Conformer1-Demo / app.py
patloeber's picture
Upload 4 files
bfb5aad
import gradio as gr
import os
from helpers import make_header, upload_file, request_transcript, wait_for_completion, make_paragraphs_string
title = """<h1 align="center">🔥AssemblyAI: Conformer-1 Demo🔥</h1>"""
subtitle = """<h2 align="center">Automatic Speech Recognition using the AssemblyAI API</h2>"""
link = """<p align="center"><a href="https://www.assemblyai.com/blog/conformer-1/">Click here to learn more about the Conformer-1 model</a></p>"""
def submit_to_AAI(api_key,
radio,
audio_file,
mic_recording):
if radio == "Audio File":
audio_data = audio_file
elif radio == "Record Audio":
audio_data = mic_recording
header = make_header(api_key)
# 1. Upload the audio
upload_url = upload_file(audio_data, header, is_file=False)
# 2. Request transcript
transcript_response = request_transcript(upload_url, header)
transcript_id = transcript_response['id']
# 3. Wait for the transcription to complete
_, error = wait_for_completion(transcript_id, header)
if error is not None:
return error
# 4. Fetch paragraphs of transcript
return make_paragraphs_string(transcript_id, header)
def change_audio_source(radio):
if radio == "Audio File":
return [gr.Audio.update(visible=True),
gr.Audio.update(visible=False)]
elif radio == "Record Audio":
return [gr.Audio.update(visible=False),
gr.Audio.update(visible=True)]
with gr.Blocks(css = """#col_container {width: 1000px; margin-left: auto; margin-right: auto;}
#chatbot {height: 520px; overflow: auto;}""") as demo:
gr.HTML('<center><a href="https://www.assemblyai.com/"><img src="file/images/logo.png" width="180px"></a></center>')
gr.HTML(title)
gr.HTML(subtitle)
gr.HTML(link)
gr.HTML('''<center><a href="https://huggingface.co/spaces/assemblyai/Conformer1-Demo?duplicate=true"><img src="https://bit.ly/3gLdBN6" alt="Duplicate Space"></a>Duplicate the Space and run securely with your AssemblyAI API Key</center>''')
with gr.Column(elem_id="col_container"):
api_key = gr.Textbox(type='password', label="Enter your AssemblyAI API key here")
with gr.Box():
# Selector for audio source
radio = gr.Radio(["Audio File", "Record Audio"], label="Audio Source", value="Audio File")
# Audio object for both file and microphone data
audio_file = gr.Audio()
mic_recording = gr.Audio(source="microphone", visible=False)
gr.Examples([os.path.join(os.path.dirname(__file__),"audio/audio.mp3")], audio_file)
btn = gr.Button("Run")
out = gr.Textbox(placeholder="Your formatted transcript will appear here ...", lines=10)
# Changing audio source changes Audio input component
radio.change(fn=change_audio_source,
inputs=[radio],
outputs=[audio_file, mic_recording])
# Clicking "submit" uploads selected audio to AssemblyAI, performs requested analyses, and displays results
btn.click(fn=submit_to_AAI,
inputs=[api_key,radio,audio_file,mic_recording],
outputs=out)
demo.launch(debug=True)