Releases: Sefaria/Sefaria-Project
Releases · Sefaria/Sefaria-Project
helm-chart-0.63.8-remove-migrate-from-startup.1
helm-chart-v0.63.8-remove-migrate-from-startup.1 (2024-07-03)
helm-chart-0.63.8-possible-fixes-for-migrate.2
helm-chart-v0.63.8-possible-fixes-for-migrate.2 (2024-07-03)
helm-chart-0.63.8-possible-fixes-for-migrate.1
helm-chart-v0.63.8-possible-fixes-for-migrate.1 (2024-07-03)
v5.106.2
helm-chart-0.64.0-migration-init.2
helm-chart-v0.64.0-migration-init.2 (2024-06-27)
helm-chart-0.64.0-migration-init.1
helm-chart-v0.64.0-migration-init.1 (2024-06-27)
v5.106.1
v5.106.0
5.106.0 (2024-06-26)
Features
- add many non-bulk functions to complement bulk functions (10300f3)
- expose linker from library class (facd588)
- finalize bulk_link function (1797888)
- linker: add functions to linker_index_converter.py that will aid adding new indexes to linker. (0eaab15)
- linker: add param to delete existing term (872d36f)
- linker: all control over what type of entities are returned from linker (27d5db7)
- linker: allow for multiple topics to match (899e59e)
- linker: basic NamedEntityResolver (0005b82)
- linker: fully support all the attributes of ArrayMapNode that can alter how the mapping should be made. (537ffed)
- linker: fully support traversing ArrayMapNode children (588c92b)
- linker: initialize namedEntityResolver in library (f8c915f)
- linker: linker API now works in both English and Hebrew. (86a15c9)
- linker: make rawref a subclass of named entity (c212a58)
- linker: make struct node in alt structures be new AltStructNode. (7761eb2)
- linker: split input by newline which seems to improve models performance (f350698)
- linker: split input by newline which seems to improve models performance (2a2d892)
- linker: update tokenizer to include more punctuation. (873c496)
- scaffolding for linker being central class to handle refs and named entities (f022c23)
Bug Fixes
- linker: add alignment_mode to allow for differentiation in how ranges are dealt with in NormalizerComposer (1976dde)
- linker: add lots of dashes to tokenizer (b84aeeb)
- linker: add missing parameter (48186fd)
- linker: Allow citations like Parshat Vayikra 2:3 (0a00e1d)
- linker: allow topics map to be empty (f1676ab)
- linker: be more lenient with span alignment (6d9e834)
- linker: change build_ref_resolver() to build_linker() (a99a7e2)
- linker: change field name from raw_ref to raw_entity to match ResolvedRef's field name (8655a6a)
- linker: change parameter order to match the way we're passing parameters. (a60f866)
- linker: curr_removed can also be negative (feb1f95)
- linker: dont allow for partial overlaps when merging. (fce8d94)
- linker: dont fallback on re now that re2 is required (4a8373b)
- linker: dont pull context if previous non cts ref is ambiguous (658b363)
- linker: dont reset ibid history every paragraph for bulk resolution (b7973b0)
- linker: dont say match is unique enough if it hasn't increased the search by at least one word. (738c93c)
- linker: early return if context is a complex text (a10d8ca)
- linker: find text beyond element boundaries by treating all whitespace as the same (ea721dd)
- linker: find text beyond element boundaries by treating all whitespace as the same (d95462c)
- linker: fix bugs with outputting named entities. (2ec609b)
- linker: fix early return if context is a complex text (10df8a7)
- linker: fix how parts are offset in link_by_paragraph() (8d46aa4)
- linker: fully support ref being None in ResolvedRef. In addition, better logic for removing duplicate refs. (befd778)
- linker: ignore certain term slugs for resetting ibid history (df270cb)
- linker: improve ability of ner resolver to resolve topics (bd7c331)
- linker: keep track of substitution end indices which seem to be critical to get normalization mapping to work in all cases (5c97e57)
- linker: make https optional in WebPage.normalize_url so that the function is more reusable (dd32508)
- linker: make span alignment more flexible (c230512)
- linker: map normalized indices to unnormalized as last step (forgot to do that). (5cef34b)
- linker: more precisely remove duplicate matches at book level. (b65108f)
- linker: normalize the URL before querying for a domain name. (b07d4f5)
- linker: only add CORS headers when debugging on localhost (e0a45eb)
- linker: properly collect all named entities from linkedDocs (10178b5)
- linker: remove extra param 'self' (c15ee64)
- linker: revert linker API so it still uses v2 results for English. (08bbf82)
- linker: use bulk_link to improve speed. better splitting regex that matches spacy's tokenizer. (5e6bb95)
- linker: WIP allow perek to be child of parsha node. still need to figure out how to do subref on parsha. (1ff1690)
- remove AmbiguousNamedEntity which isnt currently needed (fdb9e80)
- remove special casing for normalizer. (a567d4e)
- search bar: enable search with input from virtual keyboard (c60f166)
- searchbar: cleaner cod...
v5.105.4
v5.105.3
5.105.3 (2024-06-24)
Bug Fixes
- admin editor: only use Sefaria.normRef() which also does encodeURIComponent() (c4b08ce)
- admin editor: refactor usages of requestWithCallback() to use Sefaria.adminEditorApiRequest. (eb396e2)
- sidebar toc: ignore section ref for depth 1 nodes (fed4e75)
- topic: encode ref in case it has question marks or other problematic characters for a URI (8459910)
- topics: overwrite RefTopicLink._sanitize. (d4f4379)