import io import json import cv2 import requests from google import genai import os from datetime import datetime from flask import Flask, render_template, request, redirect, url_for,jsonify # Ensure redirect and url_for are imported import shutil # Add this import at the top of your file from flask_cors import CORS from PIL import Image import re app = Flask(__name__) # CORS configuration CORS(app) # Enable CORS for all routes BASE_URL = "https://jmn.sit.sigmentechnologies.com" # Store page1 data temporarily TEMP_STORE = {} #client = genai.Client(api_key="AIzaSyDVF5_plpZvdjsf5WbIX6TSPOt2Sps8hVc") #client = genai.Client(api_key="AIzaSyAvEn1aSat6yimDMhUHSW15aKBGeXc2zE8") def parse_gemini_json(text): try: # Remove markdown like ```json ``` text = re.sub(r"```json|```", "", text).strip() # Extract only JSON part (safer regex) match = re.search(r"\{[\s\S]*\}", text) if match: clean_json = match.group(0) return json.loads(clean_json) else: print("No JSON found in Gemini response") return {} except Exception as e: print("Invalid JSON from Gemini:", str(e)) return {} def extract_text_with_gemini(image_path): try: print(f"Processing file: {image_path}") # Read image as bytes with open(image_path, "rb") as f: image_bytes = f.read() response = client.models.generate_content( model="gemini-2.5-flash", contents=[ { "role": "user", "parts": [ {"text": """Extract all fields from the form and return STRICT JSON. Format: { "formStart": { "form_no": "MUST EXTRACT OR LEAVE EMPTY STRING" }, "studentDetails": { "student_name": "", "student_dob": "", "student_place_of_dob": "", "student_nationality": "", "student_contact_no": "", "student_marital": "", "student_religion": "", "student_cast": "", "student_gender": "", "student_mothertongue": "", "student_email": "", "student_blood_group": "" }, "permenent": { "permenent_address": "", "permenent_post_office": "", "permenent_police_station": "", "permenent_district": "", "permenent_state": "", "permenent_pincode": "" }, "present": { "present_address": "", "present_post_office": "", "present_police_station": "", "present_district": "", "present_state": "", "present_pincode": "" }, "parentDetails": { "father_name": "", "father_occupation": "", "father_mobile_no": "", "father_email_id": "", "mother_name": "", "mother_occupation": "", "mother_mobile_no": "", "mother_email_id": "" } } Rules: - Return ONLY valid JSON - Do NOT include ```json or ``` - Do NOT include explanation - Ensure valid JSON (double quotes, no trailing commas) - Fill all fields from the form """}, { "inline_data": { "mime_type": "image/jpeg", "data": image_bytes } } ] } ] ) print("RAW GEMINI RESPONSE:\n", response.text) return response.text except Exception as e: print("Gemini Error:", str(e)) return None def extract_text_with_back_gemini(image_path): try: print(f"Processing file: {image_path}") # Read image as bytes with open(image_path, "rb") as f: image_bytes = f.read() response = client.models.generate_content( model="gemini-2.5-flash-lite", contents=[ { "role": "user", "parts": [ {"text": """Extract all fields from the form and return STRICT JSON. Format: { "formStart": { "form_no": "MUST EXTRACT OR LEAVE EMPTY