Arekit

- [x] :fire: 296

Improvements
- [x] 527
- [x] automate `NoFolding` support, easy API usage :fire: 466
- [x] :wrench: 417
- [x] :wrench: 489
- [x] :wrench: 502 because of 501 were fixed
- [x] :wrench: 510
- [x] 503
- [x] 507
- [x] remove everything related to applications and related framework if everything will be OK with the paper (0.23.1 as well)
- [x] 520
- [x] :wrench: 526

Generalization
- [x] Long delayed issue: 284
- [x] API: doc renaming 457
- [x] 522

Changes and Simplifications
- [x] :x: 517
- [x] :x: Drop support of reading grouped Opinions (491 and 492 related) [the related unit-test was optional and has been removed as well)
- [x] :x: 483
- [x] :x: 376

Minor
- [x] 521
- [x] 505
- [x] 514
- [x] 534
- [x] Make to string for `doc_now` parameter:


v0.23.1-rc-p1
Main Updates

- [x] 439
- [x] 447
- [x] fixed : 440
- [x] new: :fire: 459
- [x] moving `evaluation` module outside :wrench: 449 (new separate project)
- [x] utils: 467
- [x] universal API for proof-of-concept

[Full Changelog](https://github.com/nicolay-r/AREkit/compare/v0.23.1-rc...HEAD)


**Implemented enhancements:**

- `NativeCsvWriter` -- sync deliimiter with other CSV formatters [\486](https://github.com/nicolay-r/AREkit/issues/486)

[v0.23.1-rc](https://github.com/nicolay-r/AREkit/tree/v0.23.1-rc) (2023-06-02)

[Full Changelog](https://github.com/nicolay-r/AREkit/compare/v0.23.0-rc...v0.23.1-rc)

**Implemented enhancements:**

- `filters=[]` -- consider the case of None by default \[Paper feedback\] [\479](https://github.com/nicolay-r/AREkit/issues/479)
- `opinions=[]` -- simplify usage of API \[paper feedback\] [\478](https://github.com/nicolay-r/AREkit/issues/478)
- `BaseSerializerPipelineItem` -- required by `arekit-ss` [\476](https://github.com/nicolay-r/AREkit/issues/476)
- `Neural Network Serializer` -- `rows_provider` should be declared outside \[paper backlog/arekit\_ss project\] [\475](https://github.com/nicolay-r/AREkit/issues/475)
- Streaming -- support `JSON` output format [\474](https://github.com/nicolay-r/AREkit/issues/474)
- `RuAttitudesDocumentProvider` -- refactor to follow the structure of the rest resources [\470](https://github.com/nicolay-r/AREkit/issues/470)
- Support `None` for `get_doc_existed_opinion_func` \[user/paper feedback\] [\469](https://github.com/nicolay-r/AREkit/issues/469)
- `SynonymsCollection` -- setup default value of `iter_group_values_lists` to `[]` [\468](https://github.com/nicolay-r/AREkit/issues/468)
- `DOC_ID` column -- remove `int` type limitation [\463](https://github.com/nicolay-r/AREkit/issues/463)
- Streaming -- provide header column names for CSV [\462](https://github.com/nicolay-r/AREkit/issues/462)
- `tqdm` -- display amount of processed documents in progress-bar \[Project Gutenberg backlog\] [\461](https://github.com/nicolay-r/AREkit/issues/461)
- `OpinionCollection` -- `iter_sentiment` method is not in use anymore [\456](https://github.com/nicolay-r/AREkit/issues/456)
- `OpinionCollection` -- the case of `None` for `opinion` results in incomplete initialization [\455](https://github.com/nicolay-r/AREkit/issues/455)
- `OpinionCollection` -- `copy` method is not in use anymore [\454](https://github.com/nicolay-r/AREkit/issues/454)
- `OpinionCollection` -- consider `opinions=[]` by default in, i.e. empty collection. [\453](https://github.com/nicolay-r/AREkit/issues/453)
- synonyms.py -- is empty and might be removed \[QUICK check and fix\] [\451](https://github.com/nicolay-r/AREkit/issues/451)
- Pandas -- completely remove dependencies [\450](https://github.com/nicolay-r/AREkit/issues/450)
- `BertTextBTemplates` -- switch name to prompts [\446](https://github.com/nicolay-r/AREkit/issues/446)
- RuSentRel -- embed train and test indices in collection [\444](https://github.com/nicolay-r/AREkit/issues/444)
- SentiNEREL -- entity filter [\443](https://github.com/nicolay-r/AREkit/issues/443)
- SentiNEREL -- move from another project \[NIVTS project backlog, RuSentNE competitions\] [\439](https://github.com/nicolay-r/AREkit/issues/439)

**Fixed bugs:**

- `Network` module -- context constant has a predefined `text` value which is limited for networks only [\485](https://github.com/nicolay-r/AREkit/issues/485)
- `read_ruattitudes_to_brat_in_memory` -- case of `keep_doc_ids_only==True` causes exception [\482](https://github.com/nicolay-r/AREkit/issues/482)
- `prompt` -- object non subscriptable [\481](https://github.com/nicolay-r/AREkit/issues/481)
- `fill` -- in case of `None` rows count `tqdm` throws exception [\458](https://github.com/nicolay-r/AREkit/issues/458)
- `create_sample_provider` -- misused parameter [\445](https://github.com/nicolay-r/AREkit/issues/445)
- `CroppedBertSampleRowProvider` -- might crop with references outside of the bounds \[googletranslate-feedback\] [\440](https://github.com/nicolay-r/AREkit/issues/440)

**Closed issues:**

- Shortening to `RuSentRelOpinions.iter_from_doc` [\480](https://github.com/nicolay-r/AREkit/issues/480)
- `InputTextOpinionProvider` -- rename to `ContentsProvider` [\473](https://github.com/nicolay-r/AREkit/issues/473)
- `RuSentiFramesCollection.read` -- rename method `read_collection` to `read` \[paper feedback\] [\472](https://github.com/nicolay-r/AREkit/issues/472)
- `DocumentOperation` -- provide directory-based document provider by default \[Project Gutenberg feedback\] [\467](https://github.com/nicolay-r/AREkit/issues/467)
- Stream writing [\459](https://github.com/nicolay-r/AREkit/issues/459)
- `dist_in_sent=0` by default [\452](https://github.com/nicolay-r/AREkit/issues/452)
- Evaluation -- is not a part of the AREkit soon [\449](https://github.com/nicolay-r/AREkit/issues/449)
- Prompting -- collect base classes that allows such input processing [\447](https://github.com/nicolay-r/AREkit/issues/447)
- SentiNEREL -- move `split_fixed.txt` into the data SentiNEREL data archive. [\442](https://github.com/nicolay-r/AREkit/issues/442)
- What's new in 0.23.0 [\401](https://github.com/nicolay-r/AREkit/issues/401)

**Merged pull requests:**

- CVE-2007-4559 Patch [\412](https://github.com/nicolay-r/AREkit/pull/412) ([TrellixVulnTeam](https://github.com/TrellixVulnTeam))



\* *This Changelog was automatically generated by [github_changelog_generator](https://github.com/github-changelog-generator/github-changelog-generator)*

What's new: Globalization and Internalization


Globalization for any language is the major aspect of 0.23.0, since we annou
nce `AREnets` and `sample-transfer`
We tend to generalize some aspects in order to consider **other languages than original one** (Russian).
We introduce `CompoundEntities` which may include other entities.

