CollateX is a collation tool.
CollateX is a software to
read multiple (>= 2) versions of a text, splitting each version into parts (tokens) to be compared,
identify similarities of and differences between the versions (including moved/transposed segments) by aligning tokens, and
output the alignment results in a variety of formats for further processing, for instance to support the production of a critical apparatus or the stemmatic analysis of a text’s genesis.
Free software: GPLv3 license
Partially non-progressive multiple-sequence alignment
Multiple output formats: alignment table, variant graph
Near matching (optional)
Supports Python 3
Supports unicode (Python 3 only)
How to install:
Mac/Linux: pip install collatex
if you don’t have pip installed, install it first with: easy_install pip
For near matching functionality python-levenshtein C library is required. Install it with (on Mac OS X and Linux): pip install python-levenshtein. Windows users may need a precompiled binary distribution of this library if they want to use near matching.
from collatex import * collation = Collation() collation.add_plain_witness("A", "The quick brown fox jumps over the dog.") collation.add_plain_witness("B", "The brown fox jumps over the lazy dog.") alignment_table = collate(collation) print(alignment_table)
When running from the command shell, run the example script with:
Create documentation for CollateX Python
TEI output writes “t” values instead of “n” values
TEI output uses minidom instead of etree
TEI output uses same namespaces and wrapper as CollateX Java
Add “csv” and “tsv” output options
Use graphviz Python bindings instead of PyGraphviz for Windows compatibility
Update networkx compatibility from 1.11 to 2.1
Replace pygraphviz bindings with graphviz for Windows compatibility
Update near matching to add near-matching edges and adjust rank in SVG output
New version of the alignment algorithm (which we call the MatchCube approach) to reduce order effects during multiple witness alignment.
Added the new SVG renderer which uses the pygraphviz bindings instead of the graphviz bindings.
Thanks to David J. Birnbaum for the patch.
Fixed the CalledProcessError bug that the previous renderer caused when used with Python 3.
Added the ability to use output=”svg_simple” next to output=”svg”. The “svg_simple” option gets you the n-property
based graph, so just the normalized version of the tokens, which will hide any variation in the t-property.
Thanks to Joris van Zundert for the patch.
Changed the colour scheme of the “html2” output option, to aid those with Red/Green colour-blindness.
Thanks to Melodee H. Beals for the patch.
Bug fix release
Fixed a bug in the new near match functionality that would cause tokens to go missing in the alignment table.
Thanks to Torsten Hiltmann for reporting it and providing a test case.
Official release for the Dixit code and collation workshop in Amsterdam
This release contains the new near match functionality implemented as a post process after alignment. Same as RC1.
It also contains the multiple short witnesses bugfix done in 2.0.1
Bug fix release for the Dixit code and collation workshop in Amsterdam
Fixed index out of range bug when multiple very short witnesses (= one token) were collated
Disabled debug statements for near matching
New near match functionality, implemented as a post process after alignment.
First official release for the Dixit code and collation workshop in Amsterdam
Added XML as an output format
Added TEI parallel segmentation as an output format
Tokenizer: retain whitespace in the t-property of preceding token
Witness: added normalization: strips whitespace
Merged old collate_pretokenized_json() function into collate() function
JSON output contains full JSON representation of the tokens
Enabled segmentation support for all input formats, and for SVG output
Enhanced SVG output to include “n” value and all “t” values of JSON input
JSON output is raw Unicode, instead of escaped characters
Test suite updated
Rename of TokenIndex.py was not in effect in the uploaded files. Fixed now.
Renamed TokenIndex.py module to tokenindex.py to follow conventions.
Moved all the block and suffix, LCP interval code to new class TokenIndex.
Added output option ‘html2’ for colored alignment table rendering.
Fix a bug that was caused by the fact that a dash was stored in empty cells of the AlignmentTable. Now None is stored (this resolved a TODO). Plain text and HTML rendering of the table render a dash for empty cells. JSON output now returns null for empty cells. Fixes bug when a token with a dash in the content was screwing the rendering of the alignment table (caused of by one errors).
Further improved blockification of witnesses.
Added properties_filter option to enable users to influence matching based on properties of tokens.
Improved blockification of witnesses.
Added SVG output option to the collate function. For this functionality to work the graphviz python library needs to be installed.
Bug-fix: collate_pretokenize_json function should not re-tokenized the content. Thanks to Tara L. Andrews.
Allow near-matching for plain as well as for pre-tokenized content. Thanks to Tara L. Andrews.
Added HTML option to collate function for the output as an alignment table represented as HTML.
Added support for Unicode character encoding
Ported codebase from Python 2 to Python 3
Separated IPython display logic from functional logic. No longer will the collate function try to determine whether you are running an environment that is capable of display HTML or SVG.
Added near matching option to collate function.
Added variant or invariant status to columns in alignment table object and JSON output.
Added experimental A* decision graph search optimization.
Added WordPunctuationTokenizer (treats punctuation as separate tokens).
Combined suffix array and edit graph aligner approaches into one collation algorithm.
Fixed handling of segmentation parameter in pretokenized JSON function.
Added Windows support. Thanks to David J. Birnbaum.
Fixed handling of IPython imports.
Added JSON output to collate method.
Added option to collate method to enable or disable parallel segmentation.
Added table output to collate_pretokenized_json method, next to the already existing JSON output.
Cached the suffix and LCP arrays to prevent unnecessary recalculation
Fixed handling of empty cells in JSON output of pretokenized JSON.
Fixed compatibility issue when rendering HTML or SVG with IPython 2.1 instead of IPython 0.13.
Corrected RST syntax in package info description.
Added pretokenized JSON support.
Added JSON visualization for the alignment table.
Fixed imports in init.py, “from collatex import *” now works correctly.
Added IPython HTML support for alignment table.
Added IPython SVG support for variant graph.
Added convenience constructors on Collation object.
Added horizontal layout for the alignment table visualization, next to vertical one.
Removed max 6 witness limit in aligner, now n number of witnesses can be aligned.
Added transposition detection.
Added alignment table plus plain text visualization.
Added collate convenience function.
First release on PyPI.
First pure Python development release of CollateX.
New collation algorithm, which does non progressive multiple witness alignment.
Release history Release notifications | RSS feed
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Hashes for collatex-2.2-py2.py3-none-any.whl