Update app.py

app.py

@@ -12,25 +12,16 @@ client = Groq(
 model_name = "llama-3.1-70b-versatile"
 chat_groq = ChatGroq(model=model_name)
 
-def transcribe_audio(audio_filepath):
-    transcription_text = ""
-    if audio_filepath is None:
-        print("No se proporcionó audio.")
-        return transcription_text
-    try:
-        print(f"Transcribiendo audio desde: {audio_filepath}")
-        with open(audio_filepath, "rb") as file:
-            transcription = client.audio.transcriptions.create(
-                file=file,
-                model="whisper-large-v3",
-                response_format="json",
-                temperature=0.0
-            )
-        print(f"Respuesta de transcripción: {transcription}")
-        transcription_text = transcription.text
-    except Exception as e:
-        print(f"Error en transcripción de audio: {e}")
-    return transcription_text
+def transcribe_audio(audio):
+    filename = audio
+    with open(filename, "rb") as file:
+        transcription = client.audio.transcriptions.create(
+            file=(filename, file.read()),
+            model="whisper-large-v3",
+            response_format="json",
+            temperature=0.0
+        )
+    return transcription.text
 
 
 def extract_text_from_pdf(pdf_path):
@@ -79,7 +70,7 @@ def organize_clinical_record(current_text, transcription_text, pdf_text):
     {clinical_record_template}
 
     Borrador Actual del Registro Clínico:
-    {current_text}
+    {iterative_output}
 
     Nueva Información de Audio:
     {transcription_text}
@@ -91,52 +82,16 @@ def organize_clinical_record(current_text, transcription_text, pdf_text):
     """
 
     organized_text = chat_groq.invoke(prompt)
-    return organized_text.content
-
-def process_input(audio_filepath, pdfs, current_text):
-    transcription_text = transcribe_audio(audio_filepath)
-    # Definimos el límite máximo de palabras
-    max_words_per_prompt = 4500
-    debug_info = ""
+    return organized_text
 
-    # Preparamos la lista de textos de PDFs
-    pdf_texts = []
+def process_input(audio, pdfs):
+    transcription_text = transcribe_audio(audio)
+    pdf_text = ''
     if pdfs:
-        for pdf in pdfs:
-            pdf_content = extract_text_from_pdf(pdf)
-            pdf_texts.append((os.path.basename(pdf), pdf_content))
-            debug_info += f"Leído PDF: {pdf}\n"
-    else:
-        debug_info += "No se proporcionaron PDFs.\n"
-
-    # Combinamos los textos y contamos las palabras
-    updated_text = current_text
-    combined_texts = []
-    total_words = len(updated_text.split()) + len(transcription_text.split())
-
-    for pdf_name, pdf_content in pdf_texts:
-        pdf_words = len(pdf_content.split())
-        if total_words + pdf_words > max_words_per_prompt:
-            # Procesamos los textos actuales
-            pdf_text_combined = "\n".join([f"Contenido del PDF ({name}):\n{content}" for name, content in combined_texts])
-            updated_text = organize_clinical_record(updated_text, transcription_text, pdf_text_combined)
-            debug_info += f"Procesado lote de PDFs: {[name for name, _ in combined_texts]}\n"
-            # Reiniciamos los textos
-            combined_texts = [(pdf_name, pdf_content)]
-            total_words = len(updated_text.split()) + len(transcription_text.split()) + pdf_words
-        else:
-            combined_texts.append((pdf_name, pdf_content))
-            total_words += pdf_words
-
-    # Procesamos el último lote si hay PDFs pendientes
-    if combined_texts:
-        pdf_text_combined = "\n".join([f"Contenido del PDF ({name}):\n{content}" for name, content in combined_texts])
-        updated_text = organize_clinical_record(updated_text, transcription_text, pdf_text_combined)
-        debug_info += f"Procesado lote de PDFs: {[name for name, _ in combined_texts]}\n"
-
-    debug_info += f"Transcripción de Audio: {transcription_text}\n"
-
-    return updated_text, debug_info
+        pdf_text = extract_texts_from_pdfs(pdfs)
+    combined_text = transcription_text + "\n" + pdf_text
+    organized_record = organize_clinical_record(combined_text)
+    return organized_record.content
 
 theme = gr.themes.Base(
     primary_hue=gr.themes.Color(
@@ -149,15 +104,7 @@ theme = gr.themes.Base(
     neutral_hue="neutral",
 )
 
-current_state = gr.State()
-
-
-with gr.Blocks(theme=theme) as iface:
-    gr.Markdown("# Aplicación de Procesamiento de Audio y PDFs")
-    
-    iterative_output = gr.Textbox(
-        label="Registro Clínico Organizado",
-        value="""
+iterative_output = """
         MOTIVO DE CONSULTA: 
         
         ENFERMEDAD ACTUAL:
@@ -178,13 +125,19 @@ with gr.Blocks(theme=theme) as iface:
         ** Medicamentos:
         
         AYUDAS DIAGNOSTICAS:
-        """,
+        """
+
+with gr.Blocks(theme=theme) as iface:
+    gr.Markdown("# Aplicación de Procesamiento de Audio y PDFs")
+    
+    iterative_output = gr.Textbox(
+        label="Registro Clínico Organizado",
+        value= iterative_output,
         lines=20,
             )
     # Move the State inside the Blocks context
-    current_state = gr.State(value=iterative_output.value)
-    
-    audio_filepath = gr.Audio(sources=["microphone"], type="filepath", label="Entrada de Audio")
+    current_state = gr.State(value=iterative_output)
+    audio = gr.Audio(sources=["microphone"], type="filepath", label="Entrada de Audio")
     pdf_files = gr.File(file_types=[".pdf"], label="Subir PDFs (puedes subir múltiples archivos)", file_count="multiple")
     debug_output = gr.Textbox(label="Información de Depuración", lines=10)