Major
- [x] Nested/Compound entities support! 398
- [x] Detaching `networks` contrib module 423 -> AREnets
- [x] Appearance of transfer: https://github.com/nicolay-r/arekit-googletrans-sampler

Fixed bugs
- [x] Refactored BRAT parser, fixed bugs for other languages/collections.

Minor
- [x] 375
- [x] Internalization (435)

[Full Changelog](https://github.com/nicolay-r/AREkit/compare/v0.22.1-rc...HEAD)

**Implemented enhancements:**

- `PipelineContext` -- support `parent` contexts in case of the nested pipelines. [\433](https://github.com/nicolay-r/AREkit/issues/433)
- Idle mode -- provide such flag into main pipeline [\432](https://github.com/nicolay-r/AREkit/issues/432)
- `MapPipelineItem` -- provide `ctx` parameter in order to reach out parent Pipeline Context \[Idle mode\] [\431](https://github.com/nicolay-r/AREkit/issues/431)
- NetworkSerializer -- support the case of `Vectorizers==Null` \[Without embedding, google-trans-sampler backlog\] [\430](https://github.com/nicolay-r/AREkit/issues/430)
- ParsedRow -- depends on `pandas`, while it might be switched to `dict` type instead \[AREnets backlog\] [\427](https://github.com/nicolay-r/AREkit/issues/427)
- Remove unused code after AREnets movement [\425](https://github.com/nicolay-r/AREkit/issues/425)
- `AREnets` -- separated project for `networks` contrib part, which provides NN implementation based on Tensorflow [\423](https://github.com/nicolay-r/AREkit/issues/423)
- `Entity` -- Adopt `DisplayValue` property for CSV serialization [\419](https://github.com/nicolay-r/AREkit/issues/419)
- TsvWriter -- Remove `Dataframe` dependency [\408](https://github.com/nicolay-r/AREkit/issues/408)
- OpenNREJsonWriter -- `df.sort` is not an inplace by default [\407](https://github.com/nicolay-r/AREkit/issues/407)
- NeuralNetworkModelIO -- simplify implementation [\406](https://github.com/nicolay-r/AREkit/issues/406)
- Brat -- support nested entities \(`CompoundEntity` type\) \[simple implementation\] [\398](https://github.com/nicolay-r/AREkit/issues/398)
- What's New -- 0.22.1 Release [\323](https://github.com/nicolay-r/AREkit/issues/323)

**Fixed bugs:**

- Brat -- incorrect parsing approach may sometimes results in a wrong value might be mismatched \(use `t`\) [\437](https://github.com/nicolay-r/AREkit/issues/437)
- `VocabRepositoryUtils` -- `numpy` API considers `` by default in vocabulary on load [\428](https://github.com/nicolay-r/AREkit/issues/428)
- LabelsScaler -- uint dict and dict might have different sizes [\426](https://github.com/nicolay-r/AREkit/issues/426)

**Closed issues:**

- `read_ruattitudes_to_brat_in_memory` -- no need to pass label scaler [\436](https://github.com/nicolay-r/AREkit/issues/436)
- `PosTags` -- make them optional parameter for neural networks [\435](https://github.com/nicolay-r/AREkit/issues/435)
- RuSentiFrames -- clarify `tqdm` caption when loading \(ARElight backlog\) [\434](https://github.com/nicolay-r/AREkit/issues/434)
- Sync with AREnets updates [\429](https://github.com/nicolay-r/AREkit/issues/429)
- `BERT` -- provide cropped sampler [\422](https://github.com/nicolay-r/AREkit/issues/422)
- `googletrans` -- move to the separeted project [\421](https://github.com/nicolay-r/AREkit/issues/421)
- `_provide_sentence_terms` -- consider `s_ind` and `t_ind` as well since they may combined with and modified at the same time \[nivts\_project backlog\] [\420](https://github.com/nicolay-r/AREkit/issues/420)
- Entity -- provide DisplayValue property \(which is `Value` by default\) [\418](https://github.com/nicolay-r/AREkit/issues/418)
- `googletrans` -- TranslatorPipelineItem for parsed texts [\416](https://github.com/nicolay-r/AREkit/issues/416)
- Instant downloading -- simplify data downloading [\413](https://github.com/nicolay-r/AREkit/issues/413)
- PandasBasedRowsStorage -- implement the nested type from the `BaseRowsStorage` [\410](https://github.com/nicolay-r/AREkit/issues/410)
- Readers/Writers -- make a part of the contrib [\409](https://github.com/nicolay-r/AREkit/issues/409)
- TextOpinion Annotation -- particular filtering rules for SentiNEREL and Russian texts. \[pipeline items\] [\404](https://github.com/nicolay-r/AREkit/issues/404)
- Evalution -- enhancing error log analysis [\400](https://github.com/nicolay-r/AREkit/issues/400)
- Statistical Folding provided via file [\399](https://github.com/nicolay-r/AREkit/issues/399)
- Balancing as a side part of the Storage [\380](https://github.com/nicolay-r/AREkit/issues/380)

**Merged pull requests:**

- CVE-2007-4559 Patch [\412](https://github.com/nicolay-r/AREkit/pull/412) ([TrellixVulnTeam](https://github.com/TrellixVulnTeam))



\* *This Changelog was automatically generated by [github_changelog_generator](https://github.com/github-changelog-generator/github-changelog-generator)*

Release Notes :tada:

[Full Changelog](https://github.com/nicolay-r/AREkit/compare/v0.22.0-rc-p1...HEAD)

**WHAT'S NEW:**
- :notebook: Provide `BRAT-based reader` (refactoring) of documents and mentioned entities in it! :partying_face:
- :wrench: Provide verbose treatment of values for SynonymsCollection (327)
- :wrench: Fixed embedding issues for `Entity` type for neural networks (308)
- :wrench: Refactoring `RuSentRel` reader, which is now repesents an ontop build over BRAT. (287)
- :wrench: Attitude annotation performed on a fly within a pipeline! (281)
- :wrench: Opinion annotation does not depend on the experiment (250)
- :wrench: 347
- :new: added `utils` contrib part and there were moved :partying_face:
- evaluation (2-3 scale)
- cv-splittings (324)
- entity formatters
- synonyms collections templates: stemmer-based
- experiment handlers (325)
- np_utils -- utils to interact with np-serialized data (348)
- **pipelines** :loop: for opinions extraction and data serialization, text processing: we are now able to declare a custom pipeline and adopt serialization for a variety of RE tasks
(322),
(326)
(351)
- :new: API for conversion of external `text_opinions` into `parsed_news` (338)
- :new: API for a variety of pipelines for data preparation, depending on `DataType` (343)
- :new: `DataType` now includes `Dev` and `Etalon` by default (345)
- :new: Evaluation refactoring, and support `TextOpinion` level results evaluation (355)
- :wastebasket: `experimential_rusentrel` contrib part removed (321)
- :wastebasket: `OpinionRowsProvider` should be removed [ARElight backlog] (282)
- fixed: 356


