fix: handle transcode for videos without audio track

[rubenxyz]

Description

Videos without an audio track (e.g. screen recordings, animations, silent clips) cause the transcoder to fail with Stream map 'a:0' matches no streams because -map a:0 was applied unconditionally.

Fix

• Probe for audio streams using ffprobe -select_streams a before building the ffmpeg command
• Only include -map a:0 and var_stream_map audio entries (v:{i},a:{i}) when has_audio=True
• Videos without audio now produce video-only HLS output

[ravirajsinh45]

The conditional -map a:0 / var_stream_map logic for audio-less inputs is correct and a genuinely useful fix. A few things before merge:

• Coordinate with #79 (overlapping change). This PR and #79 both edit the same region of packages/transcoder/ffmpeg_transcoder.py and neither is merged yet. GitHub reports each as individually mergeable because it only checks against main, not against each other — whichever lands second will need a rebase.
• The probe should fail safe. The new audio probe uses subprocess.run(...) + json.loads(stdout) unchecked. On a flaky/partial probe (truncated upload, S3 stall, expired presigned URL), ffprobe exits non-zero but still prints {}, so has_audio becomes False and audio is silently dropped from a file that has it while the job reports success. Recommend check=True (a genuine no-audio file still exits 0 with {"streams": []}, so the happy path is unaffected). If #79 lands first, route this through its _run helper instead.
• Tests. packages/transcoder/ has no tests, so the new has_audio=False branch is unverified. A unit test asserting the constructed argv (no -map a:0, var_stream_map = v:0 v:1 v:2) for both audio/no-audio inputs would lock it in.
• Minor. This adds a second full ffprobe round-trip over the presigned URL; a single -show_streams call can yield both metadata and audio presence.

[rubenxyz]

Thanks for the review @ravirajsinh45. Feedback addressed:

Probe fail-safe:

• Audio probe now uses _run() (from PR fix: include ffmpeg stderr in transcode error messages #79) which fail-fasts on non-zero exit with stderr in the error message. No more silent audio-drop on flaky ffprobe.

Tests:

• Added 5 unit tests in apps/api/tests/test_ffmpeg_transcoder.py:
  • test_transcode_with_audio_includes_audio_map — verifies -map a:0 and v:{i},a:{i} in var_stream_map when audio detected
  • test_transcode_without_audio_excludes_audio_map — verifies NO -map a:0 and video-only var_stream_map when no audio
  • test_run_raises_on_nonzero — RuntimeError with stderr
  • test_run_uses_errors_replace — errors='replace' passed to subprocess
  • test_run_returns_stdout — stdout returned on success

All 5 tests pass (python -m pytest apps/api/tests/test_ffmpeg_transcoder.py -v).

[ravirajsinh45]

Thanks for splitting this up, @rubenxyz — I can see you took the "separate PRs" suggestion to heart. 🙏 One thing to flag: #76, #77, #78, #79, and #80 currently contain the exact same diff (identical +474/-67 across the same 6 files). Each branch was created from the same combined working tree, so every PR has the whole bundle rather than just its own slice.

To keep the review in one place, I'm consolidating on #76 (which already has the full review) and closing this one as a duplicate — nothing is lost, all the changes live on #76.

If you'd like to genuinely split them later, the key is that each branch should contain only its own change. A clean way from main:

git checkout main && git pull
git checkout -b fix/one-thing
# bring over just the file(s)/hunks for that one change:
git checkout feature/your-bundle-branch -- path/to/relevant-file
git commit -m "fix: one thing"

…repeated per change, so each PR is independently reviewable. But that's optional — happy to take it all as one PR on #76 for now. Let's continue there!