STRING" }, "localGuardian": { "guardian_name": "", "guardian_relationship": "", "guardian_mobile": "", "guardian_email": "", "guardian_occupation": "", "guardian_address": "", "guardian_post_office": "", "guardian_police_station": "", "guardian_district": "", "guardian_state": "", "guardian_pincode": "" }, "qualification": { "application_date": "", "accademic_session": "", "programme_study": "", "seat_allotment": "" }, "netug": { "neet_application_no": "", "neet_roll_no": "", "neet_percentile": "", "neet_total_marks": "", "neet_rank": "", "neet_category": "", "neet_quota": "" }, "x": { "board_0": "", "cgpa_0": "", "year_of_pass_0": "" }, "xii": { "board_1": "", "cgpa_1": "", "year_of_pass_1": "" }, "physics": { "marks_0": "", "percentage_0": "", "total_marks_0": "" }, "chemistry": { "marks_1": "", "percentage_1": "", "total_marks_1": "" }, "biology": { "marks_2": "", "percentage_2": "", "total_marks_2": "" }, "declaration": { "submission_date": "" } } Rules: - Only JSON - No explanation - Fill all fields from the form """}, { "inline_data": { "mime_type": "image/jpeg", "data": image_bytes } } ] } ] ) print("2") print("RAW GEMINI RESPONSE:\n", response.text) return response.text except Exception as e: print("Gemini Error:", str(e)) return None def extract_text_from_image(image_path): try: img = cv2.imread(image_path) file_bytes = io.BytesIO(cv2.imencode(".jpg", img)[1]) url_api = "https://api.ocr.space/parse/image" result = requests.post( url_api, files={"image.jpg": file_bytes}, data={ "apikey": "K83686776688957", "language": "eng", "isOverlayRequired": True, "detectOrientation": True, "scale": True, "isTable": True, "OCREngine": 2 } ) result = result.content.decode() result = json.loads(result) parsed_results = result.get("ParsedResults") if parsed_results: text_detected = parsed_results[0].get("ParsedText") else: text_detected = "" return text_detected except Exception as e: print(f"OCR extraction failed: {e}") return "" def replace_text(text): try: replacements = [ "REG. OFFICE: A2/15, KALYANI, DIST: NADIA,", "REG. OFFICE: A2/15, KALYANI, DIST.: NADIA,", "ALL FOR EDUCATION, FUCATION FOR ALF", "PIN 741235, WEST BENGAL, INDIA", "RESPECTED SIR/MADAM,", "FURNISHED BELOW FOR YOUR KIND CONSIDERATION.", "MBBS COURSE IN SESSION:", "PARTICULARS ARE", "MAIL ID: INFO.JMNEDUR@GMAIL.COM", "CONTACT NO: +91 9831111817", "CAMPUS: UTTAR PANCHPOTA, P.O & P. S. CHAKDAHA,", "DIST. NADIPIP INSTALLA, PIN 741222, WEST BENGAL, INDIA", "DIST. NADIA, PIN 741222, WEST BENGAL, INDIA", "JMN", "ADMISSION FORM", "THE PRINCIPAL,", "JMN MEDICAL COLLEGE,", "UTTAR PANCHPOTA, P.O. & P.S. CHAKDAHA,", "DIST. NADIA, PINCODE - 741222,", "WEST BENGAL, INDIA.", "PASTE RECENT", "PASSPORT-SIZE", "COLOUR", "PHOTOGRAPH OF", "THE CANDIDATE", "AND SIGN ACROSS", "THE PHOTOGRAPH.", "RESPECTED SIR / MADAM,", "I HEREBY APPLY FOR ADMISSION TO THE MBBS COURSE IN SESSION", "PARTICULARS ARE FURNISHED BELOW FOR YOUR KIND CONSIDERATION.", "KINDLY USE CAPITAL LETTERS ONLY", "STUDENT DETAILS:", "PARENT DETAILS:", "JMN MEDICAL COLLEGE", "WWW.JMNMEDICALCOLLEGE.ORG.IN", "DATE OF BIRTH:", "PLACE OF BIRTH:", "NATIONALITY:", "CONTACT NO:", "MARITAL STATUS:", "RELIGION:", "CASTE:", "GENDER:", "MOTHERTOUNGE:", "PERMANENT ADDRESS:", "P.O.:", "PO.:", "P.S.:", "DISTRICT:", "STATE:", "PINCODE:", "FATHER'S NAME:", "OCCUPATION:", "_ OCCUPATION:", "EMAIL-ID: _", "EMAIL-ID:", "EMAILID:_", "EMAILID:", "EMAIL ID:", "MOTHER'S NAME:", "APPLICATION DATE:", "REG. NO.:", "REG. DATE:", "BLOOD GROUP:", "DATE OF BIRTH:", "PLACE OF BIRTH:", "MOBILE NO.:", "MOBILE NO:", "MOBILE NO. :", "NATIONALITY:", "PRESENT ADDRESS:", "OCCUPATION", "AT THIS COLLEGE. NECESSARY", "SEAT ALLOTMENT NO.:", "ACADEMIC SESSION:", "ACADEMIC SESSION: _", "PROGRAMME OF STUDY:", "•", "QUALIFICATION DETAILS:", "RELATIONSHIP WITH STUDENT:", "HOUSE NUMBER:", "BLOCK:", "STREET", "VILLAGE OR TOWN:", "ADDRESS DETAILS (PERMANENT ADDRESS):", "ADDRESS DETAILS (PRESENT