**Implemented enhancements:**

- RuSentiFrames stat -- move script from `source` to the related UnitTest dir [\391](https://github.com/nicolay-r/AREkit/issues/391)
- Vocabulary for Embedding -- save it in `.txt` format. [\388](https://github.com/nicolay-r/AREkit/issues/388)
- BratSentence -- entities should be initialized via parameter [\383](https://github.com/nicolay-r/AREkit/issues/383)
- ModelIO -- move vocab and embedding related API to EmbeddingIO [\382](https://github.com/nicolay-r/AREkit/issues/382)
- BERT -- formatter differs only in TextB. [\381](https://github.com/nicolay-r/AREkit/issues/381)
- Provide JSON writer for OpenNRE library [\378](https://github.com/nicolay-r/AREkit/issues/378)
- ExperimentSerializationContext -- some parameters might be optional \[Remove them\] [\369](https://github.com/nicolay-r/AREkit/issues/369)
- `ExperimentSerializationContext` -- `Annotator` property is not used. [\368](https://github.com/nicolay-r/AREkit/issues/368)
- DocumentOperations -- `iter_doc_ids` actually wraps the ExperimentContext functionality [\367](https://github.com/nicolay-r/AREkit/issues/367)
- `iter_tagget_doc_ids` -- this might be treated as `iter_doc_ids` of an another instance [\366](https://github.com/nicolay-r/AREkit/issues/366)
- `ExperimentIterationHandler` -- switch to the PipelineItem for NN and BERT serialization \[Remove `ExperimentEngine` and `ExperimentHandler`\] [\365](https://github.com/nicolay-r/AREkit/issues/365)
- `FixedFolding` -- intersected parts are not supported \[NIVTS project backlog\] [\364](https://github.com/nicolay-r/AREkit/issues/364)
- `InputDataSerializationHelper` -- refactoring [\362](https://github.com/nicolay-r/AREkit/issues/362)
- `exp_io.balance_samples`-- remove Dependency from `DataType.Train` [\360](https://github.com/nicolay-r/AREkit/issues/360)
- NeuralNetwork -- for the fine-tunning it is impossible to pick a default embedding/vocabulary. [\359](https://github.com/nicolay-r/AREkit/issues/359)
- Evaluation -- support results evaluation for `TextOpinion` [\355](https://github.com/nicolay-r/AREkit/issues/355)
- `DefaultOpinionAnnotator` -- `etalon_opinion` logic might be moved outside \[Remove `DataType` dependency, backlog\] [\354](https://github.com/nicolay-r/AREkit/issues/354)
- `StatesCount`, `StateIndex` and `iter_states` of `BaseDataFolding` -- this is a part of CV-based method [\353](https://github.com/nicolay-r/AREkit/issues/353)
- Evaluator refactoring [\352](https://github.com/nicolay-r/AREkit/issues/352)
- Processing module -- Multiple Languages Scaling \[Eng/Rus\] \[Contents Relocation\] [\351](https://github.com/nicolay-r/AREkit/issues/351)
- ExperimentContext -- remove Evaluator from the base class. [\349](https://github.com/nicolay-r/AREkit/issues/349)
- `np_utils` -- move from `networks` to `utils` contrib part [\348](https://github.com/nicolay-r/AREkit/issues/348)
- `StringWithEmbeddingNetworkTermMapping` -- has hard-coded algorithms for tokens and terms embedding creation. [\347](https://github.com/nicolay-r/AREkit/issues/347)
- Existed in Embedding -- log \(remove print\) [\346](https://github.com/nicolay-r/AREkit/issues/346)
- DataType -- provide `Dev` and `Etalon` default types \[QUICK fix\] [\345](https://github.com/nicolay-r/AREkit/issues/345)
- Data Serialization -- update API that allow to provide a particular pipeline processor for each `DataType` \[Backlog\] [\343](https://github.com/nicolay-r/AREkit/issues/343)
- Model io utils -- move into `contrib` part [\342](https://github.com/nicolay-r/AREkit/issues/342)
- `Engine` -- provide states iterator as a parameter instead of `DataFolding` [\341](https://github.com/nicolay-r/AREkit/issues/341)
- Brat -- provide stability [\340](https://github.com/nicolay-r/AREkit/issues/340)
- BaseParsedNewsServiceProvider -- support conversion from `Entity` to `DocumentEntity` [\338](https://github.com/nicolay-r/AREkit/issues/338)
- OpinionEntityType -- this should be generalized [\335](https://github.com/nicolay-r/AREkit/issues/335)
- BratTextEntitiesParser and StringPartitioning -- nested entities are not supported. \[Temp fix\] [\334](https://github.com/nicolay-r/AREkit/issues/334)
- RuAttitudesLabelConverter -- required only for conversion \(not for parsing\) [\332](https://github.com/nicolay-r/AREkit/issues/332)
- SentenceOpinion -- no need to store entity values [\331](https://github.com/nicolay-r/AREkit/issues/331)
- Utils -- provide opinion converters from brat [\330](https://github.com/nicolay-r/AREkit/issues/330)
- RuAtttitudes -- move `SentenceOpinion` to brat [\329](https://github.com/nicolay-r/AREkit/issues/329)
- BratEntityCollectionHelper -- `extract_entities` considering for rows prefixed with `T` [\328](https://github.com/nicolay-r/AREkit/issues/328)
- SynonymsCollection -- `value_to_group_id_func` does not support expansion by default. [\327](https://github.com/nicolay-r/AREkit/issues/327)
- BERT and Network Serialization -- refactoring duplicated serialization implementations [\322](https://github.com/nicolay-r/AREkit/issues/322)
- `exp_joined` -- removed such experiment at `experiment_rusentrel` contrib [\321](https://github.com/nicolay-r/AREkit/issues/321)
- `rusentrel_experiment` -- organize a separated python project [\320](https://github.com/nicolay-r/AREkit/issues/320)
- "Uknown}" -- specific to RuSentRel entity case [\319](https://github.com/nicolay-r/AREkit/issues/319)
- `BertExperimentInputSerializerIterationHandler` -- Simplify API \[Blog example backlog\] [\318](https://github.com/nicolay-r/AREkit/issues/318)
- BaseRowsStorage -- consider rows shuffling \[ARElight backlog\] [\316](https://github.com/nicolay-r/AREkit/issues/316)
- EntityIds -- expected to be a part of the BaseSampleRowProvider \[ARElight backlog\] [\312](https://github.com/nicolay-r/AREkit/issues/312)
- `iter_synonym_groups` \[Sources\]-- refactor to common method \[ARElight backlog\] [\310](https://github.com/nicolay-r/AREkit/issues/310)
- term-embedding-pairs -- refactor chain of the parameter dependencies. [\304](https://github.com/nicolay-r/AREkit/issues/304)
- Move EntityFormatters outside [\302](https://github.com/nicolay-r/AREkit/issues/302)
- Sources -- RusentRel collection based on brat toolkit serialization format [\287](https://github.com/nicolay-r/AREkit/issues/287)
- `BaseOpinionsRowProvider` -- useless class and hence should be removed \[refactoring IOUtils\] [\282](https://github.com/nicolay-r/AREkit/issues/282)
- IOUtils -- replace `experiment` instance \(and dependency\) with string provider. [\252](https://github.com/nicolay-r/AREkit/issues/252)
- Annotator and algorithm is not related to experiment. [\250](https://github.com/nicolay-r/AREkit/issues/250)
- DocumentOperations -- parsed docs related API is not related to the expetiment concepts. [\249](https://github.com/nicolay-r/AREkit/issues/249)
- Remove `sep_doc_id` variable [\131](https://github.com/nicolay-r/AREkit/issues/131)
- Update Framework Description [\74](https://github.com/nicolay-r/AREkit/issues/74)

