additions

Additions and Release Notes (.NET Port)

This page tracks ZemberekDotNet-specific additions that do not exist in Java Zemberek, plus release-level API notes.

Version 0.19.2

Morphology: Extended API additions

Namespace:

• ZemberekDotNet.Morphology.Extended

Added public API:

• TurkishCase enum (ZemberekDotNet.Core.Turkish)
• SingleAnalysis extensions:
  • ExtdGetCase()
  • ExtdGetPossibleCases()
• TurkishMorphology extensions:
  • ExtdAnalyzeNumeralWithSuffix(string input)
  • ExtdAnalyzeWithRanking(string input, WordFrequencyModel frequencyModel = null)
  • ExtdSynthesize(string lemma, TurkishCase targetCase, bool plural = false)
• WordAnalysis extension:
  • ExtdGetRankedAnalyses(WordFrequencyModel frequencyModel = null)
• New helper types:
  • RankedAnalysis
  • WordFrequencyModel
  • ExtendedMorphologyContext

Embedded corpus

ZemberekDotNet.Morphology now embeds a Turkish frequency resource:

• Extended/Resources/tr_50k.txt

This corpus is used by ranking/fuzzy workflows and is packaged inside the Morphology NuGet artifact.

Correctness note

ExtdGetCase() and ExtdGetPossibleCases() now infer nominative from the last morpheme group's POS, not from any nominal morpheme anywhere in the chain.

Why this matters:

• It prevents adjective-derivation false positives such as elma + With(lı) -> Adj from being incorrectly marked as nominative.

Test coverage

Extended morphology tests include:

• numeral apostrophe analysis
• typed case extraction
• ranked analysis confidence behavior
• BK-tree fuzzy analysis behavior
• morphology synthesis helper behavior
• edge-case POS correctness for derivations (adjective, adverb, postposition, verb, gerund)

Compatibility and migration

• These APIs are additive and backward-compatible.
• Existing Morphology APIs remain unchanged.
• To use new methods, import ZemberekDotNet.Morphology.Extended.