Releases: dginev/LaTeXML-Plugin-Cortex
LaTeXML 0.8.8 and ar5iv 2024
LaTeXML 0.8.7 and arXMLiv 2022
This release marks the state of the plugin used to generate the Full Run over arXiv's sources upto December 2022.
It uses the first release candidate for LaTeXML 0.8.7.
2.0.0
Stabilized harness, worker and Dockerfile for the arXMLiv 2021 dataset conversion run.
Aside: conversion progress tracked (December 2021) here:
https://corpora.mathweb.org/corpus/arxmliv/tex%5Fto%5Fhtml
LaTeXML 0.8.5 and arXMLiv 10.2020
This release tracks the 0.8.5 release of LaTeXML, and will be used for the initial run for the 2020 version of the arXMLiv corpus ("arXiv as HTML5").
LaTeXML 0.8.4 for arXMLiv 08.2019
Various improvements for the newest set of arXiv runs, and the 0.8.4 release of 0.8.4
Robust to downtime
This release anticipates lack of connectivity to the main zmq router that dispatches the CorTeX tasks each worker consumes.
There are a variety of cases where this can become an issue, from high load / breakage on the central router server, to actual ISP downtime leading to connections going dark.
The release adds a range of robustness checks, including a debounced retry on failure to fetch a new task. It also ensures a clean reboot to any worker that fails to receive the next task after retrying after 2,4,8,16,32 and 64 seconds debounced from each failure.