Create app.py

app.py

@@ -0,0 +1,80 @@
+import gradio as gr
+import pdfminer.high_level
+import transformers
+from transformers import pipeline
+from pydub import AudioSegment
+import tempfile
+
+# Error handling function
+def handle_error(message):
+    print(f"Error: {message}")
+    return {"audio": None, "error": message}
+
+# Function to extract text from PDF
+def extract_text(pdf_path):
+    try:
+        with open(pdf_path, "rb") as file:
+            text = pdfminer.high_level.extract_text(file)
+            return text
+    except Exception as e:
+        return handle_error(f"Failed to extract text: {e}")
+
+# Function to split text into chunks
+def chunk_text(text, chunk_size=250):
+    chunks = []
+    for i in range(0, len(text), chunk_size):
+        chunk = text[i:i + chunk_size]
+        chunks.append(chunk)
+    return chunks
+
+# Function to perform text-to-speech and stitch audio
+def convert_to_speech(text_chunks, language="en", speaker="0"):
+    try:
+        model_name = "tts-es-es1"  # Replace with your chosen model
+        tts_pipeline = pipeline("text-to-speech", model=model_name)
+        audio_segments = []
+        for chunk in text_chunks:
+            audio = tts_pipeline(text=chunk, lang=language, speaker=speaker)
+            audio_segments.append(AudioSegment.from_mp3(audio["audio"]))
+        return audio_segments
+    except Exception as e:
+        return handle_error(f"Text-to-speech failed: {e}")
+
+# Function to save and return audio file
+def save_audio(audio_segments, filename, format="mp3"):
+    try:
+        combined_audio = audio_segments[0]
+        for segment in audio_segments[1:]:
+            combined_audio += segment
+        audio_path = tempfile.NamedTemporaryFile(suffix=f".{format}").name
+        combined_audio.export(audio_path, format=format)
+        return {"audio_path": audio_path}
+    except Exception as e:
+        return handle_error(f"Failed to save audio: {e}")
+
+# Gradio interface definition
+def pdf_to_speech(pdf_file):
+    # Extract text from PDF
+    text = extract_text(pdf_file)
+    if text["error"]:
+        return text["error"]
+
+    # Chunk text and convert to speech
+    text_chunks = chunk_text(text)
+    audio_result = convert_to_speech(text_chunks)
+    if audio_result["error"]:
+        return audio_result["error"]
+
+    # Save and return audio
+    audio_data = save_audio(audio_result)
+    return audio_data["audio_path"]
+
+# Create Gradio interface
+interface = gr.Interface(
+    fn=pdf_to_speech,
+    inputs=gr.FileInput(type="pdf"),
+    outputs=[gr.Audio(label="Play"), gr.File(label="Download")],
+)
+
+# Launch Gradio app
+interface.launch()