Skip to content

Latest commit

 

History

History
114 lines (85 loc) · 5.73 KB

CHANGES-1.0.0.md

File metadata and controls

114 lines (85 loc) · 5.73 KB
Raw

Overview of changes in version 1.0.0

AUTOTYP version 1.0.0 is a completely new release that focuses on usability, documentation and completeness. It has been radically overhauled compared to the earlier 0.1.x version. The sheer number of differences makes it impossible to provide a comprehensive list of changes. What follows is a quick summary of the most important of the new release as well as notes on migrating from the old database releases.

Major new features in version 1.0.0:

  • New naming conventions for datasets and variables, focusing on usability and clarity. All names now consistently follow the CamelCase convention and are based on verbose descriptions that provide more context about the variable (e.g. Position -> VerbInflectionMarkerPosition). Hundreds of variables have been renamed to fit these criteria.

  • The datasets are now organized into thematic modules, rather than each dataset constituting a module on its own.

  • Published data now includes the raw exported database data, in addition to the previously published derived aggregated tables. All aggregation scripts used to compute derived data are published as well (see aggregation-scripts). Please feel free to inspect the scripts and modify them to suit your own needs.

  • Many improvements to variable descriptions and metadata. The metadata YAML files are now simpler and more compact, which should make the documentation more accessible.

  • Overhauled the data architecture to allow nested and repeated table fields (see Data Architecture). This allows many datasets to be expressed in a more natural, conceptually simpler fashion.

  • New R and JSON exports for users who want quick access to the data using their preferred data wrangling environment.

  • Language name and glottocode is exported for every dataset in addition to the internal language ID

Major changes to individual datasets/modules:

  • GrammaticalRelations module now encompasses all data on grammatical relations and alignments. We now fully provide the underlying raw database data in addition to the aggregated alignment data and the scripts used to produce these aggregations.

  • VerbSynthesis has been overhauled to include detailed list of inflectional categories expressed on verbs

  • LocusOfMarking module now contains the raw database data in addition to the previously published aggregations.

  • GrammaticalMarkers dataset has been overhauled to include a detailed list of marker hosts and marked categories

  • MorphemeClasses replaces the previous aggregated Morpheme_types dataset and exposes the information about individual language-specific morpheme classes. The information previously available in Morpheme_types is now integrated into the improved MorphologyPerLanguage aggregated dataset.

  • New module Categories groups together datasets that provide information about selected grammatical categories

  • New module Definitions provides access to underlying definitions of categorical variables used across AUTOTYP

  • New module PerLanguageSummaries groups together various per-language aggregated summaries (code to generate these summaries is available under aggregation-scripts)

Notes on migration from older AUTOTYP release

If you have been using the AUTOTYP version 0.1.x you will notice that many datasets have been moved or renamed. The following list should help you to find the new location of the data:

  • Agreement is now exported as Categories/Agreement
  • Alienability is now exported as Categories/Alienability
  • Alignment is now exported as GrammaticalRelations/Alignment
  • Alignment_per_language is now PerLanguageSummaries/AlignmentForDefaultPredicatesPerLanguage
  • Clause_linkage is now Sentence/ClauseLinkage
  • Clause_word_order is now Sentence/ClauseWordOrder
  • Clusivity is now exported as Categories/Clusivity
  • Gender is now exported as Categories/Gender
  • Grammatical_markers is now exported as Morphology/GrammaticalMarkers
  • GR_per_language has been superseded by GrammaticalRelations/GrammaticalRelationCoverage
  • Locus_per_language is now PerLanguageSummaries/LocusOfMarkingPerLanguage
  • Locus_per_macrorelation has been superseded by Morphology/DefaultLocusOfMarkingPerMacrorelation
  • Locus_per_microrelation has been superseded by Morphology/LocusOfMarkingPerMicrorelation
  • Markers_per_language is now PerLanguageSummaries/GrammaticalMarkersPerLanguage
  • Morpheme_types has been superseded by Morphology/MorphemeClasses and PerLanguageSummaries/MorphologyPerLanguage
  • Morphology_per_language is now PerLanguageSummaries/MorphologyPerLanguage
  • NP_per_language is now PerLanguageSummaries/NPStructurePerLanguage
  • NP_structure is now NP/NPStructure
  • NP_structure_presence is now PerLanguageSummaries/NPStructurePresence
  • Numeral_classifiers is now exported as Categories/NumeralClassifiers
  • Register is still Register
  • Synthesis is now Morphology/VerbSynthesis
  • Valence_classes is now GrammaticalRelations/PredicateClasses
  • Valence_classes_per_language is now PerLanguageSummaries/PredicateClassesSemanticsPerLanguage
  • VInfl_counts_per_position is now PerLanguageSummaries/VerbInflectionAndAgreementCountsByPosition
  • VInfl_cat_* is now PerLanguageSummaries/VerbInflectionCategoriesAggregatedBy*
  • VInfl_macrocat_* is now PerLanguageSummaries/VerbInflectionMacrocategories*
  • VAgr_* is now PerLanguageSummaries/VerbAgreementAggregatedBy*
  • Word_domains is now Word/WordDomains