next - TBA
- ...
1.15.0 - 2023-11-17
- Topic modelling - Reorder methods by relevance (#1020)
- Replace pdfminer3k with pypdf (#994)
- Topic Modelling - Do not recompute when spin value doesn't change (#1019)
- Annotator - Change epsilon step and limits (#1009)
- LDAVis - Fix fail after None data (#1008)
- Corpus - remove dictionary and fix wrong types count on subsampled corpus (#990)
- Corpus - Open file dialogue at previous location (#1002)
1.14.0 - 2023-08-23
- Annotator - Add documentation (#989)
- Ontology - Enable insert in ontology with multiple roots (#940)
- Ontology: Fix 'remove word' bug (#998)
- Keywords - Handle connection error (#992)
- Twitter - Add note about plan requirements/changes to the documentation (#996)
- Reimplement concave hull (#849)
- Statistics - Fix settings (#993)
- Word Cloud - Update wordcloud javascript library (#988)
- Small refactoring (found while transtaing) (#985)
- Statistics - Fix layout (hidden + button on Windows) (#984)
1.13.1 - 2023-05-11
- import_documents - replace lxml xml parser with ElementTree (#973)
1.13.0 - 2023-04-26
- Sentiment Analysis - Language from corpus (#954)
- NYTimes - add language to corpus (#926)
- Document embedding - Use language from the corpus (#953)
- Guardian - infer language and add to corpus (#925)
- Score Documents - Use SBERT embedding instead of FastText (#930)
- Wikipedia - add language to corpus (#928)
- Keywords - replace embedding with MBERT (#932)
- Remove elements with delete/backspace key (#948)
- Semantic Viewer - show document when no words at the input (#933)
- PubMed - add language to corpus (#927)
- Score Documents - enable matching n-grams (#935)
- Create Corpus - add language to corpus (#924)
- Twitter - add language to corpus (#921)
- Import documents - language dialog and language guessing (#918)
- Add language to corpus (#916)
- Corpus - Fix contexts to be compatible between sessions (#966)
- Keywords: Fix selection and use idClicked instead of buttonClicked (#965)
- Ontology - Show labels instead of names for imported ontolgies (#936)
- Corpus widget - preserve corpus's preprocessing (#950)
- Pubmed - replace deprecated extend_corpus (#949)
- Corpus - preserve name in extend_attributes (#937)
- Make widgets PyQt6 compatible (#929)
- Corpus: Unpickle corpus without language (#919)
- Score Documents - cast bool scores to float (#913)
- Annotator: Invalidate clusters on setting change (#910)
1.12.0 - 2022-10-06
- Ontology - remove cache and other fixes (#896)
- VectorizationComputeValue - fix unpickling old pickles (#904)
- Keywords/Score Documents - fix ctrl/cmd selection (#902)
- Word Enrichment - fix PyQt6 incompatibility and sort imports (#901)
- VectorizationComputeValue - remove wrongly set "original" variable (#900)
1.11.0 - 2022-08-24
- Replace exec_ with exec and fix deprecations (#887)
- Ontology - fix cannot be called from a running event loop error (#882)
1.10.0 - 2022-07-08
- Corpus: fix ngrams_corpus (#871)
1.9.0 - 2022-06-21
- Import Documents - fix metadata matching (#870)
- Word cloud - add type to the selected words output (#868)
1.8.1 - 2022-06-20
- Score Documents - adapt to the latest changes in document embedding (#866)
- Temporary proxy fix
1.8.0 - 2022-06-09
- Create corpus - new widget (#854)
- Computation in separate thread for base vectorizer; use base vectorizer for embedding (#852)
- Mark "Words" outputs as non-dynamic (#855)
- Corpus refactoring (#767)
- Functionalities for computing concave hull around clusters of points (#816)
- Normalize - fix unpickling for Normalizers before caching was implemented (#838)
- Credential manager dialog at import (#857)
- NLTK - use proxy when proxy adresses entered (#853)
- Semantic search: fix wrong output when response is None (#826)
- Annotate documents: update to work with the latest changes in keywords (#845)
- Update embedder callbacks (#842)
- Keywords: Fix sending data while running (#836)
- Preprocess Text: set highest absolute frequency (#807)
- Wikipedia: make widget useable again (#825)
- Concave hull: fix cases when all points inline (#821)
- Keywords: Always invoke preprocessor call (#815)
1.7.0 - 2022-04-08
- Twitter: adapt to APIv2 (#780)
1.6.2 - 2022-01-10
- Solved numpy 1.22 incompatibility
1.6.1 - 2021-11-28
- Semantic Viewer: New widget (#748)
- score-documents: handle document titles with newlines (#754)
1.6.0 - 2021-11-23
- Semantic Search: increase maximal allowed chunk size, fix error when timeout (#752)
- Topic Modeling: Base selection style on itemdelegates (#751)
- Topic - remove topic computations in chunk which results in poor topics (#750)
- Semantic Search - fix callback to return proportions (#747)
- BoW: use training weights on test data (#745)
- OWScoreDocuments: Ensure unique names on output (#744)
- Sparse2CorpusSliceable: add support for np.ndarray as key (#737)
- Store ngrams_corpus correctly (#729)
1.5.3 - 2021-10-08
1.5.1 - 2021-09-16
- Fixing bug with non-working Gensim library
1.5.0 - 2021-09-13
- Keywords: Add 'Embedding' scoring method (#666)
- OW Corpus Viewer: Add annotated corpus output (#672)
- Concordance - search in the separate thread (#668)
- OWPreprocess Text: add option to filter on POS tags (#679)