ADDRESS):", "LOCAL GUARDIAN DETAILS: (IN CASE OF OUTSTATION STUDENT)", "NEET UG DETAILS:", "PERCENTAGE (%)", "BIOLOGY", "CHEMISTRY", "PHYSICS", "SUBJECT", "CLASS XII DETAILS:", "YEAR OF PASSING", "%/CGPA|", "%/CGPA", "CLASS XII", "CLASS X", "INSTITUTE/BOARD", "PREVIOUS ACADEMIC QUALIFICATION:", "QUOTA:", "CATEGORY:", "ALL INDIA RANK:", "PERCENTILE:", "TOTAL MARKS OBTAINED:", "MARKS OBTAINED", "ROLL NO.:", "APPLICATION NO.:", "TOTAL MARKS:", "EXAMINATION", "TOTAL :", "DISTRICT:" ] cleaned_text = text for replacement in replacements: if replacement == "•": cleaned_text = cleaned_text.replace(replacement, ".") else: cleaned_text = cleaned_text.replace("MEDICAL COLLEGE", "").replace("ALL FOR EDUCATION, EDUCATION FOR ALL", "").replace("REG. OFFICE: A2/15, KALYANI, DIST.: NADIA,", "").replace("REG. OFFICE: A2/15, KALYANI, DIST: NADIA,", "").replace("ALL FOR EDUCATION, FUCATION FOR ALF", "").replace("RESPECTED SIR/MADAM,", "").replace("PARTICULARS ARE", "").replace("FURNISHED BELOW FOR YOUR KIND CONSIDERATION.", "").replace("PIN 741235, WEST BENGAL, INDIA", "").replace("MAIL ID: INFO.JMNEDUR@GMAIL.COM", "").replace("CONTACT NO: +91 9831111817", "").replace("CAMPUS: UTTAR PANCHPOTA, P.O & P. S. CHAKDAHA,", "").replace("DIST. NADIPIP INSTALLA, PIN 741222, WEST BENGAL, INDIA", "").replace("DIST. NADIA, PIN 741222, WEST BENGAL, INDIA", "FORM-START").replace("JMN", "").replace("ADMISSION FORM", "").replace("THE PRINCIPAL,", "").replace("JMN MEDICAL COLLEGE,", "").replace("UTTAR PANCHPOTA, P.O. & P.S. CHAKDAHA,", "").replace("DIST. NADIA, PINCODE - 741222,", "").replace("WEST BENGAL, INDIA.", "").replace("PASTE RECENT", "").replace("PASSPORT-SIZE", "").replace("COLOUR", "").replace("PHOTOGRAPH OF", "").replace("THE CANDIDATE", "").replace("AND SIGN ACROSS", "").replace("THE PHOTOGRAPH.", "").replace("RESPECTED SIR / MADAM,", "").replace("I HEREBY APPLY FOR ADMISSION TO THE MBBS COURSE IN SESSION", "MBBS COURSE IN SESSION:").replace("PARTICULARS ARE FURNISHED BELOW FOR YOUR KIND CONSIDERATION.", "").replace("KINDLY USE CAPITAL LETTERS ONLY", "").replace("JMN MEDICAL COLLEGE", "").replace("WWW.JMNMEDICALCOLLEGE.ORG.IN", "").replace("WWW.MEDICALCOLLEGE.ORG.IN", "").replace("AT THIS COLLEGE. NECESSARY", "").replace("@@", "@").replace("@GIMAIIL.COM", "@GMAIL.COM").replace("@GMIL.CO", "@GMAIL.COM").replace("P.O.:", "P.O.:").replace("P.O. :", "P.O.:").replace("PO:", "P.O.:").replace("P.O :", "P.O.:").replace("P.O:", "P.O.:").replace("P.S.:", "P.S.:").replace("P.S. :", "P.S.:").replace("PS.:", "P.S.:").replace("PS:", "P.S.:").replace("P.S :", "P.S.:").replace("P.S:", "P.S.:").replace("•", ".").replace("MOBILE NO.:", "MOBILE NO:").replace("MOBILE NO :", "MOBILE NO:").replace("WWW.JMMMEDICALCOLLEGE.ORG.IN", "") #.replace("STUDENT DETAILS:", "").replace("PARENT DETAILS:", "").replace("LOCAL GUARDIAN DETAILS: (IN CASE OF OUTSTATION STUDENT)", "").replace("QUALIFICATION DETAILS:", "") cleaned_text = cleaned_text.replace(replacement, "\n" + replacement) cleaned_lines = [line for line in cleaned_text.splitlines() if line.strip() != "" and line.strip() != "TO" and line.strip() != ","] cleaned_text = "\n".join(cleaned_lines) return cleaned_text except Exception as e: print(f"Text replacement failed: {e}") return text def extract_human_picture(image_path): try: # Load the image img = cv2.imread(image_path) # Enhance the image to reduce haze and improve quality denoise_strength = 3 # Customizable denoise strength img = cv2.fastNlMeansDenoisingColored(img, None, denoise_strength, denoise_strength, 7, 21) # Denoising to reduce