refactor!: sdk release with new fern generated architecture #640
Conversation
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
Contributor
|
Important Review skippedMore than 25% of the files skipped due to max files limit. The review is being skipped to prevent a low-quality review. 189 files out of 299 files are above the max files limit of 100. Please upgrade to Pro plan to get higher limits. You can disable this status message by setting the Note Other AI code review bot(s) detectedCodeRabbit has detected other AI code review bot(s) in this pull request and will avoid duplicating their findings in the review comments. This may lead to a less comprehensive review. ✨ Finishing touches
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Contributor
There was a problem hiding this comment.
Pull request overview
This PR implements a major refactor of the Deepgram Python SDK, regenerating WebSocket components with a Fern-generated architecture to provide improved type safety and API alignment. The changes include removal of custom-maintained WebSocket types, telemetry infrastructure, and HTTP/client utilities in favor of auto-generated implementations.
Key Changes
- Complete removal of manually maintained WebSocket message types and event classes across Listen V1/V2, Speak V1, and Agent V1 APIs
- Deletion of telemetry infrastructure including batching handlers, protocol encoders, and instrumentation layers
- Removal of custom HTTP client wrappers and DeepgramClient implementations
- Addition of new SSE (Server-Sent Events) support with EventSource implementation
- Updates to HTTP retry logic with improved backoff strategies and rate limit handling
Reviewed changes
Copilot reviewed 294 out of 500 changed files in this pull request and generated 2 comments.
Show a summary per file
| File | Description |
|---|---|
src/deepgram/extensions/types/sockets/*.py |
Removed 23+ manually maintained WebSocket type files for Listen V1/V2, Speak V1, and Agent V1 |
src/deepgram/extensions/telemetry/*.py |
Deleted telemetry infrastructure including handlers, encoders, and models |
src/deepgram/extensions/core/*.py |
Removed HTTP/WebSocket instrumentation and telemetry event capture |
src/deepgram/client.py |
Deleted custom DeepgramClient and AsyncDeepgramClient implementations |
src/deepgram/core/http_client.py |
Enhanced retry logic with improved backoff, jitter, and rate limit handling |
src/deepgram/core/client_wrapper.py |
Updated SDK version to 5.3.1 and added async token support |
src/deepgram/core/http_sse/*.py |
Added new SSE support with EventSource, decoders, and charset handling |
src/deepgram/agent/v1/types/*.py |
Added 100+ auto-generated Agent V1 type files |
src/deepgram/core/pydantic_utilities.py |
Fixed RootModel validator mode for Pydantic v2 compatibility |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+73
to
+76
| def _add_symmetric_jitter(delay: float) -> float: | ||
| """Add symmetric jitter (±10%) for exponential backoff.""" | ||
| jitter_multiplier = 1 + (random() - 0.5) * JITTER_FACTOR | ||
| return delay * jitter_multiplier |
There was a problem hiding this comment.
The docstring states "±10%" but the implementation uses JITTER_FACTOR (0.2), which produces ±20% jitter when centered at 0.5. Update the docstring to "±20%" to match the actual behavior, or adjust the implementation if ±10% was intended.
|
|
||
|
|
||
| def _add_positive_jitter(delay: float) -> float: | ||
| """Add positive jitter (0-20%) to prevent thundering herd.""" |
There was a problem hiding this comment.
The docstring correctly states "0-20%" jitter which matches the implementation using JITTER_FACTOR (0.2). However, consider clarifying that this adds jitter in the range [delay, delay * 1.2] for precision.
