Update test_suite.py

test_suite.py

@@ -1,159 +1,130 @@
-import streamlit as st
 import pdfplumber
 import re
 from pdf_engine import extract_header_value, apply_value_replacement
 from parser_welspun import extract_welspun_items
 from parser_bkt import extract_bkt_items
 
-def render_universal_test_suite(selected_shipper):
-    st.markdown("---")
-    st.header("🧪 Universal Interactive Debugger & Test Suite")
-    st.caption("बिना एक्सेल जनरेट किए सीधे लाइव चेक करें कि किस कॉलम में exact क्या डेटा जाएगा।")
-
-    if "cached_pdf_lines" not in st.session_state or not st.session_state["cached_pdf_lines"]:
-        st.info("💡 कृपया पहले Section 2 में अपनी इनवॉइस PDF अपलोड करें, फिर यहाँ टेस्ट रन करें।")
-        return
-
-    pdf_lines = st.session_state.get("cached_pdf_lines", [])
-    pdf_text = st.session_state.get("cached_pdf_text", "")
-    
-    shipper_info = st.session_state["shipper_database"].get(selected_shipper, {})
+# 🚀 LayoutLMv3 & Document AI Integration Imports
+from transformers import LayoutLMv3Processor, LayoutLMv3ForTokenClassification
+import torch
+from PIL import Image
+import io
+
+def process_document_with_layoutlmv3(image_or_pdf_bytes):
+    """
+    ZeroGPU / GPU environment ke liye LayoutLMv3 processing function.
+    Yeh document ki visual bounding boxes aur text dono ko process karta hai.
+    """
+    try:
+        # Note: Model loading aur inference yahan ZeroGPU decorator ke antargat run hoga
+        processor = LayoutLMv3Processor.from_pretrained("microsoft/layoutlmv3-base")
+        # Model initialization aur inference logic yahan aage configure kiya ja sakta hai
+        return True, "LayoutLMv3 processor initialized successfully."
+    except Exception as e:
+        return False, str(e)
+
+def run_universal_test_suite_backend(selected_shipper, shipper_database, pdf_lines, pdf_text, test_category, target_field):
+    """
+    Streamlit-free backend test suite jo Gradio ke through single field ya row-by-row inspection run karta hai.
+    """
+    if not pdf_lines:
+        return "💡 कृपया पहले इनवॉइस PDF अपलोड करें, फिर टेस्ट रन करें।"
+
+    shipper_info = shipper_database.get(selected_shipper, {})
     header_rules = shipper_info.get("mapping_rules", {})
     item_rules = shipper_info.get("item_table_rules", {})
     assigned_parser = shipper_info.get("item_table_rule_name", "parser_welspun").strip().lower()
 