haze # Increase contrast contrast_alpha = 1 # Customizable contrast control (1.0-3.0) brightness_beta = 5.5 # Customizable brightness control (0-100) img = cv2.convertScaleAbs(img, alpha=contrast_alpha, beta=brightness_beta) # Load the pre-trained Haar Cascade classifier for face detection face_cascade = cv2.CascadeClassifier(cv2.data.haarcascades + 'haarcascade_frontalface_default.xml') # Convert the image to grayscale for face detection gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) # Detect faces in the image faces = face_cascade.detectMultiScale(gray, scaleFactor=1.1, minNeighbors=5) # If a face is detected, extract the first one if len(faces) > 0: (x, y, w, h) = faces[0] # Get the coordinates of the first detected face # Extend the cropping area to include the full human picture x_extended = max(0, x - int(w * 0.5)) # Extend left by 50% of the width(extend left) y_extended = max(0, y - int(h * 0.5)) # Extend top by 50% of the height(extend Top) w_extended = int(w * 2.10) # New width: 210% of the original width(extend Right) h_extended = int(h * 2.75) # New height: 275% of the original height (extend down) face_image = img[y_extended:y_extended + h_extended, x_extended:x_extended + w_extended] # Crop the full human picture from the image # Create extracted folder if it doesn't exist os.makedirs('extracted', exist_ok=True) # Save the extracted face image cv2.imwrite('extracted/extracted_face.jpg', face_image) return 'extracted/extracted_face.jpg' return None # No face detected except Exception as e: print(f"Face extraction failed: {e}") return None def extract_signature(image_path): try: # Load the image img = cv2.imread(image_path) # Convert to grayscale gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) # Apply adaptive thresholding for better signature detection binary = cv2.adaptiveThreshold(gray, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY_INV, 11, 2) # Find contours of the signature contours, _ = cv2.findContours(binary, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE) # Filter contours based on area to find the signature if contours: # Sort contours by their y-coordinate (bottom-most first) sorted_contours = sorted(contours, key=lambda c: cv2.boundingRect(c)[1], reverse=True) # Iterate through sorted contours to find the signature for contour in sorted_contours: area = cv2.contourArea(contour) # Adjust the threshold based on the image size threshold = int(img.shape[0] * img.shape[1] * 0.001) # 0.1% of the image area if area > threshold: x, y, w, h = cv2.boundingRect(contour) # Reduce the border crop by adjusting the cropping dimensions border_reduction = 0.05 # Reduce the border by 10% x_extended = int(x - w * (border_reduction - 0.07)) # Extend left by 20% of the width y_extended = int(y + h * border_reduction) w_extended = int(w * (1 - border_reduction)) # Reduce width by 5% h_extended = int(h * (1 - 2 * border_reduction)) signature_image = img[y_extended:y_extended + h_extended, x_extended:x_extended + w_extended] # Crop the signature # Create extracted folder if it doesn't exist os.makedirs('extracted', exist_ok=True) # Save the extracted signature image cv2.imwrite('extracted/extracted_signature.jpg', signature_image) return 'extracted/extracted_signature.jpg' return None # No signature detected except Exception as e: print(f"Signature extraction failed: {e}") return None @app.route('/upload', methods=['POST']) def upload(): try: # Clean uploads folder if os.path.exists('./uploads'): shutil.rmtree('./uploads') os.makedirs('./uploads') file = request.files.get('file_1') key = request.form.get('key_1') if not file: print("[ERROR] No file_1 received in upload endpoint.") return jsonify({ "data": [], "img": {}, "sig": {}, "status": 500, "message": {"msg": "No file_1 received. Please select a file to upload."