feat: support custom audio track input

[ao-anam]

Summary by cubic

Adds support for sending user audio via a new send_user_audio(...) API that takes raw 16-bit PCM. Default behavior is unchanged when no audio is sent.

• New Features

  • Added send_user_audio(audio_bytes, sample_rate, num_channels) to push user audio over WebRTC.
  • An audio transceiver (sendrecv) is always set up; the track is created lazily on first audio and closed on client shutdown.
  • Introduced UserAudioInputTrack with 20ms framing and backpressure; audio is queued only after connection to avoid stale data.
  • Exported AudioFrame; added pyav dependency.

• Refactors

  • Removed disable_input_audio option; no audio is sent unless send_user_audio(...) is called.

^{Written for commit a63666a. Summary will update on new commits.}

[Copilot]

Pull request overview

This PR adds support for custom audio track input to the Anam SDK, enabling developers to provide their own audio source for speech-to-text processing instead of relying on the default microphone input.

Changes:

• Added audio_input_track parameter to ClientOptions for configuring custom audio input
• Updated the WebRTC connection setup to use the custom audio track when provided
• Exported AudioStreamTrack and AudioFrame classes from the main package for easier access

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
src/anam/types.py	Added audio_input_track field to ClientOptions with documentation
src/anam/client.py	Passed audio_input_track option to the streaming client
src/anam/_streaming.py	Implemented custom audio track handling in WebRTC setup and cleanup
src/anam/init.py	Exported AudioStreamTrack and AudioFrame for public API access

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[cubic-dev-ai]

No issues found across 4 files

[cubic-dev-ai]

1 issue found across 1 file (changes from recent commits).

Prompt for AI agents (all issues)

Check if these issues are valid — if so, understand the root cause of each and fix them.


<file name="src/anam/_streaming.py">

<violation number="1" location="src/anam/_streaming.py:720">
P2: The upper-bound check uses 4800 instead of 48000, causing the warning to fire for all valid sample rates above 4.8 kHz. This makes the warning misleading and noisy for normal inputs.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.}

[robbie-anam]

This looks good to me but on the overall purpose,just clarifying that this is "user audio" e.g. like from a mic etc.

Just asking as if it's tts audio will need to go over websocket.