**Fixed bugs:**

- `StringWithEmbeddingNetworkTermMapping` -- `map_token` is expected a particular type of embedding which return embedding only [\395](https://github.com/nicolay-r/AREkit/issues/395)
- NetworksTrainingPipelineItem -- pass labels count [\379](https://github.com/nicolay-r/AREkit/issues/379)
- `BertDefaultStringTextTermsMapper` -- non masked entity values might be with ` ` separation between words [\377](https://github.com/nicolay-r/AREkit/issues/377)
- `iter_rows_linked_by_text_opinions` -- fixed bug with incorrect check. Removed doc-related check. [\356](https://github.com/nicolay-r/AREkit/issues/356)
- TextOpinion should be a part of a single sentence -- this limitation is not emphasized in any way of exceptions and assertions [\339](https://github.com/nicolay-r/AREkit/issues/339)
- BaseParsedNewsServiceProvider -- incorrect IDs assignation [\337](https://github.com/nicolay-r/AREkit/issues/337)
- Example -- Documents become mixed \[RuAttitudes Affection\] [\292](https://github.com/nicolay-r/AREkit/issues/292)
- RuAttitudes -- `extract_text_opinions_linkages` utilizes a different approach which is not covered by common impementation. [\232](https://github.com/nicolay-r/AREkit/issues/232)

**Closed issues:**

- `SamplesIO` -- view always intialized from `tsv` [\397](https://github.com/nicolay-r/AREkit/issues/397)
- `SamplesIO` -- make optional writer [\396](https://github.com/nicolay-r/AREkit/issues/396)
- NoLabel -- allow to customize so for annotators. [\393](https://github.com/nicolay-r/AREkit/issues/393)
- Source -- remove `common` labels [\392](https://github.com/nicolay-r/AREkit/issues/392)
- Tutorials [\390](https://github.com/nicolay-r/AREkit/issues/390)
- Embed SentiNEREL collection [\389](https://github.com/nicolay-r/AREkit/issues/389)
- RuSentRel and RuAttitudes data pipelines -- provide at `utils` contrib [\387](https://github.com/nicolay-r/AREkit/issues/387)
- Serialization pipelines -- move them to `utils` contrib \[pipeline part\] [\386](https://github.com/nicolay-r/AREkit/issues/386)
- Lexicons -- move to the `utils` contrib project [\385](https://github.com/nicolay-r/AREkit/issues/385)
- Remove Gensim dependency [\384](https://github.com/nicolay-r/AREkit/issues/384)
- Evaluation -- ability to extract errors \[Backlog\] [\375](https://github.com/nicolay-r/AREkit/issues/375)
- BaseSampleRowProvider -- has BERT dependencies from contrib [\374](https://github.com/nicolay-r/AREkit/issues/374)
- `BaseIOUtils` -- remove `write_opinion_collection` [\373](https://github.com/nicolay-r/AREkit/issues/373)
- `BaseExperiment` -- remove this class. [\372](https://github.com/nicolay-r/AREkit/issues/372)
- `ExperimentTrainingContext` -- this could be removed. [\371](https://github.com/nicolay-r/AREkit/issues/371)
- BaseTensorflowModel -- provide `DataType` parameter for fitting [\370](https://github.com/nicolay-r/AREkit/issues/370)
- ExperimentSerializationContext -- remove EntityFormatter \[Backlog\] [\361](https://github.com/nicolay-r/AREkit/issues/361)
- `TextOpinion` -- id may be a variety of types [\358](https://github.com/nicolay-r/AREkit/issues/358)
- `TextOpinion` -- remove `owner` field [\357](https://github.com/nicolay-r/AREkit/issues/357)
- Experiment `pipelines` to `contrib.utils` [\326](https://github.com/nicolay-r/AREkit/issues/326)
- Experiment `handlers` to `contrib.utils` [\325](https://github.com/nicolay-r/AREkit/issues/325)
- Experiment `cv` to `contrib.utils` [\324](https://github.com/nicolay-r/AREkit/issues/324)
- RuSentRelOpinionCollectionWriter -- provide encoding parameter \[ARElight backlog\] [\317](https://github.com/nicolay-r/AREkit/issues/317)
- LabelsFormatter for TextB \[BERT\] -- labels might be not supported \[ARElight backlog\] [\315](https://github.com/nicolay-r/AREkit/issues/315)
- RuSentRel experiment -- TextParser could not be customized \[ARElight backlog\] [\314](https://github.com/nicolay-r/AREkit/issues/314)
- InputSerializers \(BERT/Networks\) --` __init__` should not depend on data-related information \[ARElight backlog\] [\313](https://github.com/nicolay-r/AREkit/issues/313)
- StringEntitiesFormatter -- rename EntityType to OpnionEntityType \[QUICK\] [\307](https://github.com/nicolay-r/AREkit/issues/307)
- Annotation -- Opinion annotation should be implemented at `OpinionOperations.iter_opinions_for_extraction` [\281](https://github.com/nicolay-r/AREkit/issues/281)
- SampleView -- adopt multiple views provider \[Refactoring\] [\269](https://github.com/nicolay-r/AREkit/issues/269)

[v0.22.0-rc-p1](https://github.com/nicolay-r/AREkit/tree/v0.22.0-rc-p1) (2022-04-02)

[Full Changelog](https://github.com/nicolay-r/AREkit/compare/v0.22.0-rc-p0...v0.22.0-rc-p1)

**Implemented enhancements:**

- Remove non utilized flags in IterationHandlers \[ARElight backlog\] [\309](https://github.com/nicolay-r/AREkit/issues/309)