- Topic Modeling: Add topic evaluation scores (#687)
- OWStatistics: Add new statistics method (#676)
- Import Documents: Add conllu reader (#675)
- OWPreprocess Text: Add Lemmagen normalizer (#708)
- normalize: speedup preprocessing with caching (#709)
- Score documents: Document selection and selection output (#710)
- Import Documents: Read metas as the right type (#677)
- LDA: Add random state (#688)
- Preprocess: Filter POS tags along with tokens (#692)
- Topic Modeling: Remove tags from display of topics (#693)
- Score documents: fix word preprocessing (#707)
- Make Lemmagen lemmatizer picklable (#713)
- Document Embedder: fix default language setting (#716)
1.4.0 - 2021-05-27
- Score documents widget (#632)
- Import documents: Import from URL (#637)
- Extract Keywords: New widget (#644)
- Word List: New widget (#634)
- Fix infer text features (#645)
- Corpus fix from_numpy and from_list; modify widget to work with corpuses without text_features (#627)
- Corpus Viewer: Handle empty corpus (#628)
- Import Documents: Handle loading folder with no readable files (#626)
1.3.1 - 2021-01-22
- Revert #592 since fastText cannot be installed on Windows (#618)
1.3.0 - 2021-01-21
- Sentiment Analysis: Add SentiArt method (#605)
- Preprocess Text: update documentation (#604)
- [DOC] Sentiment Analysis: custom files and documentation (#557)
- Preprocess: Filter by absolute frequency (#601)
- Corpus: extend_attributes retain preprocessing (#599)
- owwordcloud: Work with empty token list (#588)
- Simhash: Fix error when hash function is None (#589)
1.2.0 - 2020-10-12
- Document embedders: additional languages (#565)
- Corpus Viewer: Output selected data and memorize selection (#562)
- List files that are not loaded (#560)
- Corpus - from_table: keep text feature when renamed (#585)
- Corpus - extend attributes: Fix error with renamed text feature (#574)
- Import documents: normalize imported text and file names (#568)
- Corpus to network (#559)
- Corpus: fix deprecated use of array (#563)
- Preprocess: Retain corpus ids (#553)
1.1.0 - 2020-08-07
- Preprocess: Retain corpus ids (#553)
1.0.0 - 2020-06-12
- Refactor preprocessors (#506)
0.9.1 - 2020-05-05
0.9.0 - 2020-04-29
- Word Enrichment: compute in separate thread (#492)
- Bag of Words: option to show bow features (#499)
- Word Cloud: threaded (#502)
- Corpus: remove unnecessary empty values (#505)
- Statistics widget: new widget for feature construction (#503)
- Document Embedding widget: word embedding for documents (#504)
- Corpus to Network widget: widget for computing networks from documents (#509)
- Word Cloud: don't show zero weights (#501)
0.8.0 - 2020-02-01
- Add Concordance output as text feature. (#476)
- Corpus: add Title dropdown. (#481)
- Word Cloud: show bow weights (#486)
- Topic Modeling: change output to emulate PyLDAvis (#483)
- Corpus: make unique titles (#490)
- Word Cloud: major rewrite (spacing, tests, bugfixes) (#493)
- Remove typing as dependency (#475)
- Topic Modeling: select topic is schema-only setting (#478)
- Include widgets in coverage (#487)
- Corpus: use DomainContextHandler (#491)
0.7.4 - 2019-11-11
- Remove ufal.udpipe-temp dependency (#467)
- Remove redundant webview in Word Cloud (#464)
- Rename GeoMap to Document Map (#458)
- Fix stopword filtering (#456)
- Pickling UDPipe models (#447)
- Remove validate_email dependency (#442)
0.7.3 - 2019-07-09
- Fix broken import (#448)
0.7.2 - 2019-07-02
- Compatibility with split Orange Canvas (#445)
0.7.1 - 2019-06-18
- Remove google-compute-engine dependency (#439)
0.7.0 - 2019-06-14
- Word Enrichment: FDR imported from Orange (#416)
- OWPreprocess: autocommit when changing n-grams (#436)
- Import Documents: sanitize pdfs (#435)
- OWPreprocess: fix Stanford model loading (#419)
- Corpus Viewer: mark filtered text with Python (#408)
- Topic Modelling: HDP shows no topics (#413)
- Fix PyQt to 5.11 (#410)
- Guardian: properly display Unicode characters (#406)
0.6.0 - 2018-12-06
- Duplicate Detection: output corresponding duplicate cluster (#379)
- Update Twitter API calls to retrieve extended tweets (#383)
- Twitter: fix retrieving author timeline (#389)
0.5.1 - 2018-09-27
0.5.0 - 2018-09-14
- Concordance: output concordances (#371)
- UDPipe Lemmatizer (#367)
- OWWordCloud: Add the word cloud image to the report (#364)
- Sentiment Analysis: add Slovenian language (#366)
0.4.0 - 2018-07-23
- Improved report functionality (#357)
- OWPreprocess: Add choice in FilteringModule for All Files (*) (#334)
- Concordance: clear selection on changing word (#353)
- CorpusViewer: Break long urls (#310)
- OWPreprocess: no stopword files on Windows (#355)
- Bag of words: Use vectorized 'BINARY' local weighting (#342)
- Corpus: X and Y cast as float (#330)
- Bag of words: work on document with no tokens (#356)
0.3.0 - 2017-12-05
- Corpus & Bow: Improve sparsity handling according to Orange>=3.8.0 (#281)
- Download NLTK data asynchronously (#304)
- Add Table Input to Corpus (#308)
- Corpus: Remove text features which not in metas (#325)
- Topic Modelling: Do not call get all topics table when no corpus (#322)
- Concordance: Selection settings (#249)
- Preprocess: Use default tokenizer when None (#294)