Update app.py

app.py

@@ -1,44 +1,29 @@
+import gradio as gr
 import fitz  # PyMuPDF
 import json
-import gradio as gr
 from groq import Groq
 from dotenv import load_dotenv
 import os
 
-# Load environment variables
 load_dotenv()
-GROQ_API_KEY = os.getenv("GROQ_API_KEY")
 
 # Initialize Groq Client
-client = Groq(api_key=GROQ_API_KEY)
+client = Groq(api_key=os.getenv("GROQ_API_KEY"))
 
 def extract_form_fields(pdf_bytes):
-    if not pdf_bytes or len(pdf_bytes) == 0:
-        raise ValueError("Uploaded file is empty or not a valid PDF.")
-
-    try:
-        doc = fitz.open(stream=pdf_bytes, filetype="pdf")
-    except Exception as e:
-        raise ValueError(f"Failed to open PDF: {e}")
-
+    doc = fitz.open(stream=pdf_bytes, filetype="pdf")
     form_fields = {}
     for page in doc:
-        for widget in page.widgets():
-            key = widget.field_name
-            value = widget.field_value if widget.field_value else ""
-            form_fields[key] = value
-
+        widgets = page.widgets()
+        if widgets:
+            for widget in widgets:
+                key = widget.field_name
+                value = widget.field_value if widget.field_value else ""
+                form_fields[key] = value
     return form_fields
 
 def get_pdf_text(pdf_bytes):
-    if not pdf_bytes or len(pdf_bytes) == 0:
-        raise ValueError("Uploaded file is empty or not a valid PDF.")
-
-    try:
-        doc = fitz.open(stream=pdf_bytes, filetype="pdf")
-    except Exception as e:
-        raise ValueError(f"Failed to open PDF: {e}")
-
+    doc = fitz.open(stream=pdf_bytes, filetype="pdf")
     text = ""
     for page in doc:
         text += page.get_text()
@@ -57,53 +42,97 @@ Based on this, explain the meaning or expected value of each of the following fi
 Return your output in the following JSON format:
 {{ "field_name_1": "description", "field_name_2": "description", ... }}
     """
-
     response = client.chat.completions.create(
         model="llama3-8b-8192",
         messages=[{"role": "user", "content": prompt}]
     )
+    explanation = response.choices[0].message.content
+    return explanation
 
+def fill_pdf_fields(pdf_bytes, field_values):
+    doc = fitz.open(stream=pdf_bytes, filetype="pdf")
+    for page in doc:
+        widgets = page.widgets()
+        if widgets:
+            for widget in widgets:
+                key = widget.field_name
+                if key in field_values:
+                    widget.field_value = field_values[key]
+                    widget.update()
+    return doc.write()
+
+def analyze_pdf(file):
+    if file is None:
+        return "⚠️ Please upload a valid PDF.", "", {}, None
+
+    pdf_bytes = file.read()
     try:
-        explanation = response.choices[0].message.content
-        return explanation
-    except Exception as e:
-        return f"Failed to analyze fields: {str(e)}"
+        fields = extract_form_fields(pdf_bytes)
+        text = get_pdf_text(pdf_bytes)
+        explanation = get_field_details(fields, text)
+
+        if not fields:
+            return "⚠️ No form fields detected in the PDF.", "", {}, None
 
-def analyze_form(pdf_file):
-    if not pdf_file:
-        return "❌ No file provided.", "", ""
+        return json.dumps(fields, indent=2), explanation, fields, pdf_bytes
+    except Exception as e:
+        return f"Error: {str(e)}", "", {}, None
 
+def fill_and_download(field_inputs, original_pdf_bytes):
     try:
-        with open(pdf_file.name, 'rb') as f:
-            pdf_bytes = f.read()
+        filled_pdf_bytes = fill_pdf_fields(original_pdf_bytes, field_inputs)
+        return ("filled_form.pdf", filled_pdf_bytes)
+    except Exception as e:
+        return f"Error filling PDF: {str(e)}"
+
+def app():
+    with gr.Blocks() as demo:
+        gr.Markdown("## 📄 Smart Form Filler & Analyzer")
+
+        with gr.Row():
+            file_input = gr.File(label="Upload a PDF", file_types=[".pdf"])
+            analyze_btn = gr.Button("🔍 Analyze PDF")
+
+        extract_output = gr.Code(label="📋 Extracted Form Fields")
+        description_output = gr.Code(label="💡 Field Descriptions")
+
+        inputs_group = gr.Group(visible=False)
+        field_values = {}
+        input_boxes = []
+
+        original_pdf_bytes = gr.State()
+        filled_pdf_output = gr.File(label="⬇️ Download Filled PDF")
+        fill_btn = gr.Button("📝 Fill and Download PDF")
+
+        def display_fields_and_inputs(file):
+            extracted, desc, fields, pdf_bytes = analyze_pdf(file)
+            inputs_group.visible = bool(fields)
+            input_boxes.clear()
+            if fields:
+                for key, value in fields.items():
+                    box = gr.Textbox(label=key, value=value, interactive=True)
+                    input_boxes.append(box)
+                inputs_group.children = input_boxes
+            return extracted, desc, inputs_group.update(visible=True), pdf_bytes
+
+        analyze_btn.click(
+            fn=display_fields_and_inputs,
+            inputs=[file_input],
+            outputs=[extract_output, description_output, inputs_group, original_pdf_bytes]
+        )
 
-        fields = extract_form_fields(pdf_bytes)
-        pdf_text = get_pdf_text(pdf_bytes)
-        explanation = get_field_details(fields, pdf_text)
+        def get_filled_pdf(*args):
+            field_values = {box.label: value for box, value in zip(input_boxes, args[:-1])}
+            pdf_bytes = args[-1]
+            return fill_and_download(field_values, pdf_bytes)
 
-        field_output = (
-            json.dumps(fields, indent=2)
-            if fields else "⚠️ No form fields detected in this PDF."
+        fill_btn.click(
+            fn=get_filled_pdf,
+            inputs=input_boxes + [original_pdf_bytes],
+            outputs=[filled_pdf_output]
         )
-        return "✅ Analysis complete.", field_output, explanation
-    except Exception as e:
-        return f"❌ Error: {str(e)}", "", ""
-
-# Gradio Interface
-description = "📄 Upload a tax or registration form PDF. This tool extracts form fields and explains what each one likely means or requires using LLM."
-
-iface = gr.Interface(
-    fn=analyze_form,
-    inputs=gr.File(label="Upload PDF", type="filepath"),
-    outputs=[
-        gr.Textbox(label="Status"),
-        gr.Code(label="📋 Extracted Form Fields (JSON)", language="json"),
-        gr.Code(label="💡 Field Descriptions (JSON)", language="json"),
-    ],
-    title="📄 Form Field Analyzer",
-    description=description,
-    theme="default"
-)
+
+    return demo
 
 if __name__ == "__main__":
-    iface.launch()
+    app().launch()