uzbek_morph — Rule-based morphological analyzer for the Uzbek language

Author: Ulugbek Salaev Model Page: 🌐 http://uzmorph.uz/models/uzbek_morph

uzbek_morph is a high-precision hybrid morphological analyzer and lemmatizer for the Uzbek language. It features a rule-based engine and a corpus of 145,844 unique surface forms.

🌟 Key Features

1. 📏 Rule-Based Analysis Engine:

   • Step-by-step affix stripping using linguistic and morphophonemic rules to accurately extract canonical lemmas for out-of-vocabulary words.

2. 📚 Extensive Vocabulary Coverage (145,844+ Unique Forms):

   • Comprehensive grammatical and lexical coverage across Nouns, Verbs, Adjectives, Adverbs, Pronouns, Numerals, and Conjunctions.

3. 🔄 Morphophonemic Sound Change Rules:

   • Consonantal Lenition:
     • g' ➔ q: qulog'im ➔ quloq, tarmog'i ➔ tarmoq, qobig'ida ➔ qobiq
     • g ➔ k: eshigim ➔ eshik, yuragida ➔ yurak, bilagidan ➔ bilak
   • Vowel Syncope / Elision:
     • burni ➔ burun, shahrimizga ➔ shahar, qornida ➔ qorin, ko'nglimiz ➔ ko'ngil, og'zidan ➔ og'iz

4. 🔗 Compound & Auxiliary Verb Contextual Parser:

   • Intelligent rule chain for Uzbek converb (-ib/-b) + auxiliary verb constructions (uxlab oldi, aytib keldi, yozib berdi, kulib yubordi), disambiguating auxiliary verbs like oldi into olmoq (verb) rather than noun or adjective.

5. 👤 Pronoun & Numeral Paradigms:

   • Complete inflected personal/demonstrative pronouns (menga, senga, unga, bunga, shunga, bunda, shunda, bundan, shundan) and numerals (birinchi, ikkinchi, ikkala, uchala).

6. ⚡ Flexible & Unified API (analyze):

   • Single analyze() entry point that seamlessly processes single words, compound phrases, or complete sentences.

📦 Installation

pip install uzbek_morph

💻 Usage Code Snippets

1. Quickstart (Default Tabular Display):

import uzbek_morph

# Single Word:
uzbek_morph.analyze("kitoblarimizga")

# Compound Phrase:
uzbek_morph.analyze("uxlab oldi")

# Full Sentence:
uzbek_morph.analyze("Men kecha go'zal shahrimizga kitob sotib olgani bordim")

2. Programmatic Data Processing:

Set print_result=False to receive a list of structured dictionaries without printing the console table:

from uzbek_morph import analyze

data = analyze("Men kecha go'zal shahrimizga bordim", print_result=False)

for item in data:
    print(f"Word: {item['word']} | Lemma: {item['lemma']} | Tag: {item['tags']} | Source: {item['source']}")

📄 License & Author

Author: Ulugbek Salaev Distributed under the MIT License.

Note: Data schema adheres to UniMorph morphosyntactic specifications.