Chanda

Sanskrit Meter Identification and Analysis Library

A comprehensive Python library for identifying and analyzing Sanskrit poetic meters. Supports 200+ meters including Sama-vṛtta, Ardhasama-vṛtta, Viṣama-vṛtta, and Mātrā-vṛtta.

• Chandojñānam Web Interface: https://sanskrit.iitk.ac.in/jnanasangraha/chanda/
• Chandojñānam Web Interface Source Code: https://github.com/hrishikeshrt/chandojnanam/

Features

• 200+ Meter Database: Comprehensive coverage of Sanskrit meters

  • Sama-vṛtta (meters with identical padas)
  • Ardhasama-vṛtta & Viṣama-vṛtta (meters with varying padas)
  • Mātrā-vṛtta (matra-based meters like Āryā, Gīti)

• Smart Identification:

  • Exact pattern matching
  • Fuzzy matching with syllable-level correction suggestions
  • Verse-level analysis (4-line grouping)

• Multi-Script Support: 15+ scripts and transliteration schemes

  • Devanagari, IAST, ITRANS, Harvard-Kyoto, SLP1, WX, Velthuis, and more

• Detailed Analysis:

  • Syllable segmentation
  • Laghu-Guru (light-heavy) marking
  • Gana notation conversion
  • Mātrā (morae) counting
  • Jāti (classification) identification

Installation

From PyPI

pip install chanda

From Source

git clone https://github.com/hrishikeshrt/chanda.git
cd chanda
pip install -e .

Quick Start

Python API

from chanda import analyze_line

# Simple example - Anuṣṭubh (most common meter)
text = "को न्वस्मिन् साम्प्रतं लोके गुणवान् कश्च वीर्यवान्"
result = analyze_line(text)
meters = [name for name, _ in result.chanda]
print(meters)  # Output: ['अनुष्टुभ्']

# Gana-based meter example - shows gana pattern
text = "नमस्ते सदा वत्सले मातृभूमे"
result = analyze_line(text)
meters = [name for name, _ in result.chanda]
print(meters)  # Output: ['भुजङ्गप्रयात']
print(result.gana)            # Output: यययय

Results are objects (ChandaResult / TextAnalysisResult). Use result.to_dict() or result.to_json() for serialization.

Command-Line Interface

# Analyze a single line
chanda "को न्वस्मिन् साम्प्रतं लोके गुणवान् कश्च वीर्यवान्"

# Analyze a file
chanda -f bhagavad_gita.txt --verse --summary

# Interactive mode
chanda -i

# Get help
chanda --help

Detailed Usage

Single Line Analysis

from chanda import analyze_line

text = "रामो राजमणिः सदा विजयते रामं रमेशं भजे"
result = analyze_line(text, fuzzy=True)

if result.found:
    print(f"Meter(s): {[name for name, _ in result.chanda]}")
    print(f"Syllables: {result.syllables}")
    print(f"Pattern (LG): {' '.join(result.lg)}")
    print(f"Gana: {result.gana}")
    print(f"Syllable count: {result.length}")
    print(f"Mātrā count: {result.matra}")
else:
    # Check fuzzy matches
    if result.fuzzy:
        best_match = result.fuzzy[0]
        from chanda import format_chanda_list
        print(f"Closest meter: {format_chanda_list(best_match['chanda'])}")
        print(f"Similarity: {best_match['similarity']:.2%}")

Verse Analysis

from chanda import analyze_text

verse = """को न्वस्मिन् साम्प्रतं लोके गुणवान् कश्च वीर्यवान्।
धर्मज्ञश्च कृतज्ञश्च सत्यवाक्यो दृढव्रतः॥"""

results = analyze_text(verse, verse_mode=True, fuzzy=True)

for verse_result in results.result.verse:
    if not verse_result.chanda:
        continue

    best_meters, _score = verse_result.chanda
    # is_partial is True when fewer lines than expected were available
    partial_note = " (partial)" if verse_result.is_partial else ""
    print(f"Verse meter: {' / '.join(best_meters)}{partial_note}")

    # Detailed per-meter breakdown via MeterScore
    winner = verse_result.scores[0]
    print(f"  Match extent: {winner.match_extent:.0%}")  # fraction of padas explained
    for ev in winner.evidence:
        line_idx  = ev['line_idx']   # None for mātrā evidence
        mtype     = ev['match_type'] # 'exact', 'fuzzy', or 'matra'
        sim       = ev['similarity'] # 1.0 for exact
        pada      = ev['pada']       # e.g. ['1', '2']
        valid_pos = ev['pada_position_valid']
        print(f"    line {line_idx}: {mtype}, similarity={sim:.2f}, pada={pada}, pos_valid={valid_pos}")

Note on aliases: meters with multiple traditional names are joined with =, e.g. "कलहंस = कुटजा = सिंहनाद = नन्दिनी". A tie between distinct meters appears as separate entries in verse_result.scores.

Note on score vs match_extent: score in MeterScore is an internal ranking accumulator and should not be displayed. Use match_extent (0.0–1.0) to show users how well a meter fits the verse.

Documentation

Full documentation is available at chanda.readthedocs.io

Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes with tests
4. Run the test suite
5. Submit a pull request

Citation

If you use Chandojñānam in your research, please cite:

@inproceedings{terdalkar2023chandojnanam,
    title = "Chandojnanam: A {S}anskrit Meter Identification and Utilization System",
    author = "Terdalkar, Hrishikesh  and
      Bhattacharya, Arnab",
    booktitle = "Proceedings of the Computational {S}anskrit {\&} Digital Humanities: Selected papers presented at the 18th World {S}anskrit Conference",
    month = jan,
    year = "2023",
    address = "Canberra, Australia (Online mode)",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2023.wsc-csdh.8",
    pages = "113--127",
}

License

GNU Affero General Public License v3.0 - see LICENSE file for details.

Acknowledgments

• Based on traditional chandaśāstra texts
• Meter definitions compiled from various classical sources
• Built on the excellent sanskrit-text and indic-transliteration libraries

Links

• PyPI: https://pypi.org/project/chanda/
• GitHub: https://github.com/hrishikeshrt/chanda
• Documentation: https://chanda.readthedocs.io
• Issues: https://github.com/hrishikeshrt/chanda/issues