typecast-python 0.3.8__py3-none-any.whl → 0.3.10__py3-none-any.whl

typecast/async_client.py

@@ -1,3 +1,5 @@
+from __future__ import annotations
+
 import asyncio
 from pathlib import Path
 from typing import AsyncIterator, BinaryIO, Optional, Union
@@ -208,7 +210,8 @@ class AsyncTypecast:
                 seed=seed,
             )
         )
-        await asyncio.to_thread(output_path.write_bytes, response.audio_data)
+        loop = asyncio.get_running_loop()
+        await loop.run_in_executor(None, output_path.write_bytes, response.audio_data)
         return response
 
     async def text_to_speech_stream(
@@ -255,7 +258,7 @@ class AsyncTypecast:
                 error_text = await response.text()
                 self._handle_error(response.status, error_text)
 
-            async for chunk in response.content.iter_chunked(chunk_size):
+            async for chunk in response.content.iter_chunked(chunk_size):  # pragma: no branch
                 yield chunk
 
     async def text_to_speech_with_timestamps(

typecast/client.py

@@ -1,3 +1,5 @@
+from __future__ import annotations
+
 from pathlib import Path
 from typing import BinaryIO, Iterator, Optional, Union
 from urllib.parse import quote
@@ -195,10 +197,11 @@ class Typecast:
         if response.status_code != 200:
             self._handle_error(response.status_code, response.text)
 
+        content_type = response.headers.get("Content-Type", "audio/wav").lower()
         return TTSResponse(
             audio_data=response.content,
             duration=response.headers.get("X-Audio-Duration", 0),
-            format=response.headers.get("Content-Type", "audio/wav").split("/")[-1],
+            format="mp3" if "mp3" in content_type or "mpeg" in content_type else "wav",
         )
 
     def compose_speech(self) -> SpeechComposer:
@@ -207,7 +210,23 @@ class Typecast:
         Text passed to ``say()`` may include pause markup such as ``<|0.3s|>``.
         ``pause(seconds)`` also uses seconds, e.g. ``0.3`` for 300 ms.
         """
-        return SpeechComposer(self.text_to_speech)
+        return SpeechComposer(self.compose_text_to_speech)
+
+    def compose_text_to_speech(self, segments: list[dict]) -> TTSResponse:
+        response = self.session.post(
+            f"{self.host}/v1/text-to-speech/compose",
+            json={"segments": segments},
+            headers=self._request_headers(),
+            timeout=(10, 300),
+        )
+        if response.status_code != 200:
+            self._handle_error(response.status_code, response.text)
+        content_type = response.headers.get("Content-Type", "audio/wav").lower()
+        return TTSResponse(
+            audio_data=response.content,
+            duration=response.headers.get("X-Audio-Duration", 0),
+            format="mp3" if "mp3" in content_type or "mpeg" in content_type else "wav",
+        )
 
     def generate_to_file(
         self,

typecast/composer.py

@@ -1,10 +1,7 @@
 from __future__ import annotations
 
-import io
 import math
 import re
-import struct
-import wave
 from dataclasses import dataclass
 from typing import Callable, Literal, Optional, Union
 
@@ -31,19 +28,6 @@ class _SpeechPart:
     settings: "_ComposerSettings"
 
 
-@dataclass(frozen=True)
-class _WavSpec:
-    sample_rate: int
-    channels: int
-    sample_width: int
-
-
-@dataclass(frozen=True)
-class _ParsedWav:
-    spec: _WavSpec
-    samples: list[int]
-
-
 @dataclass(frozen=True)
 class _ComposerSettings:
     voice_id: Optional[str] = None
@@ -55,8 +39,8 @@ class _ComposerSettings:
 
 
 class SpeechComposer:
-    def __init__(self, text_to_speech: Callable[[TTSRequest], TTSResponse]):
-        self._text_to_speech = text_to_speech
+    def __init__(self, compose: Callable[[list[dict]], TTSResponse]):
+        self._compose = compose
         self._defaults = _ComposerSettings()
         self._parts: list[Union[_SpeechPart, _PausePart]] = []
 
@@ -124,41 +108,29 @@ class SpeechComposer:
         if not any(isinstance(part, _SpeechPart) for part in plan):
             raise ValueError("at least one speech segment is required")
 
