Skip to content
This repository has been archived by the owner on Dec 13, 2021. It is now read-only.


Folders and files

Last commit message
Last commit date

Latest commit



11 Commits

Repository files navigation

Note: texrex is no longer maintained, will re-emerge as PyRex

As of December 2021, I've officially dropped support/maintenance for texrex. It's unnecessarily complex due to some quirks of working with Pascal in the 2020s. Also, as much as I like Pascal and the FreePascal compiler, the FreePascal ecosystem is outright annoying. Instead, I'll be working on PyRex, a modular Python 3 reimplementation of texrex with a focus on analysing CommonCrawl data. If things go well, NLP modules for German and English (such as currently available in COWTek) might be added.

The ClaraX crawler will not be reimplemented in Python, but I'm planning to include some bias-removal tools in PyRex if CommonCrawl data allows for a good guess of at least some properties of the web graph.

The following is the old texrex README…

texrex web page cleaning system

Web document cleaning/crawl file processing tools written by Roland Schäfer for the COW project.

Current milestone: texrex-behindthecow (2015)

texrex is a free software for processing data files from crawls and turn them into a corpus of web documents. Currently, it is limited to reading ARC and WARC files, but other input modules can be developed quickly.

It performs the following processing steps:

  • read WARC or ARC files document by document
  • filter perfect duplicates using a Bloom filter
  • strip HTML, scripts, stylesheets
  • extract meta information from crawl headers
  • normalize encodings to UTF-8 (using ICU), optionally treating all ISO-8859-1 input as Win-1252
  • convert all HTML entities to appropriate codepoints (including rogue Win-1252)
  • detect, remove, and/or annotate boilerplate blocks using state-of-the-art classifier (
  • assess the text quality of the documents by looking at frequencies of short frequent word (
  • create w-shingling document fingerprints and filter near-duplicate documents
  • perform in-document deduplication (remove repeated paragraphs, insert a backreference to first copy)
  • perform additional normalization (e.g., reduce diverse Unicode dashes and hyphens to the basic codepoint)
  • write standard-compliant XML output
  • add server IP geolocation meta information (country, region, city – currently based on GeoLite)

Technologically, the main features of texrex are:

  • written in FreePascal (Object FPC mode)
  • licensed under permissive 2-clause BSD license
  • uses multi-threading for single-machine parallelization
  • uses simple INI files to configure processing jobs for the main tool
  • can be run in the background, using an included IPC client to control the process
  • depends only on two additional libraries: ICU and FANN

New tools included since texrex-neuedimensionen (June 2014):

  • HyDRA hard hyphenation remover
  • rofl tool to fix run-together sentences

Papers to read about texrex and related technology:

This repo was moved to GitHub from SourceForge r. 662 on May 1, 2016. For reference:


texrex web page cleaning & ClaraX random walk crawler







No releases published


No packages published