feat: simple demo

app.py

@@ -1,9 +1,10 @@
 import os
-import requests
+from time import time
 
 import gradio as gr
+import requests
+
 from languages import LANGUAGES
-from time import time
 
 GLADIA_API_KEY = os.environ.get("GLADIA_API_KEY")
 
@@ -21,8 +22,6 @@ ACCEPTED_LANGUAGE_BEHAVIOUR = [
 
 def transcribe(
     audio: str = None,
-    language_behaviour: str = ACCEPTED_LANGUAGE_BEHAVIOUR[2],
-    language: str = "english",
 ) -> dict:
     """
     This function transcribes audio to text using the Gladia API. 
@@ -30,26 +29,14 @@ def transcribe(
     Get your api key at gladia.io !
 
     Parameters:
-    audio_url (str): The URL of the audio file to transcribe. If audio_url is provided, audio file will be ignored.
     audio (str): The path to the audio file to transcribe.
-    video (str): The path to the video file. If provided, the audio field will be set to the content of this video.
-    language_behaviour (str): Determines how language detection should be performed. 
-        Must be one of [
-            "manual",
-            "automatic single language",
-            "automatic multiple languages"
-            ]
-        If "manual", the language field must be provided and the API will transcribe the audio in the given language.
-        If "automatic single language", the language of the audio will be automatically detected by the API 
-        but will force the transcription to be in a single language.
-        If "automatic multiple languages", the language of the audio will be automatically detected by the API for 
-        each sentence allowing code-switching over 97 languages.
-        
-    language (str): The language of the audio file. This field is ignored if language_behaviour is set to "automatic*".
 
     Returns:
     dict: A dictionary containing the transcribed text and other metadata about the transcription process. If an error occurs, the function returns a string with an error message.
     """
+    DEFAULT_MANUAL_LANGUAGE = "english"
+
+    language_behaviour = ACCEPTED_LANGUAGE_BEHAVIOUR[2]
 
     # if video file is there then send the audio field as the content of the video
     files = {
@@ -59,14 +46,12 @@ def transcribe(
     # priority given to the audio or video
     if audio:
         files["audio"] = (audio, open(audio, "rb"), "audio/wav")
-    else:
-        files["audio_url"] = ((None, audio_url),)
 
     # if language is manual then send the language field
     # if it's there for language_behaviour == automatic*
     # it will ignored anyways
     if language_behaviour == "manual":
-        files["language"] = (None, language)
+        files["language"] = (None, DEFAULT_MANUAL_LANGUAGE)
 
     start_transfer = time()
     response = requests.post(
@@ -109,28 +94,12 @@ iface = gr.Interface(
     """,
     fn=transcribe,
     inputs=[
-        gr.Audio(label="or Audio file to transcribe", source="upload", type="filepath"),
-        gr.Dropdown(
-            label="""Language transcription behaviour:\n
-        If "manual", the language field must be provided and the API will transcribe the audio in the given language.
-        If "automatic single language", the language of the audio will be automatically detected by the API 
-        but will force the transcription to be in a single language.
-        If "automatic multiple languages", the language of the audio will be automatically detected by the API for 
-        each sentence allowing code-switching over 97 languages.
-            """,
-            choices=ACCEPTED_LANGUAGE_BEHAVIOUR,
-            value=ACCEPTED_LANGUAGE_BEHAVIOUR[1]
-        ),
-        gr.Dropdown(
-            choices=sorted([language_name for language_name in LANGUAGES.keys()]),
-            label="Language (only if language behaviour is set to manual)",
-            value="english"
-        ),
+        gr.Audio(label="Audio file", source="upload", type="filepath"),
     ],
     outputs="json",
     examples=[
-        ["examples/good.will.hunting.wav", ACCEPTED_LANGUAGE_BEHAVIOUR[1], "english"],
-        ["examples/wolf.of.wall.street.wav", ACCEPTED_LANGUAGE_BEHAVIOUR[1], "english"],
+        ["examples/good.will.hunting.wav"],
+        ["examples/wolf.of.wall.street.wav"],
     ],
 )
 iface.queue()