Chandassu

<div align= "center">

</div>

Python library implementing "Chandassu", metrical poetry in global languages. </br> Currently supported language(s):

• Telugu

arXiv Preprint: Read Here

Benchmarks

Our algorithm achieves 91.73% accuracy on the proposed Chandassu Score.

Usage

Install package using

pip install chandassu

Check padyam

from chandassu.telugu.laghuvu_guruvu import LaghuvuGuruvu
from chandassu.telugu.padya_bhedam import check_padyam

data= """తొండము నేక దంతమును తోరపు బొజ్జయు వామ హస్తమున్
మెండుగ మ్రోయు గజ్జెలును మెల్లని చూపులు మందహాసమున్
కొండొక గుజ్జు రూపమున కోరిన విద్యలకెల్ల నొజ్జవై
యుండెడి పార్వతీ తనయ యోయి గణాధిప నీకు మ్రొక్కెదన్"""

# Generate LaghuvuGuruvu Sequence
lg_data= LaghuvuGuruvu(data= data).generate()

score= check_padyam( 
                        lg_data= lg_data ,
                        type= "vutpalamaala",
                        return_micro_score= True, 
                        verbose= False
                    )

print(score)

Computational Architecture

• nidhi.py: Telugu alphabet classification with phonological categorizations </br>
• panimuttu.py: Vowel-consonant operations utility functions </br>
• laghuvu_guruvu.py: Prosodic analysis engine with tokenizer and syllabic weight classifier </br>
• ganam.py: Pattern repository mapping LaghuvuGuruvu sequences to metrical units </br>
• padyam_config.py: Configuration database for padyam type constraints and rules </br>
• check_lakshanam.py: Validation module for yati (caesura) and prasa (alliteration) </br>
• padya_bhedam.py: Main integration module for metrical assessment and evaluation </br>

Notebooks:

1. web_scrapping.ipynb: Scraper scripts to collect data from andhrabharati.com
2. Data_Analysis.ipynb: Analyzes data using different visualizations
3. Chandassu_Generation.ipynb: Generates chandassu for the scraped data
4. Score_Generation.ipynb: Computes scores by after chandassu generation
5. Accuracy_Evaluation.ipynb: Benchmarking and accuracy (Chandassu Score) evaluation

Eager to Contribute?

Any data, features, refactoring or your innovative thought, please check <a href= "https://github.com/BodduSriPavan-111/chandassu/blob/main/CONTRIBUTING.md">here</a>.

Acknowledgements

Special thanks to Sesha Sai Vadapalli and Kalepu Nagabhushana Rao, maintainers of andhrabharati.com. </br> Appana Mohan Naga Phani Kumar for proofreading our article. </br> Sincere gratitude to our parents and family members for their continuous support throughout this work. </br> This work was undertaken with the grace of Sri Ramalinga Chowdeswari Devi.

Citation

@misc{pavan2025computationalsociallinguisticstelugu, </br> title={Computational Social Linguistics for Telugu Cultural Preservation: Novel Algorithms for Chandassu Metrical Pattern Recognition}, </br> author={Boddu Sri Pavan and Boddu Swathi Sree}, </br> year={2025}, </br> eprint={2510.01233}, </br> archivePrefix={arXiv}, </br> primaryClass={cs.CL}, </br> url={https://arxiv.org/abs/2510.01233}, </br> }

Thank You !