Final Release

This release stands as a final version.

The original version of this software (https://github.com/MattWellie/Final-Referencer) was a project carried out at the West Midlands Regional Genetic Laboratory (WMRGL), Birmingham, UK. At the time of writing, the staff at WMRGL had to generate MS-Word documents, one per gene tested as part of service, annotating on:

• exon boundaries
• base position & numbers, and the encoded amino acid for each triplet codon
• flanking intron
• available primers & batch numbers

This was a long and error-prone process, requiring manual work to be manually checked prior to a gene becoming part of routine service.

My service improvement project was to take the transcript reference files (GenBank or LRG format) and automate the generation of a complete, typeset document. The original version was implemented in service at WMRGL, requiring a specific VM to be run in a specific hypervisor, on a specific desktop in the lab, to satisfy the python dependencies and LaTex runtime).

A few years later I returned, re-wrote the application in Python3 syntax, and had a crack at containerising the process. For reasons lost to time, the strategy I went with was a 2-container chain:

• the first contained the python application, generating a raw .tex file
• the second contained a TexLive installation, and generated the final PDF
• the two containers were connected by a Makefile, which handled mounting of final and intermediate files between the two containers

Now, in 2026, and with the help of Claude, this has had a complete rewrite:

• The refparse software is now distributed and installed as a package, with one pyproject.toml containing all dependencies
• Instead of multiple Docker containers there's one - a TexLive base and a package install on top
• Makefile is out, now there's a simple script to handle the input/output/cleanup process
• The code has had a bit of a spring clean, removing some redundant/silly methods and attributes, and updating some syntax

I have no use for this project other than mild vanity, and there is no need for further improvement/updates. This is probably the final update to the package. It's been fun. I sincerely hope WMRGL no longer uses these reference sequence files.