kokoro_tts: skip phonemes the vocab has no id for, as KModel does

[christopherthompson81]

Review copy on the fork. Upstream issue: 0xShug0#556.

encode_input_ids_and_count threw on any symbol missing from Kokoro's 114-entry vocab — including symbols the model's own G2P had just produced — which aborted the whole request.

The reference implementation does not. hexgrad/Kokoro's KModel tokenizes with input_ids = list(filter(None, map(vocab.get, phonemes))): it drops phonemes with no token id. Dropping gives a correct reading; throwing gives no audio at all.

Malformed UTF-8 above still throws — that is a real error. An unknown but well-formed phoneme is not. Each skip is recorded as a debug trace (kokoro.skipped_phoneme), so the information is kept without failing the request.

What it costs today, measured

40 real corpus sentences per language through the release kokoro_82m_q8_0, with the throw still in place:

language	result	symbol
en-us / en-gb	40 ok / 0 threw	—
es	35 ok / 5 threw	« U+00AB
fr-fr	34 ok / 6 threw	« U+00AB
it	39 ok / 1 threw	̪ U+032A
hi / pt-br	40 ok / 0 threw	—

So roughly 12% of Spanish and 15% of French prose failed outright — the vocab carries ( and ) but not the guillemets. English is affected too, just more rarely: eSpeak glottalises /t/ before a syllabic nasal, so button, kitten, written and forgotten produce a U+0329 the vocab lacks (100 of the first 40,000 dictionary words — but one anywhere aborts the request).

Verification

Every previously-failing case renders, with durations in line with words that always worked (button 1.00 s beside hidden 1.07 s), and 114/114 distinct English golden sentences render.

gate	result
C# bindings model test	ran=4 skipped=1 failures=0
audiocpp_c_api_path / _exports	Passed
audiocpp_c_api_model	Passed (146 s)
audiocpp_c_api_parity	PARITY OK across 5 families

kokoro_tts audio MATCH cli=79800@24000Hz peak=0.3383 api=79800@24000Hz peak=0.3383

79800 frames at 24 kHz is 3.325 s — byte-identical to what the unpatched engine produced for that sentence, and identical between the CLI and the C API. The fix is inert on text that never had an out-of-vocab symbol.

Notes for the reviewer

• Developed against pin 3eccab5, cherry-picked onto b46fe6b. The touched file and every dependent header are byte-identical between the two bases.
• The per-language numbers came from restoring the throw in a patched build and running real corpus text — not from diffing espeak-ng --ipa against the vocab. That proxy is misleading in both directions: it predicts 507 hyphen failures in French that never occur (espeak_text() ends with replace(out, "-", "")), and misses the guillemets that actually fail.
• A separate PR follows for the external-phonemizer option; it wants an ABI addition and is kept apart from this fix deliberately.

🤖 Generated with Claude Code

https://claude.ai/code/session_01EkxqpYvUbjCpRDnFiNiVfx

[christopherthompson81]

Folded into #9 as one PR, per review. Both commits are there unchanged — the bug fix is still the first commit and still reviewable on its own.

The reason the split stopped being worth it: the ABI addition turned out to be completing something already half-designed rather than proposing a new concept. string_list / float_list / path_list / audio_path_list have been in the spec's type vocabulary since schema v1, and the schema validates their declarations — there was simply no transport, and no spec had ever used one.