Add randomize all button, improve UI interaction

Update choose_providers to make option B Hume only 30% of the time and remove check for presence of character description

Relax timeout configs for Hume TTS calls, add exponential backoff, add logs for logging call times

Fix missing expected argument warning caused by some lambdas not expecting an argument

Move custom css to global styles file

Fix styling for displaying results

Update UI layout and improve instructions content

Improve text length validation and make vote buttons flush with audio players

Update API integration code to be async

Update Anthropic integration to be async

Remove custom gradio theme in favor of using default theme

Update database layer and application logic for db interaction to be async

Fix types in app.py

Add '_' prefix to private methods throughout project, cleans up formatting and comments/docstrings

more explicit control flow

Fix vote button state management after vote button click

Move env var validation util to config, refactor to fix linting errors across project

Fix formatting in app.py

Move more business logic out of app.py, simplify state management for options

Remove unnecessary vote button state update on audioplayer 2 playback completing.

Remove unused imports

Improves UI content for clarity

Simplify UI logic in app.py, add logic for handling empty character description inputs.

Move business logic out of app.py, refactor code to call hume once specifying 2 generations instead of call Hume twice

Improve Hume errors, fix error message text in UI

Improve Anthropic error handling, add UnretryableAnthropicError

Add Dockerfile and .dockerignore, update server_name arg for gradio launch

improve 11labs error handling

Fix other button b bug

Fix bug with option b button

Update character description input label to sentence casing

Add missing json import in app.py

Update application flow to accept a character description and normalize a prompt for claude

Update design to accomodate easily adding other TTS providers in the future

Update README for with different start scripts, minor UI improvements to layout

Refactor tts integration functions to write audio to file and return file path, audio players to play mp3 file written to temp folder, fix audioplayer loading, remove unused imports

Update Hume integration to use OCTAVE TTS endpoint, update Elevenlabs integration to use voice design endpoint, no longer specify voice since voices are now generated

Add formatter (black) and format all code in src directory— using black default formatting configuration

Update option state to include the voice along with the provider, and update reporting of winner to include the voice name.

Restore encapsulation for ElevenLabs integration, update TTS functions to return the voice name in addition to the audio

Remove testing logic to force Hume to Hume comparison

Improve clarity of instructions, minor adjustment to UI component layout

Decouple text generation and speech synthesis and enable editing of text

Fix typo in documentation

Update UI to compare Hume vs Elevenlabs 50% of the time, and Hume vs Hume 50% of the time

Clean up docstrings across files

Remove docstring typos

Add UI constraint on prompt input for max character length

Update reset ui logic to reset audio players

Remove orphaned gradio components, clean up docstrings, fix types in app.py