-    col_category, col_field = st.columns([1, 2])
-
-    with col_category:
-        test_category = st.radio("टेस्ट कैटेगरी चुनें:", ["Header Fields Rules", "Item Table Columns (G to AB)"])
-
-    with col_field:
-        if test_category == "Header Fields Rules":
-            field_options = list(header_rules.keys())
-            if not field_options:
-                st.warning("कोई Header Field उपलब्ध नहीं है।")
-                return
-            target_field = st.selectbox("जाँचने के लिए Header Field चुनें:", field_options)
+    if test_category == "Header Fields Rules":
+        if target_field not in header_rules:
+            return f"⚠️ Header Field '{target_field}' nahi mila."
+            
+        rule_data = header_rules[target_field]
+        ky = rule_data.get("keyword", "")
+        pos = rule_data.get("position", "Right (आगे)")
+        cl = rule_data.get("cell", "").strip()
+        m_mode = rule_data.get("match_mode", "Exact Word")
+        stop_kw = rule_data.get("stop_kw", "")
+        final_flt = rule_data.get("filter", "None")
+
+        final_val = extract_header_value(pdf_lines, pdf_text, ky, pos, m_mode, stop_kw, final_flt)
+
+        display_cell = cl if cl else 'Not Set'
+        if cl and cl.isalpha():
+            display_cell = f"{cl.upper()}2 (Dynamic Auto-Increment Row)"
+
+        result_summary = (
+            f"🔍 Inspection Result: {target_field}\n"
+            f"----------------------------------------\n"
+            f"📋 Parameters:\n"
+            f"- Target Excel Cell: {display_cell}\n"
+            f"- Keyword: {ky if ky else 'N/A'}\n"
+            f"- Match Mode: {m_mode}\n"
+            f"- Filter: {final_flt}\n\n"
+            f"🎯 Extracted Value: {final_val if final_val else 'BLANK / NOT FOUND'}"
+        )
+        return result_summary
+
+    else:
+        if target_field not in item_rules:
+            return f"⚠️ Item Field '{target_field}' nahi mila."
+            
+        rule_info = item_rules[target_field]
+        col_letter = rule_info.get("col", "").upper()
+        rule_type = rule_info.get("type", "PDF Row Item")
+        rule_val = rule_info.get("rule", "")
+
+        if "bkt" in assigned_parser:
+            parsed_items = extract_bkt_items(pdf_lines)
         else:
-            field_options = list(item_rules.keys())
-            if not field_options:
-                st.warning("कोई Item Table Column उपलब्ध नहीं है।")
-                return
-            target_field = st.selectbox("जाँचने के लिए Item Field चुनिए:", field_options)
-
-    if st.button("🚀 Run Live Single Field Inspection", type="primary", use_container_width=True):
-        st.write("---")
-        st.subheader(f"🔍 Inspection Result: `{target_field}`")
-
-        if test_category == "Header Fields Rules":
-            rule_data = header_rules[target_field]
-            ky = rule_data.get("keyword", "")
-            pos = rule_data.get("position", "Right (आगे)")
-            cl = rule_data.get("cell", "").strip()
-            m_mode = rule_data.get("match_mode", "Exact Word")
-            stop_kw = rule_data.get("stop_kw", "")
-            final_flt = rule_data.get("filter", "None")
-
-            final_val = extract_header_value(pdf_lines, pdf_text, ky, pos, m_mode, stop_kw, final_flt)
-
-            display_cell = cl if cl else 'Not Set'
-            if cl and cl.isalpha():
-                display_cell = f"{cl.upper()}2 (Dynamic Auto-Increment Row)"
-
-            res_col1, res_col2 = st.columns(2)
-            with res_col1:
-                st.markdown("#### 📋 Field Parameters")
-                st.write(f"* **Target Excel Cell:** `{display_cell}`")
-                st.write(f"* **Keyword:** `{ky if ky else 'N/A'}`")
-                st.write(f"* **Match Mode:** `{m_mode}`")
-                st.write(f"* **Filter:** `{final_flt}`")
-
-            with res_col2:
-                st.markdown("#### 🎯 Extracted Excel Value")
-                if final_val:
-                    st.success(f"✅ **Value to Excel:** `{final_val}`")
+            parsed_items = extract_welspun_items(pdf_lines, pdf_text=pdf_text)
+
+        if not parsed_items:
+            return "⚠️ इस PDF में कोई आइटम रो नहीं मिली या पार्सर से डेटा एक्सट्रेक्ट नहीं हुआ।"
+
+        preview_lines = [f"📊 Row-by-Row Preview for Column {col_letter}:"]
+        for idx, item in enumerate(parsed_items):
+            excel_row_num = 2 + idx
+            cell_target = f"{col_letter}{excel_row_num}"
+            nums = item.get("nums", [])
+            r_val_lower = str(rule_val).lower().strip()
+            f_name_lower = target_field.lower().strip()
+            extracted_cell_val = ""
+
+            if rule_type == "Constant Text":
+                extracted_cell_val = apply_value_replacement(rule_val, rule_val)
+            elif rule_type == "Excel Cell Reference":
+                extracted_cell_val = f"={rule_val}"
+            elif rule_type == "Smart Detection":
+                desc = item.get("description_text", "").upper()
+                extracted_cell_val = "PCS" if ("PCS" in desc or "PC" in desc) else (rule_val if rule_val else "SET")
+            else:
+                if "igst %" in r_val_lower or "igst rate" in f_name_lower:
+                    extracted_cell_val = nums[5] if len(nums) > 5 else ""
+                elif "igst amt" in r_val_lower or "igst amount" in f_name_lower:
+                    extracted_cell_val = nums[6] if len(nums) > 6 else ""
+                elif "hs" in r_val_lower or "ritc" in f_name_lower or "hs code" in r_val_lower:
+                    extracted_cell_val = item.get("hs_code", "")
+                elif "description" in r_val_lower or "description" in f_name_lower:
+                    extracted_cell_val = item.get("description_text", "")
+                elif "dbk" in r_val_lower or "drawback" in f_name_lower or col_letter == "S":
+                    extracted_cell_val = item.get("dbk_found", "")
+                elif "weight" in r_val_lower or "net wt" in f_name_lower:
+                    extracted_cell_val = nums[0] if len(nums) > 0 else ""
+                elif "qty" in r_val_lower or "quantity" in f_name_lower:
+                    extracted_cell_val = nums[1] if len(nums) > 1 else ""
+                elif "rate" in r_val_lower:
+                    extracted_cell_val = nums[2] if len(nums) > 2 else ""
+                elif "amount" in r_val_lower or "goods value" in f_name_lower:
+                    extracted_cell_val = nums[3] if len(nums) > 3 else ""
+                elif "taxable" in r_val_lower:
+                    extracted_cell_val = nums[4] if len(nums) > 4 else ""
                 else:
-                    st.error("⚠️ **Result:** BLANK / NOT FOUND")
+                    extracted_cell_val = rule_val
 