**Fixed bugs:**

- BertExperimentInputSerializerIterationHandler -- missed `value_to_group_id_func` parameter [\311](https://github.com/nicolay-r/AREkit/issues/311)

[v0.22.0-rc-p0](https://github.com/nicolay-r/AREkit/tree/v0.22.0-rc-p0) (2022-03-29)

[Full Changelog](https://github.com/nicolay-r/AREkit/compare/v0.22.0-rc...v0.22.0-rc-p0)

**Fixed bugs:**

- Remove `,` presence assertion from Opinon `__init__` class method [\306](https://github.com/nicolay-r/AREkit/issues/306)
- ModuleNotFoundError: No module named 'arekit.common.data.input.providers.instances' [\301](https://github.com/nicolay-r/AREkit/issues/301)

**Closed issues:**

- What's New -- Release 0.22.0 [\227](https://github.com/nicolay-r/AREkit/issues/227)




v0.22.0-rc-p1
Release Notes :tada:
* Pipelines integration!
* Utilized now in text processing, which now could be deleted onto tokenization, entities assignation, frames assignation stages.
* Repositories for opinions and network input samples!
* Storage kernel customizations support for opinion and samples! Using Pandas by default.
* Opinion-related service turn into providers: pairs, opinions, text-opinions, etc.

> **NOTE:** issue [\232](https://github.com/nicolay-r/AREkit/issues/232) has been moved to the next release.
**This version does not support RuAttitudes collection news parsing!**
Will be fixed in the [upcomming project](https://github.com/nicolay-r/AREkit/projects/4).

Changelog

[v0.22.0-rc](https://github.com/nicolay-r/AREkit/tree/v0.22.0-rc) (2022-03-17)

[Full Changelog](https://github.com/nicolay-r/AREkit/compare/v0.21.0-rc...v0.22.0-rc)

