Conversations

Introducing "Conversations" - a Python package designed to
explore the true potential of conversational analysis.
With its ability to transcribe, diarise, and generate visually appealing HTML reports,
"Conversation" empowers you to delve deeper into the intricacies of human communication.
Recognizing that conversations are brimming with meaning
conveyed not just through language,
but also through key acoustic features like loudness, intonation, and speech rate,
this package brings forth an indispensable tool for understanding and
interpreting the wealth of information embedded in everyday exchanges.
Conversations
enables you to analyse and comprehend
the complex world of human interaction.
Documentation
Comprehensive documentation, including overview, API, and examples,
can be found here.
Installation
pip install conversations
Usage
from pathlib import Path
from conversations import Conversation
audio_file = Path('/path/to/audio.mp4')
speaker_mapping={"0": "Alice", "1": "Bob", "2": "Sam"}
conversation = Conversation(recording=audio_file, speaker_mapping=speaker_mapping)
conversation.transcribe()
conversation.diarise()
conversation.save()
html_report = conversation.report()
with open('conversation.html', 'w') as f:
f.write(html_report.render())
text_report = conversation.export_text()
with open('conversation.txt', 'w') as f:
f.write(text_report)
summary = conversation.summarise()
with open('conversation-summary.txt', 'w') as f:
f.write(summary)
answer = conversation.query("What was the value of operating cash flow we were discussing?")
print(f"The answer to your query is: {answer}")
print(summary)