substitution for commands without parameters made more accurate.
    • from_json for TraceableString

This tool is still work in progress.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

flachtex-0.7.0.tar.gz (53.6 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

flachtex-0.7.0-py3-none-any.whl (39.4 kB view details)

Uploaded Python 3

File details

Details for the file flachtex-0.7.0.tar.gz.

File metadata

  • Download URL: flachtex-0.7.0.tar.gz
  • Upload date:
  • Size: 53.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for flachtex-0.7.0.tar.gz
Algorithm Hash digest
SHA256 c3e8885bb529bd8d9504f1dbec79b853cbaa0b2ceda8c0b0a9c1065d7c558dc1
MD5 005827676f4fa406eedb01dd96ab8961
BLAKE2b-256 f8dff5dcc950f0d3260888b4dae582b030574f0d028861a3119e94a2e78212b7

See more details on using hashes here.

File details

Details for the file flachtex-0.7.0-py3-none-any.whl.

File metadata

  • Download URL: flachtex-0.7.0-py3-none-any.whl
  • Upload date:
  • Size: 39.4 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.13.7

File hashes

Hashes for flachtex-0.7.0-py3-none-any.whl
Algorithm Hash digest
SHA256 5f04b4a0b517ca08cb08f789961088013184ee48ce3377b6498d666d4fa2fd74
MD5 44ba5478846cdb2e36823ff4b00dc235
BLAKE2b-256 3f2e2f615804d1d2e8654b09a39bc4e0ff8c15e1d608b501f80b9edcef8e0e46

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page