[ffeldhaus]

Description

Fixes #13041

Checklist

• I have followed Sample Guidelines from AUTHORING_GUIDE.MD
• Tests pass: nox -s py-3.9 (see Test Environment Setup)
• Lint pass: nox -s lint (see Test Environment Setup)
• Please merge this PR for me once it is approved

README.rst is not updated as it seems to be auto generated.

[ghost]

Hi team,

I'm currently reviewing this pull request and will provide my feedback in a few minutes. In the meantime, here's a summary of the changes for you and other reviewers to quickly get up to speed:

This pull request implements a real-time bidirectional streaming audio interface with Google Cloud Dialogflow CX. The primary goal is to create an example that captures audio from a user's microphone, streams it to Dialogflow CX for intent detection, and plays back the synthesized audio responses.

Here's a breakdown of the changes:

• Intent: Provide a complete example implementation of Streaming Detect Intent with continuous microphone input and audio output. This addresses issue #13041.
• Changes:
  • Added a new Python file: dialogflow-cx/streaming_detect_intent_infinite.py. This file contains the core logic for the bidirectional audio streaming with Dialogflow CX. It's a substantial addition (lines 1-727).
  • Updated dialogflow-cx/requirements.txt (lines 1-7) to include necessary dependencies: termcolor==2.5.0 and pyaudio==0.2.14. The google-cloud-dialogflow-cx library was also updated to version 1.38.0.
• Activity: The author, ffeldhaus, has indicated that tests and lint checks have passed.

The example is well-documented and includes instructions on setting up the environment and running the script. It also handles potential errors and provides informative logging.

I'll be checking for things like error handling, code style, and overall clarity of the example. I'll post my detailed review shortly.

---

And now, a little haiku to lighten the mood:

Code flows like a stream,
Dialog's voice, a clear response,
Smooth talk, all in sync.

[ghost]

This pull request introduces a continuous streaming example for Dialogflow CX, addressing issue #13041. The implementation is comprehensive and well-structured, demonstrating bidirectional streaming, audio input/output, and graceful exit handling. The code is generally clean and follows best practices. However, there are a few minor suggestions for improvement.

[glasnt]

Internal b/389902797

[snippet-bot]

Here is the summary of changes.

You are about to add 1 region tag.

• dialogflow-cx/streaming_detect_intent_infinite.py:54, tag dialogflow_streaming_detect_intent_infinite

---

This comment is generated by snippet-bot.
If you find problems with this result, please file an issue at:
https://github.com/googleapis/repo-automation-bots/issues.
To update this comment, add snippet-bot:force-run label or use the checkbox below:

• Refresh this comment

[davidcavazos]

This PR also needs tests. Since it's an "infinite" streaming CLI app, there's no need to test it end-to-end, but unit tests around the most important functions would be nice.

It's a little tricky since most functions are methods depending on the internal state of an object, so each test might need to create its own object.

I'm okay with more minimal testing around AudioIO since that's what streams from the microphone and can get tricky. However the Dialogflow code should be tested, it seems it could simply use a bytes generator with some fixed data, or maybe it could be read from a file?

[ffeldhaus]

This PR also needs tests. Since it's an "infinite" streaming CLI app, there's no need to test it end-to-end, but unit tests around the most important functions would be nice.

It's a little tricky since most functions are methods depending on the internal state of an object, so each test might need to create its own object.

I'm okay with more minimal testing around AudioIO since that's what streams from the microphone and can get tricky. However the Dialogflow code should be tested, it seems it could simply use a bytes generator with some fixed data, or maybe it could be read from a file?

I added a simple test similar to the existing test in streaming_detect_intent_partial_response_test.py reusing the some hello.wav resource file as input and the same existing Dialogflow Agent.

I reused and extended the MockPyAudio from transcribe_streaming_infinite_v2_test.py.

Please review again, all comments so far should be addressed.