Remove dependency on `speechrecognition` library

[neubig]

The speechrecognition library is causing issues due to its inclusion of a tests module in its distributed package.

1. Understand the code here: https://github.com/Uberi/speech_recognition
2. Do a minimal re-implementation of only the functionality that is used by the openhands-aci directly within the openhands-aci library. Only include functionality that is currently used, include no other functionality
3. Remove the import of the speechrecognition library from pyproject.toml
4. Make sure that all github CI workflows still pass before sending a PR

[openhands-ai]

I'm on it! neubig can track my progress at all-hands.dev

[openhands-ai]

I've successfully fixed issue #136 by removing the dependency on the speechrecognition library. Here's a summary of the changes made:

1. Created a minimal implementation of the audio functionality needed by openhands-aci:

   • Implemented AudioFile, AudioData, and Recognizer classes in a new module openhands_aci/utils/audio.py
   • The implementation includes only the functionality that was being used in the codebase (audio file reading and Google speech recognition)

2. Updated the imports in md_converter.py to use our new implementation instead of the speechrecognition library

3. Removed the speechrecognition dependency from pyproject.toml

4. Added unit tests for the new audio module to ensure it works correctly

5. Verified that all unit tests pass with our implementation

6. Created a pull request (Fix issue #136: Remove dependency on speechrecognition library #137) that fixes issue Remove dependency on speechrecognition library #136

The PR has been successfully created and is available at: #137

This approach allows the openhands-aci library to maintain the same functionality without depending on the problematic speechrecognition library, which had issues with its tests module.