Update app.py

app.py

@@ -1,7 +1,43 @@
-# Inference
-
 import gradio as gr
 from huggingface_hub import InferenceClient
+import pytesseract
+from PIL import Image
+from pypdf import PdfReader
+import ocrmypdf
+import os
+
+# Image to Text
+
+def fn_image_to_text(input_image):
+    return pytesseract.image_to_string(Image.open(input_image))
+
+# PDF to Text
+
+def fn_pdf_to_text(input_pdf):
+    reader = PdfReader(input_pdf)
+    
+    output_pdf = ""
+    for page in reader.pages:
+        output_pdf+=page.extract_text()
+  
+    image_count = 0
+    for page in reader.pages:
+        image_count += len(page.images)
+
+    if image_count > 0 and len(output_pdf) < 1000:
+        input_pdf_ocr = input_pdf.replace(".pdf", " - OCR.pdf")
+        ocrmypdf.ocr(input_pdf, input_pdf_ocr, force_ocr=True)
+    
+        reader = PdfReader(input_pdf_ocr)
+        output_pdf = ""
+        for page in reader.pages:
+            output_pdf+=page.extract_text()
+
+        os.remove(input_pdf_ocr)
+  
+    return output_pdf
+
+# Inference
 
 model_text = "google/gemma-2-27b-it"
 model_vision = "google/paligemma2-3b-pt-224"
@@ -11,19 +47,21 @@ client = InferenceClient()
 def fn_text(
     prompt,
     history,
+    input,
     #system_prompt,
     max_tokens,
     temperature,
     top_p,
 ):
+    messages = [{"role": "user", "content": output}]
     #messages = [{"role": "system", "content": system_prompt}]
-    #history.append(messages[0])
+    history.append(messages[0])
     
-    #messages.append({"role": "user", "content": prompt})
-    #history.append(messages[1])
+    messages.append({"role": "user", "content": prompt})
+    history.append(messages[1])
 
-    messages = [{"role": "user", "content": prompt}]
-    history.append(messages[0])
+    #messages = [{"role": "user", "content": prompt}]
+    #history.append(messages[0])
     
     stream = client.chat.completions.create(
         model = model_text,
@@ -43,6 +81,7 @@ app_text = gr.ChatInterface(
     fn = fn_text,
     type = "messages",
     additional_inputs = [
+        gr.File(type="filepath", label="Input"),
         #gr.Textbox(value="You are a helpful assistant.", label="System Prompt"),
         gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max Tokens"),
         gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),