-        else:
-            # 🚀 SMART ROW-BY-ROW ITEM TABLE PREVIEW
-            rule_info = item_rules[target_field]
-            col_letter = rule_info.get("col", "").upper()
-            rule_type = rule_info.get("type", "PDF Row Item")
-            rule_val = rule_info.get("rule", "")
-
-            st.markdown("#### 📋 Item Rule Configuration")
-            st.write(f"* **Target Excel Column:** `{col_letter}`")
-            st.write(f"* **Rule Type:** `{rule_type}`")
-            st.write(f"* **Rule Value / Detail:** `{rule_val}`")
-
-            st.markdown("---")
-            st.markdown(f"#### 📊 Row-by-Row Preview for Column `{col_letter}`")
-
-            # सही पार्सर कॉल करके सारे आइटम्स निकालें
-            if "bkt" in assigned_parser:
-                parsed_items = extract_bkt_items(pdf_lines)
-            else:
-                parsed_items = extract_welspun_items(pdf_lines, pdf_text=pdf_text)
-
-            if parsed_items:
-                preview_table_data = []
-                for idx, item in enumerate(parsed_items):
-                    excel_row_num = 2 + idx  # मान लेते हैं डेटा Row 2 से शुरू होता है
-                    cell_target = f"{col_letter}{excel_row_num}"
-                    
-                    nums = item.get("nums", [])
-                    r_val_lower = str(rule_val).lower().strip()
-                    f_name_lower = target_field.lower().strip()
-                    extracted_cell_val = ""
-
-                    # सिमुलेशन लॉजिक ठीक वैसे ही जैसे प्रोसेसर में चलता है
-                    if rule_type == "Constant Text":
-                        extracted_cell_val = apply_value_replacement(rule_val, rule_val)
-                    elif rule_type == "Excel Cell Reference":
-                        extracted_cell_val = f"={rule_val}"
-                    elif rule_type == "Smart Detection":
-                        desc = item.get("description_text", "").upper()
-                        if "PCS" in desc or "PC" in desc:
-                            extracted_cell_val = "PCS"
-                        else:
-                            extracted_cell_val = rule_val if rule_val else "SET"
-                    else:
-                        # Standard PDF Row Item logic matching
-                        if "igst %" in r_val_lower or "igst rate" in f_name_lower:
-                            extracted_cell_val = nums[5] if len(nums) > 5 else ""
-                        elif "igst amt" in r_val_lower or "igst amount" in f_name_lower:
-                            extracted_cell_val = nums[6] if len(nums) > 6 else ""
-                        elif "hs" in r_val_lower or "ritc" in f_name_lower or "hs code" in r_val_lower:
-                            extracted_cell_val = item.get("hs_code", "")
-                        elif "description" in r_val_lower or "description" in f_name_lower:
-                            extracted_cell_val = item.get("description_text", "")
-                        elif "dbk" in r_val_lower or "drawback" in f_name_lower or col_letter == "S":
-                            extracted_cell_val = item.get("dbk_found", "")
-                        elif "weight" in r_val_lower or "net wt" in f_name_lower:
-                            extracted_cell_val = nums[0] if len(nums) > 0 else ""
-                        elif "qty" in r_val_lower or "quantity" in f_name_lower:
-                            extracted_cell_val = nums[1] if len(nums) > 1 else ""
-                        elif "rate" in r_val_lower:
-                            extracted_cell_val = nums[2] if len(nums) > 2 else ""
-                        elif "amount" in r_val_lower or "goods value" in f_name_lower:
-                            extracted_cell_val = nums[3] if len(nums) > 3 else ""
-                        elif "taxable" in r_val_lower:
-                            extracted_cell_val = nums[4] if len(nums) > 4 else ""
-                        else:
-                            extracted_cell_val = rule_val
-
-                        if "=" in str(rule_val):
-                            extracted_cell_val = apply_value_replacement(str(extracted_cell_val), str(rule_val))
-
-                    preview_table_data.append({
-                        "Item Sr No": idx + 1,
-                        "Excel Cell": cell_target,
-                        "Extracted Value": str(extracted_cell_val)
-                    })
-
-                # Streamlit में खूबसूरत टेबल दिखाना
-                st.dataframe(preview_table_data, use_container_width=True)
-                st.success(f"🎉 कुल {len(parsed_items)} आइटम्स की रो-बाय-रो रिपोर्ट सफलतापूर्वक जनरेट हो गई है!")
-            else:
-                st.warning("⚠️ इस PDF में कोई आइटम रो नहीं मिली या पार्सर से डेटा एक्सट्रेक्ट नहीं हुआ।")
+                if "=" in str(rule_val):
+                    extracted_cell_val = apply_value_replacement(str(extracted_cell_val), str(rule_val))
+
+            preview_lines.append(f"Row {idx+1} | Cell: {cell_target} | Value: {extracted_cell_val}")
+
+        return "\n".join(preview_lines)