Suggested change
| """Add positive jitter (0-20%) to prevent thundering herd.""" | |
| """ | |
| Add positive jitter (0-20%) to prevent thundering herd. | |
| The returned value will be in the range [delay, delay * 1.2]. | |
| """ |
- Fix test_listen_v1_media to pass required request parameter to transcribe_file - Fix test_speak_v1_audio to consume iterator returned by generate()
- Add client.py to fernignore (custom access_token and session ID support) - Add wire test files to fernignore (test implementation fixes) - Improve comments with clear explanations for each protected file
Add telemetry_opt_out and telemetry_handler parameters to DeepgramClient and AsyncDeepgramClient to maintain API compatibility with main branch signature. These parameters are accepted but ignored (no-op) to avoid requiring changes to generated SDK files while maintaining backwards compatibility.
lukeocodes
force-pushed
the
lo/v5-5-regen
branch
from
02f5059 to
f9f4511
Compare
- Add tests/manual to .fernignore - Add .pytest_cache/ and .env to .gitignore
Update send_listen_v_1_media, send_listen_v_2_media, and send_agent_v_1_media methods to accept bytes instead of str. Change internal implementation to use _send() instead of _send_model() to properly handle binary audio data.
Add socket client files to .fernignore with detailed comments explaining the manual fixes for binary audio data support. These files must be manually maintained because the generator incorrectly types media methods as str and uses _send_model() which fails on binary data.
Replace direct calls to private _send() method with public send_listen_v_1_media and send_listen_v_2_media methods in manual test files.
Add comprehensive WebSocket API reference documentation covering Listen V1, Listen V2, Speak V1, and Agent V1. Document method signatures, message types, event types, and usage examples for both sync and async clients. Update all media method examples to show binary data usage.
Rescue release-please setup from main branch for automated releases
- Add matrix strategy for Python 3.8-3.13 - Update setup-python action to v6 - Change mypy path to src/ - Add Docker verification step for WireMock tests - Update publish job to use POETRY_PYPI_TOKEN_PYPI
Add commitlint-based PR title validation to enforce conventional commits
- Add changelog-log workflow for webhook notifications - Add tests-daily workflow for scheduled test runs
Rescue migration documentation for v2-to-v3 and v3-to-v5 migrations
- Add 22 example files covering all major SDK functionality - Include authentication, transcription, TTS, voice agent, management API examples - Add fixtures directory with sample audio files - Add README.md with categorized examples and usage instructions - Add requirements.txt for example dependencies - Examples demonstrate streaming, error handling, and best practices - All examples validated against actual client functionality
- Fix import ordering (ruff isort) - Remove trailing whitespace - Remove trailing blank lines - Format code according to project standards
lukeocodes
changed the title
Add datetime wrapper in tests/wire/__init__.py to handle 'Z' suffix
for Python 3.8-3.10 compatibility. Python 3.10's datetime.fromisoformat()
doesn't support 'Z' suffix (added in 3.11), so we convert 'Z' to '+00:00'
automatically when the wire tests module is imported.
This allows generated test files to use datetime.fromisoformat('2024-01-15T09:30:00Z')
without modification, ensuring compatibility across all supported Python versions.
The wrapper class inherits from the original datetime class, so all other
datetime functionality remains unchanged.
Document the Python 3.8-3.10 datetime.fromisoformat compatibility fix that patches datetime.datetime.fromisoformat() to handle 'Z' suffix in ISO 8601 strings. This allows generated wire test files to work across all supported Python versions without modification.
Update version in pyproject.toml and SDK headers (User-Agent and X-Fern-SDK-Version) to 6.0.0-alpha.2.
- Add poetry to PATH after bootstrap in all jobs (compile, test, publish) to ensure poetry commands are available - Remove unnecessary matrix strategy from publish job - it should run once per tag push, not per Python version
Update version in pyproject.toml and SDK headers (User-Agent and X-Fern-SDK-Version) to 6.0.0-alpha.3.
Replace PYPI_API_TOKEN secret with OIDC authentication using pypa/gh-action-pypi-publish action in both ci.yml and release-please.yml workflows.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
This PR contains regenerations that includes major rewrites of all WebSocket components in the Deepgram Python SDK. The changes take further steps to automating all parts of the SDK, including the WebSocket implementation with a fully type-safe, generated codebase aligned with the latest API definitions.
Note: This regeneration aligns the SDK with the latest API definitions and sets the foundation for future improvements. The WebSocket rewrite provides a more robust, type-safe, and maintainable codebase.