Update app.py

app.py

@@ -1,4 +1,7 @@
+import os
 import gradio as gr 
+import pandas as pd
+
 from utilities.setup import get_files
 #import spaces
 from services.diarization import Diarizer
@@ -7,7 +10,7 @@ from speechbox import ASRDiarizationPipeline
 
 
 #@spaces.GPU
-def process_meeting(audio_input, num_speakers, speaker_names):
+def process_meeting(audio_input, num_speakers):
     """
     audio_input: filepath --> str
     num_speakers: number --> int
@@ -15,22 +18,54 @@ def process_meeting(audio_input, num_speakers, speaker_names):
     """
     
     print(audio_input) 
-    # first, pass it through a diarization stage.
-        #diarization_result, label_file = diarizer.run(audio_input, num_speakers)
+    # Get diarization and transcription pipelines
     diarization_pipeline = diarizer.get_pipeline()
-    # Next, pass it through a transctiption stage
     asr_pipeline = transcriber.get_pipeline()
-    # After, match the diarization with the transcription
+
+    # Pass it into speechbox for prediction and cleaning
     pipeline = ASRDiarizationPipeline(
         asr_pipeline=asr_pipeline, 
         diarization_pipeline=diarization_pipeline)
-    
-    # Finally, Clean up the docs.
+
     output = pipeline(audio_input,
                       num_speakers = num_speakers)
-    print(output)
-    return "a", None #diarization_result, label_file
 
+    # Clean User name
+    text = ""
+    for i in range(len(output)):
+        speaker = output[i]['speaker']
+        words = output[i]['text']
+        text += f"{speaker}: {words}\n"
+
+    return text
+
+
+def click_message():
+    return "Results loading. Go to next page!"
+
+
+def default_table():
+    return pd.DataFrame({
+        "Default": ["SPEAKER_00", "SPEAKER_01", "SPEAKER_02", "SPEAKER_03","SPEAKER_04"],
+        "Name": ["", "", "", "", ""],
+        "Title": ["", "", "", "", ""]
+    })
+
+def substitue_names(speaker_names, text):
+
+    df = speaker_names.itertuples(index=False)
+    for default, name, title in df:
+        if title <> "":
+            title = " ("+title+")"
+        
+        text = text.replace(default, f"{name}{title}")
+
+    temp_file = "transcript.txt"
+    with open(temp_file, "w") as file:
+        file.write(text)
+
+    #os.remove(temp_file)
+    return temp_file
 
 
 def main(conf):
@@ -51,30 +86,41 @@ def main(conf):
                                                   conf["session"]["max_speakers"])),
                                        label="Number of Speakers", 
                                        value=conf["session"]["min_speakers"])
-            speaker_names = gr.Dataframe(
-                label="Type your names and details. Your actual entries will be limited to the speakers you selected above.",
-                headers=["Name", "Supporting Details"],
-                datatype=["str", "str"],
-                row_count=(5,"fixed"),
-                col_count=(2, "fixed"),
-                type="pandas"
-            )
 
             process_button = gr.Button("Process")
+            output_box = gr.Textbox()
 
             
         with gr.TabItem(conf["layout"]["page_names"][2]):
             gr.Markdown("# 🔊 View and download your meeting transcript")
-            diarization_output = gr.Textbox(label="Diarization Output")
-            label_file_link = gr.File(label="Download DAW Labels")
-
+            transcription_output = gr.Textbox(label="Transcription Review")
+            
+            speaker_names = gr.Dataframe(
+                label="Match output names to desired names and titles/responsibility. Only enter values for ",
+                headers=["Default", "Name", "Title"],
+                datatype=["str", "str"],
+                row_count=(5,"fixed"),
+                col_count=(3, "fixed"),
+                type="pandas",
+                value=default_table(),
+            )
+            fix_button = gr.Button("Fix and Prepare Download")
+            label_file_link = gr.File(label="Download Cleaned Transcription")
 
         # Process
+        process_button.click(fn=click_message,
+                             outputs=output_box)
         
         process_button.click(
             fn=process_meeting,
-            inputs=[audio_input, num_speakers, speaker_names],
-            outputs=[diarization_output, label_file_link]
+            inputs=[audio_input, num_speakers],
+            outputs=[transcription_output]
+        )
+
+        fix_button.click(
+            fn=substitue_names,
+            inputs=[speaker_names, transcription_output],
+            outputs=[label_file_link]
         )
 
     demo.launch(server_name="0.0.0.0", allowed_paths=["/"])
@@ -85,7 +131,10 @@ def main(conf):
 if __name__ == "__main__":
     # get config
     conf = get_files.json_cfg()
+    
     # initialize diarizer
     diarizer = Diarizer(conf)
+    
     transcriber = Transcriber(conf)
+    # run main
     main(conf)