Create app.py

app.py

@@ -0,0 +1,84 @@
+import gradio as gr
+from transformers import pipeline
+from pydub import AudioSegment
+from docx import Document
+import fitz  # PyMuPDF
+import os
+import tempfile
+
+# Load Huggingface TTS pipeline
+tts = pipeline("text-to-speech", model="tts_en_cmu_us_slt")
+
+def extract_text_from_pdf(pdf_path):
+    """Extract text from a PDF file."""
+    pdf_document = fitz.open(pdf_path)
+    text = ""
+    for page_num in range(len(pdf_document)):
+        page = pdf_document[page_num]
+        text += page.get_text()
+    return text
+
+def extract_text_from_docx(docx_path):
+    """Extract text from a Word (.docx) file."""
+    doc = Document(docx_path)
+    text = [paragraph.text for paragraph in doc.paragraphs]
+    return "\n".join(text)
+
+def process_file(file):
+    """Process uploaded file and convert text to audio."""
+    file_extension = file.name.split('.')[-1].lower()
+    text = ""
+    
+    try:
+        # Save uploaded file temporarily
+        with tempfile.NamedTemporaryFile(delete=False) as temp_file:
+            temp_file.write(file.read())
+            temp_path = temp_file.name
+        
+        # Extract text based on file type
+        if file_extension == 'txt':
+            with open(temp_path, 'r') as f:
+                text = f.read()
+        elif file_extension == 'pdf':
+            text = extract_text_from_pdf(temp_path)
+        elif file_extension == 'docx':
+            text = extract_text_from_docx(temp_path)
+        else:
+            return "Unsupported file type!", None
+        
+        # Generate speech audio
+        audio_data = tts(text, return_tensors=True)
+        
+        # Save the audio as an MP3 file
+        output_path = os.path.join(tempfile.gettempdir(), "output.mp3")
+        audio_segment = AudioSegment.from_file(audio_data["waveform"], format="wav")
+        audio_segment.export(output_path, format="mp3")
+        
+        return "Text converted to audio successfully!", output_path
+    finally:
+        # Clean up temporary files
+        if os.path.exists(temp_path):
+            os.remove(temp_path)
+
+# Define Gradio interface
+def text_to_audio(file):
+    result_message, audio_file = process_file(file)
+    if audio_file:
+        return result_message, audio_file
+    else:
+        return result_message, None
+
+# Create Gradio interface
+interface = gr.Interface(
+    fn=text_to_audio,
+    inputs=gr.File(label="Upload your file (txt, pdf, or docx)"),
+    outputs=[
+        gr.Textbox(label="Result"),
+        gr.Audio(label="Generated Audio")
+    ],
+    title="Text-to-Audio Converter",
+    description="Upload a text, PDF, or Word file and convert its content to an audio file (American English). Download the audio after it's generated."
+)
+
+# Launch Gradio app
+interface.launch()