ninada 4.0.0__py3-none-any.whl

ninada/__init__.py

@@ -0,0 +1,76 @@
+"""ninada (निनाद): Streaming Pan-Indic & Classical Sanskrit Neural Text-to-Speech."""
+from .engine import (
+    HOP_LENGTH,
+    SAMPLE_RATE,
+    Ninada,
+    encode_wav_bytes,
+    speak,
+    split_prosodic_clauses,
+    stream,
+    synthesize_klatt_pcm,
+)
+from .g2p import (
+    BLANK_ID,
+    ID_TO_LANGUAGE,
+    ID_TO_TOKEN,
+    IPA_ID_TO_TOKEN,
+    IPA_VOCAB,
+    LANGUAGE_IDS,
+    RETROFLEX_LATERAL_ID,
+    TOKEN_TO_ID,
+    UNK_ID,
+    VOCAB_TOKENS,
+    WARMSTART_DONOR_MAP,
+    apply_phonological_rules,
+    brahmic_to_devanagari,
+    detect_script_and_lang,
+    iast_to_devanagari,
+    ipa_to_ids,
+    segment_aksharas,
+    text_to_ids,
+    text_to_ipa,
+)
+
+from .studio_g2p import (
+    STUDIO_HOP_LENGTH,
+    STUDIO_VOCABS,
+    clean_studio_text,
+    studio_text_to_ids,
+)
+
+__version__ = "4.0.0"
+
+__all__ = [
+    "__version__",
+    "SAMPLE_RATE",
+    "HOP_LENGTH",
+    "STUDIO_HOP_LENGTH",
+    "STUDIO_VOCABS",
+    "clean_studio_text",
+    "studio_text_to_ids",
+    "Ninada",
+    "speak",
+    "stream",
+    "split_prosodic_clauses",
+    "encode_wav_bytes",
+    "synthesize_klatt_pcm",
+    "VOCAB_TOKENS",
+    "TOKEN_TO_ID",
+    "ID_TO_TOKEN",
+    "IPA_VOCAB",
+    "IPA_ID_TO_TOKEN",
+    "RETROFLEX_LATERAL_ID",
+    "BLANK_ID",
+    "UNK_ID",
+    "LANGUAGE_IDS",
+    "ID_TO_LANGUAGE",
+    "WARMSTART_DONOR_MAP",
+    "detect_script_and_lang",
+    "iast_to_devanagari",
+    "brahmic_to_devanagari",
+    "apply_phonological_rules",
+    "segment_aksharas",
+    "text_to_ipa",
+    "ipa_to_ids",
+    "text_to_ids",
+]

ninada/cli.py

@@ -0,0 +1,43 @@
+"""Command-line interface for Ninada (Pan-Indic & Classical Sanskrit TTS)."""
+from __future__ import annotations
+
+import argparse
+import json
+import sys
+from pathlib import Path
+
+from .engine import SAMPLE_RATE, Ninada
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser(
+        prog="ninada",
+        description="Ninada — Pan-Indic & Classical Sanskrit Neural TTS (24 kHz Studio WAV)",
+    )
+    parser.add_argument("text", nargs="?", help="Input text in Devanagari, Kannada, Tamil, Telugu, Malayalam, or IAST")
+    parser.add_argument("-l", "--lang", default=None, help="Language code: sa | hi | kn | ta | te | ml")
+    parser.add_argument("-o", "--output", type=Path, default=Path("ninada_out.wav"), help="Output WAV file path")
+    parser.add_argument("-r", "--rate", type=float, default=1.0, help="Speaking rate multiplier (default: 1.0)")
+    parser.add_argument("--tier", choices=["auto", "neural", "klatt"], default="auto", help="Synthesis engine tier")
+    parser.add_argument("--inspect", action="store_true", help="Print Pan-Indic G2P & akshara prosody JSON")
+    args = parser.parse_args()
+
+    txt = args.text if args.text else sys.stdin.read().strip()
+    if not txt:
+        parser.error("Provide input text as an argument or via stdin.")
+
+    engine = Ninada(lang=args.lang or "sa")
+    pcm, meta = engine.synthesize(txt, lang=args.lang, rate=args.rate, tier=args.tier)
+    if args.inspect:
+        print(json.dumps({k: v for k, v in meta.items() if k != "durations"}, ensure_ascii=False, indent=2))
+        return
+
+    from .engine import encode_wav_bytes
+
+    wav = encode_wav_bytes(pcm, sr=SAMPLE_RATE)
+    args.output.write_bytes(wav)
+    print(f"Wrote {len(wav)} bytes ({len(pcm)/float(SAMPLE_RATE):.2f}s, tier={meta['tier']}, lang={meta['lang']}) -> {args.output}")
+
+
+if __name__ == "__main__":
+    main()