-        output_format = (
-            self._defaults.output.audio_format if self._defaults.output else "wav"
-        )
+        formats = {
+            part.settings.output.audio_format
+            for part in plan
+            if isinstance(part, _SpeechPart)
+            and part.settings.output
+            and part.settings.output.audio_format
+        }
+        if len(formats) > 1:
+            raise ValueError("composed speech segments must use one audio format")
+        output_format = next(iter(formats), "wav")
         if output_format not in ("wav", "mp3"):
-            raise ValueError(f"unsupported composed speech output format: {output_format}")
+            raise ValueError(
+                f"unsupported composed speech output format: {output_format}"
+            )
 
-        wav_spec: Optional[_WavSpec] = None
-        output_samples: list[int] = []
+        segments: list[dict] = []
         for part in plan:
             if isinstance(part, _PausePart):
-                if wav_spec is None:
-                    raise ValueError("pause cannot be the first composed part")
-                output_samples.extend(
-                    [0] * _seconds_to_samples(part.seconds, wav_spec.sample_rate)
-                )
+                segments.append({"type": "pause", "duration_seconds": part.seconds})
                 continue
-
-            request = _settings_to_request(part.text, part.settings)
-            response = self._text_to_speech(request)
-            wav = _parse_wav(response.audio_data)
-            if wav_spec is not None and wav.spec != wav_spec:
-                raise ValueError("all composed WAV segments must use the same PCM format")
-            wav_spec = wav.spec
-            output_samples.extend(_trim_silence(wav.samples))
-
-        final_spec = wav_spec
-        assert final_spec is not None
-        wav_bytes = _encode_wav(output_samples, final_spec)
-        if output_format == "mp3":
-            raise ValueError("ffmpeg is required to encode composed speech as mp3")
-        return TTSResponse(
-            audio_data=wav_bytes,
-            duration=len(output_samples) / final_spec.sample_rate,
-            format="wav",
-        )
+            request = _settings_to_request(part.text, part.settings, output_format)
+            segments.append({"type": "tts", **request.model_dump(exclude_none=True)})
+        return self._compose(segments)
 
     def _build_plan(self) -> list[Union[_SpeechPart, _PausePart]]:
         plan: list[Union[_SpeechPart, _PausePart]] = []
@@ -173,7 +145,9 @@ class SpeechComposer:
                 if not parsed.text.strip():
                     continue
                 if not part.settings.voice_id:
-                    raise ValueError("voice_id is required for composed speech segments")
+                    raise ValueError(
+                        "voice_id is required for composed speech segments"
+                    )
                 if not part.settings.model:
                     raise ValueError("model is required for composed speech segments")
                 plan.append(_SpeechPart(text=parsed.text, settings=part.settings))
@@ -184,10 +158,14 @@ def parse_pause_markup(text: str) -> list[Union[_TextPart, _PausePart]]:
     parts: list[Union[_TextPart, _PausePart]] = []
     last_index = 0
     for match in _PAUSE_TOKEN.finditer(text):
-        if match.start() > last_index:
-            parts.append(_TextPart(kind="text", text=text[last_index : match.start()]))
-        parts.append(_PausePart(kind="pause", seconds=float(match.group(1))))
-        last_index = match.end()
+        seconds = float(match.group(1))
+        if math.isfinite(seconds) and seconds > 0:
+            if match.start() > last_index:
+                parts.append(
+                    _TextPart(kind="text", text=text[last_index : match.start()])
+                )
+            parts.append(_PausePart(kind="pause", seconds=seconds))
+            last_index = match.end()
     if last_index < len(text):
         parts.append(_TextPart(kind="text", text=text[last_index:]))
     return parts
@@ -207,69 +185,33 @@ def _merge_settings(
     )
 
 
-def _merge_output(base: Optional[Output], override: Optional[Output]) -> Optional[Output]:
+def _merge_output(
+    base: Optional[Output], override: Optional[Output]
+) -> Optional[Output]:
     if base is None and override is None:
         return None
-    data = base.model_dump(exclude_none=True, exclude_unset=True) if base is not None else {}
+    data = (
+        base.model_dump(exclude_none=True, exclude_unset=True)
+        if base is not None
+        else {}
+    )
     if override is not None:
         data.update(override.model_dump(exclude_none=True, exclude_unset=True))
     return Output(**data)
 
