Skip to content

Release 2.2.0

Choose a tag to compare

@thiswillbeyourgithub thiswillbeyourgithub released this 04 Oct 18:24
· 1734 commits to main since this release

Commits since last release

  • [0894463] bump version 2.1.8 -> 2.2.0

  • [78ab333] better import

  • [4e490a4] rename WDoc to wdoc

  • [d72b238] rename summary.py to summarize.py

  • [1a1437e] docs: mention that faiss score fn modification changes the range of similarity

  • [299f420] typo

  • [c7efbdd] fix: actually relevancy is a similarity so goes from -1 to +1

  • [cc152d1] fix: imports were broken

  • [751baa0] docs: simpler install url

  • [4042226] update zsh completion

  • [8e92beb] new: add shell completions for bash and fish

  • [b54e1fb] docs: update completion

  • [515a273] fix: generating completions

  • [4054576] revert: forget idea about lazy loading batch loadere

  • [f0c384a] untested: use lazy loading for batchloader

  • [6e3a05b] fix: arg parsing

  • [3762bb3] fix: imports

  • [ec0b8bb] minor: fix

  • [e37350b] breaking: ename cache folders

  • [4fbb396] docs: fix incorrect help command

  • [3465a2b] enh: check env variable case insensitively

  • [193e595] major: rename project from WDoc to wdoc Lowercase is better and more humble

  • [19006a5] fix: crash when no arguments given

  • [e7a9e57] docs: remove a print

  • [cd75e07] docs: clarification

  • [39dbc3e] minor: rework the way init.py of subdirs are used

  • [6e7aa2f] enh: when calling --help, use lazy loading for faster documentation

  • [a536df5] feat: add a new env variable to enable import tricks for a way faster import time

  • [c25aeb4] fix: entrypoints It turns out that the way I set my entry points made my init called
    twice, which was adding an extra delay when calling --help and probably
    other circumstances. Thanks to claude-sonnet I was able to figure out
    how to refactor init.py and main.py to a more standard usage.
    Some changes will probably be fixed later as they create bugs as it
    required changing a lot of moving parts

  • [be6b4c6] docs: dont mention lazy imports as a feature

  • [7c1e8cd] docs: add top level docstring to env

  • [cc94309] docs: better parse_file documentation

  • [afedd6c] minor: set debug and verbose flag as arg to wdoc if parse_file called

  • [272e179] minor: md_printer should dedent message

  • [d8b4b53] minor: use Literal for some env types

  • [b6891b0] docs: typo

  • [d20d3df] minor: use Literal for typing of task

  • [646db3c] todo: mention to start using unit test

  • [81c0cb6] minor: init phrasing

  • [8136606] potentially better main

  • [5df363e] simpler and faster way to load flags

  • [213bc02] minor: docstring for wdoc_parse_file

  • [d79a140] bump version 2.1.7 -> 2.1.8

  • [10446d9] fix: better way to catch exceptions when using filetype online_media

  • [82caa42] fix: the exception cause was not opened in the debugger

  • [6657124] minor: better exception printer

  • [8f251a7] minor: info print

  • [a076e50] minor: dont change sharedmem if not multithreading anyway

  • [d141787] docs: clarificaiton

  • [001ddd3] new: if using wdoc_parse_file set loading_failure to crash by default

  • [18b20bf] minor: add comment

  • [5fd87c7] add new crash earlier if all documents failed to load

  • [ba7947b] fix: debug in parse_file now also opens debugger like in wdoc

  • [e4f7a97] fix: debug and verbose args were missing in parse_file

  • [6461252] set max token for semantic batching to 500 by default instead of 750