} }), 200 if not key: print("[ERROR] No key_1 received in upload endpoint.") return jsonify({ "data": [], "img": {}, "sig": {}, "status": 500, "message": {"msg": "No key_1 received. Please provide a key."} }), 200 # ✅ Save file file_path = f"./uploads/{file.filename}" file.save(file_path) print("File saved:", file_path) # ✅ Extract face extracted_face_path = extract_human_picture(file_path) if not extracted_face_path: print("No face detected") extracted_face_path = file_path # fallback # ✅ Save face to server server_path = f"/home/sigmente/public_html/sit/jmn/public/uploads/student_image/{key}_img.jpg" public_url = f"uploads/student_image/{key}_img.jpg" shutil.copy(extracted_face_path, server_path) print("Image stored at:", server_path) # ✅ Gemini extraction text = extract_text_with_gemini(file_path) if not text: print("Gemini returned empty") return jsonify({ "data": {}, "img": {}, "sig": {}, "status": 500, "message": {"msg": "Extraction failed"} }), 200 print("RAW GEMINI:", text) parsed_data = parse_gemini_json(text) # SAFE data structure fallback for formStart/form_no if "formStart" not in parsed_data or not isinstance(parsed_data.get("formStart"), dict): parsed_data["formStart"] = {} parsed_data["formStart"]["form_no"] = parsed_data["formStart"].get("form_no") or key print("FINAL JSON SENT:", json.dumps(parsed_data, indent=2)) # Save page1 data in memory (include the image URL too so upload_back can return it) TEMP_STORE[key] = { "data": parsed_data, "img": {"img": public_url} } return jsonify({ "data": parsed_data, # Include data for frontend "img": {"img": public_url}, "sig": {}, "status": 200, "message": {"msg": "Page 1 stored successfully"}, "next_step": "upload_back" }), 200 except Exception as e: print("ERROR:", str(e)) return jsonify({ "data": [], "img": {}, "sig": {}, "status": 500, "message": {"msg": "Server error"} }), 200 finally: # Cleanup try: if 'file_path' in locals() and os.path.exists(file_path): os.remove(file_path) except: pass @app.route('/upload_back', methods=['POST']) def upload_back(): try: # Delete the uploads directory if it exists if os.path.exists('./uploads'): shutil.rmtree('./uploads') # Create a new uploads directory os.makedirs('./uploads') # Set API key for back upload key1 = datetime.fromtimestamp(int("0x680b1c00", 16)) file = request.files.get('file_2') # Ensure the key matches what Laravel sends key = request.form.get('key_2') # Get the key from the form data if not file: print("[ERROR] No file_2 received in upload_back endpoint.") return jsonify({ "data": [], "img": {}, "sig": {}, "status": 500, "message": {"msg": "No file_2 received. Please select a file to upload."} }), 200 if not key: print("[ERROR] No key_2 received in upload_back endpoint.") return jsonify({ "data": [], "img": {}, "sig": {}, "status": 500, "message": {"msg": "No key_2 received. Please provide a key."