 
-def _settings_to_request(text: str, settings: _ComposerSettings) -> TTSRequest:
-    output = _merge_output(settings.output, Output(audio_format="wav"))
+def _settings_to_request(
+    text: str, settings: _ComposerSettings, output_format: str
+) -> TTSRequest:
+    output = _merge_output(settings.output, Output(audio_format=output_format))
     return TTSRequest(
         text=text,
         voice_id=settings.voice_id or "",
-        model=settings.model if isinstance(settings.model, TTSModel) else TTSModel(settings.model),
+        model=settings.model
+        if isinstance(settings.model, TTSModel)
+        else TTSModel(settings.model),
         language=settings.language,
         prompt=settings.prompt,
         output=output,
         seed=settings.seed,
     )
-
-
-def _parse_wav(data: bytes) -> _ParsedWav:
-    try:
-        with wave.open(io.BytesIO(data), "rb") as reader:
-            channels = reader.getnchannels()
-            sample_width = reader.getsampwidth()
-            sample_rate = reader.getframerate()
-            if channels != 1 or sample_width != 2:
-                raise ValueError("only mono 16-bit PCM WAV is supported for composed speech")
-            frames = reader.readframes(reader.getnframes())
-    except (EOFError, wave.Error) as exc:
-        raise ValueError("unsupported WAV data") from exc
-    samples = [
-        struct.unpack("<h", frames[offset : offset + 2])[0]
-        for offset in range(0, len(frames), 2)
-    ]
-    return _ParsedWav(
-        spec=_WavSpec(sample_rate=sample_rate, channels=channels, sample_width=sample_width),
-        samples=samples,
-    )
-
-
-def _encode_wav(samples: list[int], spec: _WavSpec) -> bytes:
-    payload = b"".join(struct.pack("<h", sample) for sample in samples)
-    out = io.BytesIO()
-    with wave.open(out, "wb") as writer:
-        writer.setnchannels(spec.channels)
-        writer.setsampwidth(spec.sample_width)
-        writer.setframerate(spec.sample_rate)
-        writer.writeframes(payload)
-    return out.getvalue()
-
-
-def _trim_silence(samples: list[int]) -> list[int]:
-    start = 0
-    end = len(samples)
-    while start < end and samples[start] == 0:
-        start += 1
-    while end > start and samples[end - 1] == 0:
-        end -= 1
-    return samples[start:end]
-
-
-def _seconds_to_samples(seconds: float, sample_rate: int) -> int:
-    return round(seconds * sample_rate)

typecast/models/tts.py

@@ -1,5 +1,7 @@
+from __future__ import annotations
+
 from enum import Enum
-from typing import Literal, Optional, Union
+from typing import List, Literal, Optional, Union
 
 from pydantic import BaseModel, ConfigDict, Field, model_validator
 
@@ -330,7 +332,7 @@ def _group_into_cues(
     word_mode=True: parts are joined with a single space.
     word_mode=False: parts are concatenated directly.
 
-    Returns list[(text, start, end)] tuples.
+    Returns List[(text, start, end)] tuples.
     """
     cues = []
     cur_text_parts = []
@@ -404,11 +406,11 @@ class TTSWithTimestampsResponse(BaseModel):
     audio: str = Field(description="Base64-encoded audio bytes.")
     audio_format: Literal["wav", "mp3"] = Field(description="Audio encoding format.")
     audio_duration: float = Field(description="Length of audio in seconds.")
-    words: Optional[list[AlignmentSegmentWord]] = Field(
+    words: Optional[List[AlignmentSegmentWord]] = Field(
         default=None,
         description="Word-level timestamps; null when granularity=char.",
     )
-    characters: Optional[list[AlignmentSegmentCharacter]] = Field(
+    characters: Optional[List[AlignmentSegmentCharacter]] = Field(
         default=None,
         description="Character-level timestamps; null when granularity=word.",
     )

typecast/models/voices.py

@@ -1,5 +1,7 @@
+from __future__ import annotations
+
 from enum import Enum
-from typing import Optional
+from typing import List, Optional
 
 from pydantic import BaseModel, Field
 
@@ -12,7 +14,7 @@ class VoicesResponse(BaseModel):
     voice_id: str
     voice_name: str
     model: str
-    emotions: list[str]
+    emotions: List[str]
 
 
 class GenderEnum(str, Enum):
@@ -54,7 +56,7 @@ class ModelInfo(BaseModel):
     """Model information with supported emotions"""
 
     version: TTSModel
-    emotions: list[str]
+    emotions: List[str]
 
 
 class VoiceV2Response(BaseModel):
@@ -62,10 +64,10 @@ class VoiceV2Response(BaseModel):
 
     voice_id: str
     voice_name: str
-    models: list[ModelInfo]
+    models: List[ModelInfo]
     gender: Optional[GenderEnum] = None
     age: Optional[AgeEnum] = None
-    use_cases: Optional[list[str]] = None
+    use_cases: Optional[List[str]] = None
 
 
 class RecommendedVoice(BaseModel):

