~~~
2021-07-18
* Implement throttling as per Scopus' definition.
* Enable type hints.
* Revamp documentation: Document all classes with examples on the same page in the new section "Classes"; Add new section "Access" using subsections previously listed under "Tips".
* For the following properties, type str changed to numeric type (either as scalar or within a container): `AbstractRetrieval().authorgroup.affiliation_id`, `AbstractRetrieval().authorgroup.dep_id` and `AbstractRetrieval().authorgroup.auid`, `AbstractRetrieval().authors.id`, `AbstractRetrieval().authors.auid`, `AbstractRetrieval().citedby_count`, `AbstractRetrieval().confcode` `AbstractRetrieval().identifier`, `AbstractRetrieval().openaccess`, `AbstractRetrieval().pubmed_id`, `AbstractRetrieval().source_id`, `AbstractRetrieval().subject_areas.code` `AffiliationRetrieval().author_count`, `AffiliationRetrieval().document_count`, `AffiliationRetrieval().identifier`, `AffiliationRetrieval().name_variants.doc_count`, `AffiliationSearch().affiliations.documents`, `AuthorRetrieval().affiliation_current.id` and `AuthorRetrieval().affiliation_current.parent`, `AuthorRetrieval().affiliation_history.id` and `AuthorRetrieval().affiliation_history.parent`, `AuthorRetrieval().citation_count`, `AuthorRetrieval().cited_by_count`, `AuthorRetrieval().classificationgroup`, `AuthorRetrieval().coauthor_count`, `AuthorRetrieval().document_count`, `AuthorRetrieval().historical_identifier`, `AuthorRetrieval().identifier`, `AuthorRetrieval().subject_areas.code`, `AuthorSearch().authors.documents`, `ScopusSearch().results.citedby_count` and `ScopusSearch().results.openaccess`, `SerialTitle().citescoreyearinfolist`, `SerialTitle().sjrlist`, `SerialTitle().sniplist`, `SerialTitle().source_id`, `SerialTitle().subject_area.code`.
* The type of `AbstractRetrieval().authors.affiliation` is now a string joining all affiliation IDs on ";".
* In `AffiliationSearch()` and `AuthorSearch()`, parameter `count` is deprecated and will be removed in a future release. There is no substitute.
* In `AbstractRetrieval()`, property `.correspondence` is now a list.
* In `CitationOverview()`, parameter `eid` is deprecated and will be removed in a future release. Instead, the class now accepts a list of up to 25 identifiers (Scopus ID, DOI, PII or Pubmed ID) and returns individual citation trajectories for each of the documents.
* In `CitationOverview()`, all current properties now return lists, except for `.h_index`, which remains int.
* In `CitationOverview()`, add new properties `.columnTotal`, `.grandTotal`, `.laterColumnTotal`, `prevColumnTotal`, `.rangeColumnTotal`.
* In `CitationOverview()`, change the cache file name to md5-hash of the used identifiers plus the citation type information (e.g., when self-citations are excluded).
* In `AffiliationRetrieval()`, add property `.status`.
* In `SubjectClassifications()`, the default cache file path now goes without the folder "STANDARD".
* The default folder for the configuration file become `.pybliometrics` instead of `.scopus`, but if `.scopus` exists, `pybliometrics` will use this folder. New installations will only use `.pybliometrics`. The default path for the cache file folders has become `.pybliometrics/Scopus/`.
* Allow for kwds to be passed on in the following classes: `AffiliationRetrieval()`, `AffiliationSearch()`, `AuthorSearch()`, `CitationOverview()`, `SubjectClassifications()`.
* For search classes, change error message when the search result size exceeds 5k. The possiblity to change this number has been removed.
* In all classes, raise ValueError if parameter `refresh` is neither int nor bool.
* In all classes, harmonize error message documentation and order of parameters.
* Require `tqdm` package to print progress bars.
* Class `ContentAffiliationRetrieval()` has been removed.
* Fix bug that cached an empty file when `download=False` using any Search class.