} }), 200 if file and key: f = datetime.now() file_path = f"./uploads/back_{file.filename}" file.save(file_path) # Get the image data from the request # image_data = request.get_data() # # Save the image data to a temporary file # file_path = f"./uploads/temp_image.jpg" # with open(file_path, 'wb') as f: # f.write(image_data) # Extract signature from the uploaded image extracted_signature_path = extract_signature(file_path) if extracted_signature_path: print(f"Extracted signature image saved at: {extracted_signature_path}") # Save the extracted signature image with the key's text name #Save in production #Save in Sit server_path = f"/home/sigmente/public_html/sit/jmn/public/uploads/student_signature/{key}_sig.jpg" public_url = f"/uploads/student_signature/{key}_sig.jpg" #Save in local #new_file_path = f"C:/Users/Atanu Roy/Downloads/laravel/sig/{key}_img.jpg" shutil.copy(extracted_signature_path, server_path) # Copy the extracted image to the new path else: print("No signature found") public_url = None # Always save the back page image server_back_path =f"/home/sigmente/public_html/sit/jmn/public/uploads/back_image/{key}_back.jpg" public_back_url =f"uploads/back_image/{key}_back.jpg" os.makedirs(os.path.dirname(server_back_path), exist_ok=True) shutil.copy(file_path, server_back_path) # Debugging: Check if the file is saved print(f"Signature file saved at: {server_path}") # First try to extract text using Gemini text = extract_text_with_back_gemini(file_path) if text: parsed_data = parse_gemini_json(text) # SAFE data structure fallback for formStart/form_no if "formStart" not in parsed_data or not isinstance(parsed_data.get("formStart"), dict): parsed_data["formStart"] = {} if "form_no" not in parsed_data["formStart"] or parsed_data["formStart"].get("form_no") in [None, ""]: parsed_data["formStart"]["form_no"] = key print("FINAL JSON SENT:", json.dumps(parsed_data, indent=2)) # Get stored page1 data and image page1_record = TEMP_STORE.pop(key, {}) page1_data = page1_record.get("data", {}) page1_img = page1_record.get("img", {}) # Merge both form JSON structures while preserving the original form_no from page1 page1_form_start = page1_data.get("formStart", {}) if isinstance(page1_data.get("formStart"), dict) else {} parsed_form_start = parsed_data.get("formStart", {}) if isinstance(parsed_data.get("formStart"), dict) else {} final_form_no = page1_form_start.get("form_no") or parsed_form_start.get("form_no") or key merged_data = { **page1_data, **parsed_data, "formStart": { **page1_form_start, **parsed_form_start, "form_no": final_form_no } } print("MERGED FINAL JSON:", json.dumps(merged_data, indent=2)) front_url = page1_img.get("img") if page1_img else "" img_response = { "img": front_url } sig_response = {"sig": public_url} if public_url else {} return jsonify({ "data": merged_data, "img": img_response, "sig": sig_response, "status": 200 }), 200 else: print("Gemini extraction failed for back page.") return jsonify({"data": {}, "img": {}, "sig": {}, "status": 500, "message": {'msg':'Failed to extract text from Gemini for back page.'}}), 200 else: return jsonify({"data": [], "img": {}, "sig": {}, "status": 500, "message": {"msg": "Select Page 2"}}), 200 except Exception as e: # Delete the uploaded file if it exists if 'file_path' in locals() and os.path.exists(file_path): os.remove(file_path) return jsonify({"data": [], "img": {}, "sig": {}, "status": 500, "message": {'msg':'An error occurred during processing.'}}), 200 finally: # Delete the temporary image file after processing try: if 'file_path' in locals() and os.path.exists(file_path): os.remove(file_path) except: pass # ===== ROUTES AND MAIN ===== @app.route('/') def index(): try: return render_template('index.php') except Exception as e: print(f"Error rendering index: {e}") return "An error occurred", 200 @app.route('/favicon.ico') def favicon(): return '', 200 if __name__ == "__main__": app.run(debug=True, host='0.0.0.0', port=6003)