{typecast_python-0.3.8.dist-info → typecast_python-0.3.10.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: typecast-python
-Version: 0.3.8
+Version: 0.3.10
 Summary: Official Typecast Python SDK - Convert text to lifelike speech using AI-powered voices
 Project-URL: Homepage, https://typecast.ai
 Project-URL: Documentation, https://typecast.ai/docs/overview
@@ -217,26 +217,30 @@ Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: Apache Software License
 Classifier: Operating System :: OS Independent
 Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.8
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
 Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
 Classifier: Topic :: Multimedia :: Sound/Audio :: Speech
 Classifier: Topic :: Software Development :: Libraries :: Python Modules
-Requires-Python: >=3.11
-Requires-Dist: aiohttp>=3.14.0
+Requires-Python: <3.15,>=3.8
+Requires-Dist: aiohttp>=3.9.0
 Requires-Dist: pydantic>=2.0.0
 Requires-Dist: requests>=2.28.0
 Requires-Dist: typing-extensions>=4.0.0
 Provides-Extra: dev
 Requires-Dist: aioresponses>=0.7.6; extra == 'dev'
-Requires-Dist: black>=23.0.0; extra == 'dev'
-Requires-Dist: flake8>=6.0.0; extra == 'dev'
+Requires-Dist: black<25.0.0,>=23.0.0; extra == 'dev'
+Requires-Dist: flake8<6.0.0,>=5.0.0; extra == 'dev'
 Requires-Dist: isort>=5.0.0; extra == 'dev'
-Requires-Dist: mypy>=1.0.0; extra == 'dev'
-Requires-Dist: pytest-asyncio>=0.21.0; extra == 'dev'
-Requires-Dist: pytest-cov>=7.0.0; extra == 'dev'
+Requires-Dist: mypy<1.11.0,>=1.0.0; extra == 'dev'
+Requires-Dist: pytest-asyncio<1.0.0,>=0.21.0; extra == 'dev'
+Requires-Dist: pytest-cov<6.0.0,>=5.0.0; extra == 'dev'
 Requires-Dist: pytest-mock>=3.14.0; extra == 'dev'
-Requires-Dist: pytest>=7.0.0; extra == 'dev'
+Requires-Dist: pytest<9.0.0,>=7.0.0; extra == 'dev'
 Description-Content-Type: text/markdown
 
 <div align="center">
@@ -250,7 +254,7 @@ Convert text to lifelike speech using AI-powered voices
 [![PyPI version](https://img.shields.io/pypi/v/typecast-python.svg?style=flat-square)](https://pypi.org/project/typecast-python/)
 [![coverage](https://img.shields.io/badge/coverage-100%25-brightgreen.svg?style=flat-square)](../docs/coverage-policy.md)
 [![License](https://img.shields.io/badge/license-Apache--2.0-blue.svg?style=flat-square)](LICENSE)
-[![Python](https://img.shields.io/badge/Python-3.9+-3776ab.svg?style=flat-square&logo=python&logoColor=white)](https://www.python.org/)
+[![Python](https://img.shields.io/badge/Python-3.8--3.14-3776ab.svg?style=flat-square&logo=python&logoColor=white)](https://www.python.org/)
 
 [Documentation](https://typecast.ai/docs) | [API Reference](https://typecast.ai/docs/api-reference) | [Get API Key](https://typecast.ai/developers/api/api-key)
 

