[Fix] 220-django-channels-live-stt-python — pin deepgram-sdk to v6.1.1

[github-actions]

Summary

• Pin deepgram-sdk from loose >=4.0.0 to exact ==6.1.1 in requirements.txt
• The consumer code already uses the v6 API surface (listen.v1.connect, EventType, ListenV1Results, send_media, send_close_stream), so no source changes were needed
• All tests pass with the pinned version

Test plan

• Installed dependencies with pinned SDK version
• Ran test_example.py — all 5 tests pass
• Reviewer: confirm CI passes on this branch

🤖 Generated with Claude Code

[github-actions]

Code Review

Overall: APPROVED

Tests ran ✅

============================= test session starts ==============================
platform linux -- Python 3.11.15, pytest-9.0.2, pluggy-1.6.0
collected 5 items

tests/test_example.py::test_django_imports PASSED                        [ 20%]
tests/test_example.py::test_consumer_imports PASSED                      [ 40%]
tests/test_example.py::test_consumer_source_uses_deepgram PASSED         [ 60%]
tests/test_example.py::test_consumer_audio_forwarding PASSED             [ 80%]
tests/test_example.py::test_template_exists PASSED                       [100%]

============================== 5 passed in 0.72s ===============================

Integration genuineness

Pass. This is a Deepgram + Django Channels example. consumer.py imports AsyncDeepgramClient from the official SDK and makes a real live STT connection via listen.v1.connect(). The .env.example correctly lists DEEPGRAM_API_KEY. No third-party platform is involved — Django Channels is a framework, not an external API service.

Code quality

• ✅ Official Deepgram SDK (deepgram-sdk==6.1.1) — matches required Python SDK version
• ✅ tag="deepgram-examples" present on the Deepgram connection
• ✅ No hardcoded credentials — key read from os.environ["DEEPGRAM_API_KEY"]
• ✅ Error handling: on_error event handler, disconnect cleanup with try/except
• ✅ Tests import from src/ and verify the consumer's structure and configuration
• ✅ Credential check runs first (top of test file, before any SDK imports)
• ✅ No word-list assertions — tests verify structural properties, not specific transcripts

Documentation

• ✅ README covers what you'll build, env vars with where-to-get links, install/run instructions
• ✅ .env.example present and complete
• ✅ Architecture diagram explains the audio pipeline clearly

Change summary

Minimal, correct fix: pins deepgram-sdk>=4.0.0 → deepgram-sdk==6.1.1 in requirements.txt. The source code already used the v6 API surface (listen.v1.connect, EventType, ListenV1Results, send_media, send_close_stream), so no code changes were needed.

---

✓ All checks pass. Ready for merge.

---

Review by Lead on 2026-04-04

[github-actions]

Code Review

Overall: APPROVED

Tests ran ✅

tests/test_example.py::test_django_imports PASSED                        [ 20%]
tests/test_example.py::test_consumer_imports PASSED                      [ 40%]
tests/test_example.py::test_consumer_source_uses_deepgram PASSED         [ 60%]
tests/test_example.py::test_consumer_audio_forwarding PASSED             [ 80%]
tests/test_example.py::test_template_exists PASSED                       [100%]

5 passed in 0.71s

Integration genuineness

Pass — consumer.py imports AsyncDeepgramClient from the official SDK, opens a real live STT WebSocket via listen.v1.connect() with model nova-3, forwards audio via send_media(), and handles transcription results through EventType.MESSAGE callbacks. .env.example lists DEEPGRAM_API_KEY. Tests exit code 2 on missing credentials.

Code quality

• Official Deepgram SDK used (deepgram-sdk==6.1.1) — correctly pinned to the required version ✅
• tag="deepgram-examples" present on the Deepgram connection ✅
• No hardcoded credentials — DEEPGRAM_API_KEY read from os.environ ✅
• Error handling: on_error callback sends error JSON to browser, disconnect() cleans up with try/except ✅
• Tests import TranscriptionConsumer from src/ and verify structure, config, and source patterns ✅
• No word-list transcript assertions — tests are structural, appropriate for a WebSocket consumer ✅
• Credential check runs first in test_example.py before any imports ✅

Documentation

• README covers what you'll build, env vars with console link, install/run instructions, architecture diagram ✅
• .env.example present and complete ✅

---

✓ All checks pass. Ready for merge.

The PR correctly pins deepgram-sdk from >=4.0.0 to ==6.1.1 — the only change is in requirements.txt. All existing tests pass with the pinned version.

---

Review by Lead on 2026-04-04

[github-actions]

Code Review

Overall: APPROVED

Tests ran ✅

tests/test_example.py::test_django_imports PASSED                        [ 20%]
tests/test_example.py::test_consumer_imports PASSED                      [ 40%]
tests/test_example.py::test_consumer_source_uses_deepgram PASSED         [ 60%]
tests/test_example.py::test_consumer_audio_forwarding PASSED             [ 80%]
tests/test_example.py::test_template_exists PASSED                       [100%]

5 passed in 0.63s

Integration genuineness

Pass — Django Channels AsyncWebsocketConsumer handles browser WebSocket connections; audio is forwarded to Deepgram via the official SDK (AsyncDeepgramClient.listen.v1.connect()). No bypass or raw protocol usage.

Code quality

• ✅ Official Deepgram SDK used (deepgram-sdk==6.1.1 — correct pinned version)
• ✅ tag="deepgram-examples" present on the Deepgram connection
• ✅ No hardcoded credentials — DEEPGRAM_API_KEY read from environment
• ✅ Error handling covers connection errors and disconnect cleanup
• ✅ Tests import from src/ and verify actual consumer structure
• ✅ Credential check runs first (exits code 2 before SDK imports)
• ✅ No word-list assertions — tests verify source structure, not transcript content

Documentation

• ✅ README includes what-you'll-build, env vars with links, run instructions, architecture diagram
• ✅ .env.example present and complete

Change reviewed

The only change is pinning deepgram-sdk>=4.0.0 → deepgram-sdk==6.1.1 in requirements.txt. The source code already uses the v6 API surface (listen.v1.connect, EventType, ListenV1Results, send_media, send_close_stream) so no code changes were needed.

---

✓ All checks pass. Ready for merge.

---

Review by Lead on 2026-04-04