Changes

**Implemented enhancements:**

- `create_term_embedding` -- Embedding algorithm based on parts requires useless check [\298](https://github.com/nicolay-r/AREkit/issues/298)
- UnitTests -- BertOntoNotes is no longer below the core processing [\293](https://github.com/nicolay-r/AREkit/issues/293)
- SingleLabelScaler -- provide \[QUICK\] [\291](https://github.com/nicolay-r/AREkit/issues/291)
- BRAT visualization -- support processing in case of multiple documents. [\286](https://github.com/nicolay-r/AREkit/issues/286)
- Entity -- IDs Refactoring [\280](https://github.com/nicolay-r/AREkit/issues/280)
- BaseSampleRowProvider -- provide sentence id [\279](https://github.com/nicolay-r/AREkit/issues/279)
- BRAT tool -- adopt ui as a callback for the predict pipeline [\275](https://github.com/nicolay-r/AREkit/issues/275)
- ExperimentIterationHandler -- add Labeled Output Samples convertion to OpinionCollection [\270](https://github.com/nicolay-r/AREkit/issues/270)
- InferenceContext -- split bags and samples extraction from a single method \[Quick\] [\268](https://github.com/nicolay-r/AREkit/issues/268)
- DataFolding -- organize united data folding. [\267](https://github.com/nicolay-r/AREkit/issues/267)
- BaseDataFolding -- iter\_index is not related to the base implementation [\266](https://github.com/nicolay-r/AREkit/issues/266)
- DataFolding -- move into experiment context [\264](https://github.com/nicolay-r/AREkit/issues/264)
- DataIO \(exp\_data var\) -- rename it to `ExperimentContext` [\263](https://github.com/nicolay-r/AREkit/issues/263)
- ExperimentIterationHandler \(Callback before\) -- organize ExperimentEvaluationCallback [\262](https://github.com/nicolay-r/AREkit/issues/262)
- NetworkCallback -- this callback should not inherit experiment base Callback [\261](https://github.com/nicolay-r/AREkit/issues/261)
- Neural Network Hidden states writers and providers refactoring [\260](https://github.com/nicolay-r/AREkit/issues/260)
- TrainingCallback -- separate onto `TrainingTerminationCallback` and `HiddenWriterCallback` classes. [\259](https://github.com/nicolay-r/AREkit/issues/259)
- BaseTensorflowModel -- simplify `fit` and `predict` operations. [\258](https://github.com/nicolay-r/AREkit/issues/258)
- LabeledCollection -- remove `is_empty` and `reset_labels` api [\257](https://github.com/nicolay-r/AREkit/issues/257)
- NetworkCallback -- move train/predict notification info into callback [\256](https://github.com/nicolay-r/AREkit/issues/256)
- Tensorflow saver -- move the related logic outside of the model implementation [\255](https://github.com/nicolay-r/AREkit/issues/255)
- DefaultSingleLabelAnnotationAlgorithm -- single label is not a part of the algo [\244](https://github.com/nicolay-r/AREkit/issues/244)
- `ThreeScaleTaskAnnotator` -- rename and move into core. [\243](https://github.com/nicolay-r/AREkit/issues/243)
- Data/output -- create pipelines directory with the related output processing [\240](https://github.com/nicolay-r/AREkit/issues/240)
- Examples -- document parsing executes twicely [\239](https://github.com/nicolay-r/AREkit/issues/239)
- Might be utilized pipeline implementation [\238](https://github.com/nicolay-r/AREkit/issues/238)
- OpinionsProvider -- performs two actions, including ids assignation [\236](https://github.com/nicolay-r/AREkit/issues/236)
- entity\_to\_group\_func -- `BaseExperiment` should not provide this method. [\235](https://github.com/nicolay-r/AREkit/issues/235)
- TextOpinionHelper -- to news/parsed/providers \(implement the latter as a provider\) [\233](https://github.com/nicolay-r/AREkit/issues/233)
- DefaultSingleLabelAnnotationAlgorithm -- iter\_opinion duplicates the generalized pair opinion pair creation approach [\231](https://github.com/nicolay-r/AREkit/issues/231)
- Common `languages` dir -- move its contents into processing contrib. [\229](https://github.com/nicolay-r/AREkit/issues/229)
- Linked Text Opinions Refactoring. [\228](https://github.com/nicolay-r/AREkit/issues/228)
- Lemmatization should be a part of the frames processing pipeline stage [\226](https://github.com/nicolay-r/AREkit/issues/226)
- DefaultTextParser -- this class is actually a Tokenizer [\225](https://github.com/nicolay-r/AREkit/issues/225)
- News -- text-opinions provider and entities access API might be a part of a `ParsedNews` by means of `NewsParser` \(new class\) [\224](https://github.com/nicolay-r/AREkit/issues/224)
- StringLabelsFormatter -- switch to label\_types instead of label instances. [\223](https://github.com/nicolay-r/AREkit/issues/223)
- AnnotationAlgorithm -- iter\_opinions requires EntitiesCollection while the latter utilized for entities iteration [\222](https://github.com/nicolay-r/AREkit/issues/222)
- TextParseOptions -- add `keep_tokens` [\221](https://github.com/nicolay-r/AREkit/issues/221)
- FrameVariantsParser -- return modified terms only [\218](https://github.com/nicolay-r/AREkit/issues/218)
- FramesAnnotation -- `is_inverted` flag and processing shoult be a pipeline item [\217](https://github.com/nicolay-r/AREkit/issues/217)
- FramesCollection -- use `FrameConnotationProvider` instead [\216](https://github.com/nicolay-r/AREkit/issues/216)
- FrameVariantsParser -- move into processing subfolder. [\215](https://github.com/nicolay-r/AREkit/issues/215)
- OpinionOperations -- remove `try_read_annotated_opinion_collection` [\213](https://github.com/nicolay-r/AREkit/issues/213)
- DocumentOperation -- unify iter\_doc\_ids operation into one with `tag` parameter. [\212](https://github.com/nicolay-r/AREkit/issues/212)
- OpinionOperations -- move readers\* into IO. [\211](https://github.com/nicolay-r/AREkit/issues/211)
- OpinionCollectionsProvider -- serialization should not be a part of this class [\210](https://github.com/nicolay-r/AREkit/issues/210)
- data -- separate data-related information from the experiment [\209](https://github.com/nicolay-r/AREkit/issues/209)
- BaseInputReader -- class stores `_df`, however it should replaced with `BaseRowsStorage` [\207](https://github.com/nicolay-r/AREkit/issues/207)
- Repositories -- fill method should be a part of a `storage` rather than provider. [\204](https://github.com/nicolay-r/AREkit/issues/204)
- BaseStorage -- exclude `save` method into separated class BaseRowsWriter [\202](https://github.com/nicolay-r/AREkit/issues/202)
- Experiments -- rename `formats` to `api` \(QUICK\) [\201](https://github.com/nicolay-r/AREkit/issues/201)
- Embedding and Vocabulary -- organize Storage/Repository with `serialize`/`load` operations. [\200](https://github.com/nicolay-r/AREkit/issues/200)
- Sample -- remove dependency from DefaultNetworkConfig. [\199](https://github.com/nicolay-r/AREkit/issues/199)
- BaseOutputFormatter -- both provider and formatter mixes `df` usage [\198](https://github.com/nicolay-r/AREkit/issues/198)
- OpinionProvider -- remove dependency from Opinion and Document Operation instances. [\197](https://github.com/nicolay-r/AREkit/issues/197)
- Repositiories -- add this class which unite all the providers for data writing [\195](https://github.com/nicolay-r/AREkit/issues/195)
- Add column providers [\194](https://github.com/nicolay-r/AREkit/issues/194)
- NetworkSampleFormatter -- switch to provider [\193](https://github.com/nicolay-r/AREkit/issues/193)
- BaseSampleStorage -- use `store_labels` instead of `data_type` passing \(QUICK\) [\192](https://github.com/nicolay-r/AREkit/issues/192)
- NetworkOutputEncoder -- separate formatting from serialization. [\191](https://github.com/nicolay-r/AREkit/issues/191)
- BaseSampleFormatter -- `__create_row` is not relted to the Formatter, should be moved. [\190](https://github.com/nicolay-r/AREkit/issues/190)
- BaseDocumentStatGenerator -- provider depends on IO files. [\189](https://github.com/nicolay-r/AREkit/issues/189)
- OpinonFormatter -- use the latter in experiment io. [\188](https://github.com/nicolay-r/AREkit/issues/188)
- News -- remove `return_text` parameter from iter\_sentences method \(QUICK\) [\187](https://github.com/nicolay-r/AREkit/issues/187)
- BaseRowsFormatter -- move `format` method in another class [\185](https://github.com/nicolay-r/AREkit/issues/185)
- BaseSampleFormatter -- `_iter_sentence_terms` should not be a part of this class. \(QUICK\) [\184](https://github.com/nicolay-r/AREkit/issues/184)
- BaseSampleFormatter -- `_provide_rows` behavior depends on row\_ids\_provider instance type. [\182](https://github.com/nicolay-r/AREkit/issues/182)
- BaseSampleFormatter -- remove `data_type` parameter from ctor [\181](https://github.com/nicolay-r/AREkit/issues/181)
- BaseObjectParser -- `parse` method should return object of the same type as `sentence` [\179](https://github.com/nicolay-r/AREkit/issues/179)
- News -- remove `entities_parser` instance from News class. [\178](https://github.com/nicolay-r/AREkit/issues/178)
- BaseEntitiesParser -- generalize to BaseObjectsParser. [\177](https://github.com/nicolay-r/AREkit/issues/177)
- Provide SHA checksums utilization for downloaded resources. [\176](https://github.com/nicolay-r/AREkit/issues/176)
- OpinionCollectionsFormatter -- use it as instance, created within `with` block [\175](https://github.com/nicolay-r/AREkit/issues/175)
- BaseOutput -- move `_csv_to_dataframe` out of this class. [\174](https://github.com/nicolay-r/AREkit/issues/174)
- DataIO -- remove `Stemmer` instance [\172](https://github.com/nicolay-r/AREkit/issues/172)
- BaseRowsFormatter -- `formatter_type_log_name` mehod should be removed. [\171](https://github.com/nicolay-r/AREkit/issues/171)
- BaseOpinionsFormatter -- leave `save` method implementation for inheritor classes. [\170](https://github.com/nicolay-r/AREkit/issues/170)
- BaseSampleFormatter -- leave `save` method implementation for inheritor classes. [\169](https://github.com/nicolay-r/AREkit/issues/169)
- BaseIOUtils -- remove dependencies from file/\(path\) based data storage format [\168](https://github.com/nicolay-r/AREkit/issues/168)
- BaseIOUtils -- `get_input_sample_filepath` `get_input_opinions_filepath` are limit possible storage abilities. [\166](https://github.com/nicolay-r/AREkit/issues/166)
- perform\_reading\_and\_initialization -- provide samples reader. [\165](https://github.com/nicolay-r/AREkit/issues/165)
- perform\_reading\_and\_initialization -- remove dependency from `doc_ops` [\164](https://github.com/nicolay-r/AREkit/issues/164)
- NetworkInputSampleReader -- remove inheritance from TSV-based reader. [\163](https://github.com/nicolay-r/AREkit/issues/163)
- OpinionCollectionsFormatter -- use `save_to` and `load_from` notation for method names with source provider \(file/archive/storage, etc.\). [\142](https://github.com/nicolay-r/AREkit/issues/142)
- RuSentRelOpinionCollectionFormatter -- move all the opinion iteration during saving/loading into base class [\141](https://github.com/nicolay-r/AREkit/issues/141)
- news\_id or doc\_id -- normalize class and field names [\133](https://github.com/nicolay-r/AREkit/issues/133)
- embeddings subdir -- considered to be a part of networks contrib [\132](https://github.com/nicolay-r/AREkit/issues/132)
- Sentiment frame polarity \(A0-\>A1\) considered to be a part of the related experiment. [\118](https://github.com/nicolay-r/AREkit/issues/118)
- EnumServices -- provide a base class with string to Enum conversion functionality [\117](https://github.com/nicolay-r/AREkit/issues/117)
- EntityFormaters -- Move formaters into the particular experiment implementation [\116](https://github.com/nicolay-r/AREkit/issues/116)
- \_create\_parse\_options -- remove this method from DocumentOperations across all the experiments. [\112](https://github.com/nicolay-r/AREkit/issues/112)
- NewsParseOptions -- provide this options for the particular `DefaultParser` derived from `TextParser` [\111](https://github.com/nicolay-r/AREkit/issues/111)
- TextParser -- Provide a separated class with a text processing algorithm implementation API [\75](https://github.com/nicolay-r/AREkit/issues/75)
- Providing all the logging information into log\_utils.py [\30](https://github.com/nicolay-r/AREkit/issues/30)

