Upload folder using huggingface_hub

pipe.py

@@ -89,27 +89,21 @@ class AudioSpeechNERPipeline:
 
         return transcription, entities
 
-def create_ner_html(entities):
-    """
-    Create HTML representation of named entities
-    """
-    if not entities:
-        return "No named entities found."
-    
-    html = "<div style='background-color:#f0f0f0; padding:10px; border-radius:5px;'>"
-    html += "<h3>Named Entities:</h3>"
-    html += "<table style='width:100%; border-collapse:collapse;'>"
-    html += "<tr><th style='border:1px solid #ddd; padding:8px;'>Word</th><th style='border:1px solid #ddd; padding:8px;'>Entity Type</th></tr>"
+def replace_ner(entities):
+    processed_entities = []
     
     for entity in entities:
-        new_entity = labels[int(entity['entity'].split("_")[-1])]
-        html += f"<tr>" \
-                f"<td style='border:1px solid #ddd; padding:8px;'>{entity['word']}</td>" \
-                f"<td style='border:1px solid #ddd; padding:8px;'>{new_entity}</td>" \
-                f"</tr>"
-    
-    html += "</table></div>"
-    return html
+        number = int(entity['entity'].split("_")[-1])
+        
+        # Skip entities with number 0
+        if number == 0:
+            continue
+        
+        # Create a copy of the entity and update the label
+        updated_entity = entity.copy()
+        updated_entity['entity'] = labels[number]
+        processed_entities.append(updated_entity)
+    return processed_entities
 
 def process_audio_pipeline(audio):
     """
@@ -121,11 +115,9 @@ def process_audio_pipeline(audio):
     try:
         # Process the audio
         transcription, entities = pipeline.process_audio(audio)
+        entities = replace_ner(entities)
         
-        # Create HTML for entities
-        entities_html = create_ner_html(entities)
-        
-        return transcription, entities_html
+        return transcription, entities
     
     except Exception as e:
         return f"Error processing audio: {str(e)}", ""