typecast_python-0.3.10.dist-info/RECORD

@@ -0,0 +1,18 @@
+typecast/__init__.py,sha256=L6m0Z1jn1UgGjPp9AFiTwhHk9DVV_B4WrLfQDEfbcZI,1093
+typecast/_user_agent.py,sha256=R_0eSKpuP9jLp_wDJcJGFvFtUh9BZwEgsT1buVeHRAA,1786
+typecast/_voice_clone.py,sha256=TN2tbB3b5lC5uFCBnbERhz37bNJlbv6VZ-vj70EfTs4,3464
+typecast/async_client.py,sha256=Gidl-gj0zGMzZCPzmMRcr2GiYFi0cBmZOwrLhSVRL_Y,21700
+typecast/client.py,sha256=Ac6hcginhJheh6CcstzaMtyXzRbnSfUerI8rt6l_f58,21695
+typecast/composer.py,sha256=pkRSVIv99uuAwPyOXw7KSW7NAkqvAUK4hCS5pmBb7bA,7313
+typecast/conf.py,sha256=yfOHYZDvIshWWMriN5wE0GFPR8AX3zv4Rg2655OvA9Q,724
+typecast/exceptions.py,sha256=Y0ZzYebe8zOSOSAHbXfKR0G_RJgdmZXxi15Z7ZxPLIk,1568
+typecast/utils.py,sha256=XuNuX7gW8_CGKqZ-cv_tKlPVMPBluAYJBw2clwmjIMI,708
+typecast/models/__init__.py,sha256=OmaFdGpH6gK7EJyNgvbwqZHXVxjnWKJzynn7D3pfULo,1249
+typecast/models/error.py,sha256=XomIjx7jvlCjItqzJuCAT4mXC9jwTjxR8lLDUk6P8KA,152
+typecast/models/subscription.py,sha256=EIaAAo3cCRw8LYT_O6D9AVwxqIHrWCijzl4UTx7FZB8,894
+typecast/models/tts.py,sha256=xiO1e936vFr7vZqUErTEhexydoaMem6BcY_x1A1zFow,16495
+typecast/models/voices.py,sha256=3UeBLwAlaNbQQUrt7FO18zqczLoP7jmi2x68yaUtllU,2687
+typecast_python-0.3.10.dist-info/METADATA,sha256=z8W8o3SeRtx4C9PWgEvRCJ3fOcfoVTy2Utp0mqUvkOs,25786
+typecast_python-0.3.10.dist-info/WHEEL,sha256=lCkmxWfQsSc9CfIClYeavTdQeEX2toPqufh9gI35EQA,87
+typecast_python-0.3.10.dist-info/licenses/LICENSE,sha256=HvtJ-S89uUkuYmt-OvVk4MRxmzwtbn84__qJtSrGU2Q,11348
+typecast_python-0.3.10.dist-info/RECORD,,

{typecast_python-0.3.8.dist-info → typecast_python-0.3.10.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: hatchling 1.30.1
+Generator: hatchling 1.31.0
 Root-Is-Purelib: true
 Tag: py3-none-any

typecast_python-0.3.8.dist-info/RECORD

@@ -1,18 +0,0 @@
-typecast/__init__.py,sha256=L6m0Z1jn1UgGjPp9AFiTwhHk9DVV_B4WrLfQDEfbcZI,1093
-typecast/_user_agent.py,sha256=R_0eSKpuP9jLp_wDJcJGFvFtUh9BZwEgsT1buVeHRAA,1786
-typecast/_voice_clone.py,sha256=TN2tbB3b5lC5uFCBnbERhz37bNJlbv6VZ-vj70EfTs4,3464
-typecast/async_client.py,sha256=H3hCoGb6h0UyrnYUlQ0lhnpaCV-hxnZtsQKYlqaGPSk,21592
-typecast/client.py,sha256=VPZVRnB9_T4VkSlrHpKLRahhEaxIF05XUAk-skqYKz8,20850
-typecast/composer.py,sha256=r1kpRGlSJ1s4C1PkE7yIpGwnKtxYfIdFSLWaf8v1RAQ,9422
-typecast/conf.py,sha256=yfOHYZDvIshWWMriN5wE0GFPR8AX3zv4Rg2655OvA9Q,724
-typecast/exceptions.py,sha256=Y0ZzYebe8zOSOSAHbXfKR0G_RJgdmZXxi15Z7ZxPLIk,1568
-typecast/utils.py,sha256=XuNuX7gW8_CGKqZ-cv_tKlPVMPBluAYJBw2clwmjIMI,708
-typecast/models/__init__.py,sha256=OmaFdGpH6gK7EJyNgvbwqZHXVxjnWKJzynn7D3pfULo,1249
-typecast/models/error.py,sha256=XomIjx7jvlCjItqzJuCAT4mXC9jwTjxR8lLDUk6P8KA,152
-typecast/models/subscription.py,sha256=EIaAAo3cCRw8LYT_O6D9AVwxqIHrWCijzl4UTx7FZB8,894
-typecast/models/tts.py,sha256=W2Bp9M-xPJPODlk89P9CnLA2fwb__4SR4yoMkG_O8kA,16453
-typecast/models/voices.py,sha256=3yJD78QbZe3pTZv61GnB8up8qlIHKd1iSe5Bow-4ocU,2645
-typecast_python-0.3.8.dist-info/METADATA,sha256=5Gbl-QxbRCGnqUwn_8Yr2xssRMZD_3GvZghgWsUxByg,25530
-typecast_python-0.3.8.dist-info/WHEEL,sha256=mffPy8wBnZQn2VnJUU5jE99KsxaSfiyMHV9Yt0aLVxs,87
-typecast_python-0.3.8.dist-info/licenses/LICENSE,sha256=HvtJ-S89uUkuYmt-OvVk4MRxmzwtbn84__qJtSrGU2Q,11348
-typecast_python-0.3.8.dist-info/RECORD,,