**Fixed bugs:**

- ModuleNotFoundError: No module named 'arekit.common.data.input.providers.instances' [\301](https://github.com/nicolay-r/AREkit/issues/301)
- UnitTests -- Discard RuAttitudes-v1.2 support due to `index out of range` exception on reading [\295](https://github.com/nicolay-r/AREkit/issues/295)
- text\_opinions\_iter\_pipeline -- ids assigments varies after multiple calls [\278](https://github.com/nicolay-r/AREkit/issues/278)
- EntitiesParser -- provide doc\_level ids [\277](https://github.com/nicolay-r/AREkit/issues/277)
- DeepPavlovNER -- BertOntoNotes entities annotation \[Treating string and list-based text representation simultaneously\] [\274](https://github.com/nicolay-r/AREkit/issues/274)
- Examples -- get\_index\_by\_term of Vocabulary failed [\271](https://github.com/nicolay-r/AREkit/issues/271)
- Annotator Performance -- keeps all possible pairs between entities. [\253](https://github.com/nicolay-r/AREkit/issues/253)
- Network SampleID -- has type `unicode`, but expected to be integer type [\248](https://github.com/nicolay-r/AREkit/issues/248)
- Example -- given two sentences results in samples of only last of them. [\246](https://github.com/nicolay-r/AREkit/issues/246)
- UnitTests -- Incorrect labels formatter \(QUICK\) [\186](https://github.com/nicolay-r/AREkit/issues/186)
- test\_samples\_iter.py -- incorrect API usage in Tensorflow contrib. [\158](https://github.com/nicolay-r/AREkit/issues/158)

**Closed issues:**

- Transfer examples folder into separated project \[ARElight\] [\300](https://github.com/nicolay-r/AREkit/issues/300)
- RuSentRel Experiment -- Text is lemmatized irrespect of the save\_lemmas parameter in parser \[OK\] [\297](https://github.com/nicolay-r/AREkit/issues/297)
- Experiment -- refactor inference pipeline implementation [\290](https://github.com/nicolay-r/AREkit/issues/290)
- Example -- reorganize infer folder \(experiment\) [\289](https://github.com/nicolay-r/AREkit/issues/289)
- Experiment -- Organize pipeline stages as items of the BasePipeline [\285](https://github.com/nicolay-r/AREkit/issues/285)
- BaseSampleRowProvider -- provide entity values and entity types. \[QUICK\] [\283](https://github.com/nicolay-r/AREkit/issues/283)
- DeepPavlov NER -- adopt BERTontonotes. [\272](https://github.com/nicolay-r/AREkit/issues/272)
- NeuralNetworks -- graph and tf session should be initialized before the `predict` method call. [\247](https://github.com/nicolay-r/AREkit/issues/247)
- NewsServiceCollection -- implement [\245](https://github.com/nicolay-r/AREkit/issues/245)
- numpy 1.19.5 -- returns int64 by default [\242](https://github.com/nicolay-r/AREkit/issues/242)
- Organize unit tests for Output to Opinion conversion pipeline [\241](https://github.com/nicolay-r/AREkit/issues/241)
- Iter\_opinions\_collection -- complicated, considering pipeline processing instead [\237](https://github.com/nicolay-r/AREkit/issues/237)
- EntitiesCollection -- provide `value_to_group` function instead of SynonymsCollection. [\230](https://github.com/nicolay-r/AREkit/issues/230)
- BaseTextParser -- `parse_news` is not related to the text parsing concepts and should be a part of the another class [\220](https://github.com/nicolay-r/AREkit/issues/220)
- DocumentOperations -- `_get_text_parser` should not be a part of this API [\219](https://github.com/nicolay-r/AREkit/issues/219)
- Create simple parser for text with mentioned \[entities\] [\214](https://github.com/nicolay-r/AREkit/issues/214)
- NetworkInputHelper -- performing `serialize_missed_collections` during writing process [\208](https://github.com/nicolay-r/AREkit/issues/208)
- RowIDs -- should be common for input and output [\206](https://github.com/nicolay-r/AREkit/issues/206)
- SampleRowBalancerHelper -- simplify by using `pandas` group sampling [\203](https://github.com/nicolay-r/AREkit/issues/203)
- convert\_output\_to\_opinion\_collections -- pass opinion reader into parameters. [\167](https://github.com/nicolay-r/AREkit/issues/167)
- Experiment -- Separate TSV-based formater from based one for samples and opinions [\162](https://github.com/nicolay-r/AREkit/issues/162)
- Switch to Python3.6 [\160](https://github.com/nicolay-r/AREkit/issues/160)
- RuSentRel Experiment Contrib -- update description [\153](https://github.com/nicolay-r/AREkit/issues/153)
- Provide Cache for data sources [\151](https://github.com/nicolay-r/AREkit/issues/151)
- SynonymsCollection considered in ReadOnly mode only [\5](https://github.com/nicolay-r/AREkit/issues/5)

**Merged pull requests:**

- 0.21.1 rc [\234](https://github.com/nicolay-r/AREkit/pull/234) ([nicolay-r](https://github.com/nicolay-r))
- 0.21.1 rc [\196](https://github.com/nicolay-r/AREkit/pull/196) ([nicolay-r](https://github.com/nicolay-r))
- 0.21.0 rc [\159](https://github.com/nicolay-r/AREkit/pull/159) ([nicolay-r](https://github.com/nicolay-r))
- 0.21.0 rc [\157](https://github.com/nicolay-r/AREkit/pull/157) ([nicolay-r](https://github.com/nicolay-r))
- 0.21.0 rc [\152](https://github.com/nicolay-r/AREkit/pull/152) ([nicolay-r](https://github.com/nicolay-r))

Changelog

[v0.21.0-rc](https://github.com/nicolay-r/AREkit/tree/v0.21.0-rc) (2021-08-15)

[Full Changelog](https://github.com/nicolay-r/AREkit/compare/v0.20.5-rc...v0.21.0-rc)

