Skip to content

Continue HTTP TTS reads after partial iteration - #313

Open
codebanditssss wants to merge 1 commit into
AssemblyAI:masterfrom
codebanditssss:audit/round3-python-tts-partial-read
Open

codebanditssss wants to merge 1 commit into
AssemblyAI:masterfrom
codebanditssss:audit/round3-python-tts-partial-read

Conversation

@codebanditssss

Copy link
Copy Markdown

Fixes #312

Changes

  • Retain one httpx body cursor per synchronous/asynchronous HTTP TTS stream so
    partial iteration can continue through read(), aread(), or a new wrapper.
  • Add seven real local HTTP regression/control cases covering continuation,
    EOF, independent streams, and cancelled-read context cleanup.
  • Bump the version to 1.10.16. Queued SDK PRs use 1.10.13–1.10.15, so merge order
    and final version need maintainer coordination.

No extra requests, new dependencies, or full-body buffering.

Validation

  • Identical upstream/fixed reproduction: 5 failed, 2 passed → 7 passed.
  • Python 3.12/latest installed dependencies: 91 focused, 1,041 offline unit tests passed.
  • Python 3.11/httpx 0.22/pydantic 1.10/websockets 11 floors: same counts passed.
  • Ruff 0.3.5 format/check, mypy 1.5.1 changed source, and diff whitespace checks passed.

No live provider tests ran. Cancellation coverage checks abort/context cleanup,
not resumable cancelled transports.

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

HTTP TTS read fails after partial audio iteration

1 participant