Update app.py

app.py

@@ -1,4 +1,4 @@
-# app.py — HTR Space (Full Clean Version with Progress Fix)
+# app.py — HTR Space (Compact Version)
 import os, time
 from threading import Thread
 import gradio as gr
@@ -130,59 +130,29 @@ def save_as_audio(text):
 # ---------------- Gradio Interface ----------------
 with gr.Blocks(theme=gr.themes.Soft()) as demo:
     gr.Markdown("## ✍🏾 wilson Handwritten OCR")
-
     model_choice = gr.Radio(choices=list(MODEL_PATHS.keys()), value=list(MODEL_PATHS.keys())[0], label="Select OCR Model")
-
-    with gr.Row():  # Two-column layout
-        with gr.Column(scale=1):  # Inputs
-            query_input = gr.Textbox(label="Custom Prompt (optional)", placeholder="Leave empty for RAW structured output")
-            image_input = gr.Image(type="pil", label="Upload / Capture Handwritten Image", sources=["upload", "webcam"])
-
-            with gr.Accordion("⚙️ Advanced Options", open=False):
-                max_new_tokens = gr.Slider(1, 2048, value=MAX_NEW_TOKENS_DEFAULT, step=1, label="Max new tokens")
-                temperature = gr.Slider(0.1, 2.0, value=0.1, step=0.05, label="Temperature")
-                top_p = gr.Slider(0.05, 1.0, value=1.0, step=0.05, label="Top-p (nucleus)")
-                top_k = gr.Slider(0, 1000, value=0, step=1, label="Top-k")
-                repetition_penalty = gr.Slider(0.8, 2.0, value=1.0, step=0.05, label="Repetition penalty")
-
-            extract_btn = gr.Button("📤 Extract RAW Text", variant="primary")
-            clear_btn = gr.Button("🧹 Clear")
-
-        with gr.Column(scale=1):  # Outputs
-            gr.Markdown("## Output")
-            output_clean = gr.Textbox(label="Clean OCR Output", interactive=False, lines=18, show_copy_button=True)
-            with gr.Accordion("(Result.md)", open=False):
-                markdown_output = gr.Markdown()
-            pdf_file = gr.File(label="PDF File")
-            word_file = gr.File(label="Word File")
-            audio_file = gr.File(label="Audio File")
-            pdf_btn = gr.Button("⬇️ Download as PDF")
-            word_btn = gr.Button("⬇️ Download as Word")
-            audio_btn = gr.Button("🔊 Download as Audio")
-
-    # ---------------- Functions ----------------
-    def ocr_and_clean(image, model_choice, query, max_new_tokens, temperature, top_p, top_k, repetition_penalty, progress=None):
-        raw = ocr_image(
-            image, model_choice, query, max_new_tokens, temperature, top_p, top_k, repetition_penalty, progress=progress
-        )
-        clean_text = raw.replace("<|im_end|>", "").strip()
-        return clean_text, clean_text
-
-    # Connect OCR extraction with Progress
-    extract_btn.click(
-        fn=ocr_and_clean,
-        inputs=[image_input, model_choice, query_input, max_new_tokens, temperature, top_p, top_k, repetition_penalty, gr.Progress(track_tqdm=True)],
-        outputs=[output_clean, markdown_output]
-    )
-
-    # Export buttons
-    pdf_btn.click(fn=save_as_pdf, inputs=[output_clean], outputs=[pdf_file])
-    word_btn.click(fn=save_as_word, inputs=[output_clean], outputs=[word_file])
-    audio_btn.click(fn=save_as_audio, inputs=[output_clean], outputs=[audio_file])
-
-    # Clear button
-    clear_btn.click(fn=lambda: ("", None, "", MAX_NEW_TOKENS_DEFAULT, 0.1, 1.0, 0, 1.0, 1.0), 
-                    outputs=[output_clean, image_input, query_input, max_new_tokens, temperature, top_p, top_k, repetition_penalty])
+    with gr.Tab("🖼 Image Inference"):
+        query_input = gr.Textbox(label="Custom Prompt (optional)", placeholder="Leave empty for RAW structured output")
+        image_input = gr.Image(type="pil", label="Upload / Capture Handwritten Image", sources=["upload", "webcam"])
+        with gr.Accordion("⚙️ Advanced Options", open=False):
+            max_new_tokens = gr.Slider(1, 2048, value=MAX_NEW_TOKENS_DEFAULT, step=1, label="Max new tokens")
+            temperature = gr.Slider(0.1, 2.0, value=0.1, step=0.05, label="Temperature")
+            top_p = gr.Slider(0.05, 1.0, value=1.0, step=0.05, label="Top-p (nucleus)")
+            top_k = gr.Slider(0, 1000, value=0, step=1, label="Top-k")
+            repetition_penalty = gr.Slider(0.8, 2.0, value=1.0, step=0.05, label="Repetition penalty")
+        extract_btn = gr.Button("📤 Extract RAW Text", variant="primary")
+        clear_btn = gr.Button("🧹 Clear")
+        raw_output = gr.Textbox(label="📜 RAW Structured Output (exact as written)", lines=18, show_copy_button=True)
+        pdf_btn = gr.Button("⬇️ Download as PDF")
+        word_btn = gr.Button("⬇️ Download as Word")
+        audio_btn = gr.Button("🔊 Download as Audio")
+        pdf_file, word_file, audio_file = gr.File(label="PDF File"), gr.File(label="Word File"), gr.File(label="Audio File")
+
+        extract_btn.click(fn=ocr_image, inputs=[image_input, model_choice, query_input, max_new_tokens, temperature, top_p, top_k, repetition_penalty], outputs=[raw_output], api_name="ocr_image")
+        pdf_btn.click(fn=save_as_pdf, inputs=[raw_output], outputs=[pdf_file])
+        word_btn.click(fn=save_as_word, inputs=[raw_output], outputs=[word_file])
+        audio_btn.click(fn=save_as_audio, inputs=[raw_output], outputs=[audio_file])
+        clear_btn.click(fn=lambda: ("", None, "", MAX_NEW_TOKENS_DEFAULT, 0.1, 1.0, 0, 1.0), outputs=[raw_output, image_input, query_input, max_new_tokens, temperature, top_p, top_k, repetition_penalty])
 
 if __name__ == "__main__":
     demo.queue(max_size=50).launch(show_error=True)