**Implemented enhancements:**

- Sources -- clarify `do_overwrite` and refactor `check_uniqueness` flags RuSentiFrames [\150](https://github.com/nicolay-r/AREkit/issues/150)
- Compose Python Library [\145](https://github.com/nicolay-r/AREkit/issues/145)
- Sources -- provide local storage at home directory [\144](https://github.com/nicolay-r/AREkit/issues/144)
- Enum -- clarify enum34 package using instead of the enum. [\143](https://github.com/nicolay-r/AREkit/issues/143)
- OpinionCollectionsFormatter -- support to save/load only supported by label\_formatter opinions [\139](https://github.com/nicolay-r/AREkit/issues/139)
- UnitTests -- gather all tests into single folder [\125](https://github.com/nicolay-r/AREkit/issues/125)
- BaseAnnotator -- intialize method is useless as the passed parameters requires only at `serialize_missed_collections` method. [\123](https://github.com/nicolay-r/AREkit/issues/123)
- NeutralAnnotator -- Rename to annotator, as neutral prefix is related to a specifics of the particular task [\122](https://github.com/nicolay-r/AREkit/issues/122)
- NeutralAnnot -- use a predefined template for names, based on labels count, instead of Name property [\121](https://github.com/nicolay-r/AREkit/issues/121)
- DefaultNeutralAlgo -- provide dist in sentence parameter [\120](https://github.com/nicolay-r/AREkit/issues/120)
- NeutralAnnot -- Two/Three scale annotators considered to be a part of the related experiment [\119](https://github.com/nicolay-r/AREkit/issues/119)
- Evaluation Metrics -- such functions considered to be a part of the particular experiment [\115](https://github.com/nicolay-r/AREkit/issues/115)
- Embedding -- set\_stemmer method is not declared in base class [\114](https://github.com/nicolay-r/AREkit/issues/114)
- FrameVariantsCollection -- remove stemmer from \_\_init\_\_ params. [\113](https://github.com/nicolay-r/AREkit/issues/113)
- Bag \(NeuralNetworks\) -- label could be presented as uint. [\110](https://github.com/nicolay-r/AREkit/issues/110)
- experiment\_rusentrel -- Group all folders by a single `exp` prefix [\108](https://github.com/nicolay-r/AREkit/issues/108)
- BaseModel -- Replace epochs\_count parameter with generalized parameter structure. [\107](https://github.com/nicolay-r/AREkit/issues/107)
- OpinionCollection -- provide set of supported labels \(opinion filtration by labels\) [\106](https://github.com/nicolay-r/AREkit/issues/106)
- LabelCalculationMode -- make it enum [\105](https://github.com/nicolay-r/AREkit/issues/105)
- BaseModel -- replace epochs\_count with model options [\104](https://github.com/nicolay-r/AREkit/issues/104)
- ThreeLabelsScaler -- remove dependecies of the latter in NeuralNetwork contrib [\103](https://github.com/nicolay-r/AREkit/issues/103)
- RuAttitudes -- use int\_to\_label function instead of label scaler [\102](https://github.com/nicolay-r/AREkit/issues/102)
- Labels -- Move Scaler into common/labels [\101](https://github.com/nicolay-r/AREkit/issues/101)
- Labels -- Provide a unique labels for the partucular experiment in contrib [\100](https://github.com/nicolay-r/AREkit/issues/100)
- Experiments -- reorganize rusentrel experiments data within the related new folder [\97](https://github.com/nicolay-r/AREkit/issues/97)

**Fixed bugs:**

- RuAttitudes-v1.2. -- fix downloading link [\155](https://github.com/nicolay-r/AREkit/issues/155)
- sources -- Remove data folder [\149](https://github.com/nicolay-r/AREkit/issues/149)
- Entity -- type could be `None` while there is no restriction for that [\148](https://github.com/nicolay-r/AREkit/issues/148)
- RuSentRelOpinionCollectionFormatter -- label could not be found during neural network training. [\137](https://github.com/nicolay-r/AREkit/issues/137)
- frame\_variant -- label scaler receives `NoLabel` while experiment based on `NeutralLabel` [\136](https://github.com/nicolay-r/AREkit/issues/136)
- BaseEvaluator -- opinion labels might be incompatible with the one utilized in ResultEvaluator. [\124](https://github.com/nicolay-r/AREkit/issues/124)

**Closed issues:**
- UnitTests -- Run all unit tests via bash script [\156](https://github.com/nicolay-r/AREkit/issues/156)
- Remove release\_notes.md file and move the related content into Releases descriptions. [\146](https://github.com/nicolay-r/AREkit/issues/146)
- Tutorial -- Clarify on how we perform optimization [\90](https://github.com/nicolay-r/AREkit/issues/90)

Release

Fixed:
* Using custom check of duplicated opinions during `OpinionCollection` initialization.
Changes:
* Speed-up and engine optimizations:
* Optionally loading neutral annotator.
* Multi-Instance networks: now we consider that the next appered context always continues the prior.
(check out multi-instance bags creation for details)
* Now shuffling in models performed for bags, not for bag groups.
* Networks: added `allow_growth=True` flag for tensorflow based neural networks.
Memory fraction parameter has been removed.

Collection of parsed news become dispatched from text opinions collection.
* **News parsing now is assumed to be performed using `TextParser.parse(news, options)` call. Related refactoring.**
* Stemmer application from `RuAttitiudes` parser has been removed.
* Removed dependency from `RelatedParsedNewCollection` in TextOpinionCollection.
* Labeling now separated from LinkedTextOpinion collection.
* `ParsedText` class has been refactored, removed unused methods. Keep tokens has been discarded.
* BERT tsv-format-encoders are now in a Factory (at contrib directory).
* Fixed: `RuSentRelTextOpinion` replaced with `TextOpinion`, and independent from `OpinionRef`.
* `Single`/`Multi` models now are not exist, as the latter prefixes affects only onto batch types selection. Refactoring.

Release Notes

* Labels conversion `to_str` and `from_str` now a part of external formatters (unique for each source, experiment, etc.).
* Added labels-scaler, and labels casing (to int or uint) now depends on scaler;
* Added bert exporter in contribution folder: with related formatters according to the [[paper]](https://www.aclweb.org/anthology/N19-1035.pdf):
* **NLI** -- (Natural language inference) format, assumes to provide an additional sentence, which describes
attitude should be extracted
* **QA** -- (Question answering) provides an additional question onto attitude sentiment.
With Label encoding in following format:
* **Multiple** -- all the supported sentiment labels (positive, negative, neutral)
* **Binary** -- (YES, NO) according to mention (additional sentence), provided by **NLI** and **QA** formatters.
* Refactoring experiments in order to apply the latter also for classifiers (models from scikit-learn)
* Updated nn-engine API
* Refactoring tf-based neural network implementation.
* Bert now moved into separated folder from `contrib` directory.
* frame_variants moved to `frames` directory.
* Frame variants labeling in news now performed during `parse` operation.
* `DataType` now enumeration. List of Supported data-types now a part of experiment
The latter were moved onto sample level.
* `Service` folder removed as the latter assumes to be apart of this repository.

v0.20.3-wims-rc
Release Notes
* Experiments now supports two-scale and three-scale task representation with the related evaluation formats.