Hello there! I've been hard at work analyzing 38 events from the past week to bring you this update.

This week, the team introduced new Whisper transcribers for audio files and made enhancements to the PromptNode object, while also squashing some pesky bugs and refactoring code.

Code Refactoring and Improvements 🧹

• Removed support for knowledge graph and updated dependencies (Massimiliano)
• Refactored code to generate eval results in a separate method (Bogdankostic)

Bug Squashing 🐛

• Fixed a bug causing invocation layers to fail when retrying on custom status codes (Silvano)
• Fixed a bug in CohereInvocationLayer that caused it not to return the resized prompt (Silvano)

PromptNode Enhancements 📝

• Added a new optional parameter to customize text generation in PromptNode object (Fanli)
• Simplified tests for PromptNode generation_kwargs (Silvano)

Whispering Transcribers 🎙️

• Added RemoteWhisperTranscriber for transcribing audio files using OpenAI's Whisper API (ZanSara)
• Introduced LocalWhisperTranscriber for transcribing audio files using OpenAI's Whisper model locally (ZanSara)