Upload 14 files
Browse files- gemini_requestCR7.py +103 -0
- gemini_requestcr4.py +135 -0
- json-B.py +96 -0
- json-CR1.py +143 -0
- json-CR2.py +164 -0
- json-CR3.py +165 -0
- json-CR4.py +145 -0
- json-CR5.py +0 -0
- json-CR6.py +138 -0
- json-CR7.py +112 -0
- json-V2.py +97 -0
- json-V3.py +104 -0
- json-v1.py +92 -0
- orch.py +73 -0
gemini_requestCR7.py
ADDED
|
@@ -0,0 +1,103 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
|
| 7 |
+
# Your Gemini API key
|
| 8 |
+
API_KEY = "your_gemini_api_key_here"
|
| 9 |
+
|
| 10 |
+
#Path to the document image you want to extract data from
|
| 11 |
+
image_path = r"C:\Users\ASUS\Downloads\OneDrive_1_7-30-2025\CR7\CS015431_CR.jpg"
|
| 12 |
+
|
| 13 |
+
# قراءة الصورة وتحويلها إلى base64
|
| 14 |
+
with open(image_path, "rb") as f:
|
| 15 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 16 |
+
|
| 17 |
+
prompt = """
|
| 18 |
+
يرجى استخراج الحقول التالية من مستند السجل التجاري (CR7) بالصورة، باللغة العربية فقط:
|
| 19 |
+
|
| 20 |
+
- اسم المنشأة
|
| 21 |
+
- نوع السجل
|
| 22 |
+
- حالة السجل
|
| 23 |
+
- الرقم الموحد للمنشأة
|
| 24 |
+
- رقم السجل التجاري
|
| 25 |
+
- اسم المالك
|
| 26 |
+
- نوع الكيان
|
| 27 |
+
- تاريخ الاصدار
|
| 28 |
+
- تاريخ الانتهاء
|
| 29 |
+
- قائمة المدراء
|
| 30 |
+
- المدينة
|
| 31 |
+
- الموقع الإلكتروني
|
| 32 |
+
- الانشطة التجارية
|
| 33 |
+
|
| 34 |
+
أرجو إعادة النتيجة بصيغة JSON بهذه المفاتيح فقط، وإذا أي حقل غير موجود فضع قيمته null:
|
| 35 |
+
|
| 36 |
+
{
|
| 37 |
+
"اسم المنشأة": null,
|
| 38 |
+
"نوع السجل": null,
|
| 39 |
+
"حالة السجل": null,
|
| 40 |
+
"الرقم الموحد للمنشأة": null,
|
| 41 |
+
"رقم السجل التجاري": null,
|
| 42 |
+
"اسم المالك": null,
|
| 43 |
+
"نوع الكيان": null,
|
| 44 |
+
"تاريخ الاصدار": null,
|
| 45 |
+
"تاريخ الانتهاء": null,
|
| 46 |
+
"قائمة المدراء": null,
|
| 47 |
+
"المدينة": null,
|
| 48 |
+
"الموقع الإلكتروني": null,
|
| 49 |
+
"الانشطة التجارية": null
|
| 50 |
+
}
|
| 51 |
+
"""
|
| 52 |
+
|
| 53 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 54 |
+
|
| 55 |
+
headers = {
|
| 56 |
+
"Content-Type": "application/json"
|
| 57 |
+
}
|
| 58 |
+
|
| 59 |
+
data = {
|
| 60 |
+
"contents": [
|
| 61 |
+
{
|
| 62 |
+
"role": "user",
|
| 63 |
+
"parts": [
|
| 64 |
+
{"text": prompt},
|
| 65 |
+
{
|
| 66 |
+
"inline_data": {
|
| 67 |
+
"mime_type": "image/jpeg",
|
| 68 |
+
"data": image_b64
|
| 69 |
+
}
|
| 70 |
+
}
|
| 71 |
+
]
|
| 72 |
+
}
|
| 73 |
+
]
|
| 74 |
+
}
|
| 75 |
+
|
| 76 |
+
response = requests.post(url, headers=headers, json=data)
|
| 77 |
+
|
| 78 |
+
try:
|
| 79 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 80 |
+
|
| 81 |
+
match = re.search(r"```json\s*(\{.*?\})\s*```", response_text, re.DOTALL)
|
| 82 |
+
if match:
|
| 83 |
+
json_text = match.group(1)
|
| 84 |
+
result = json.loads(json_text)
|
| 85 |
+
|
| 86 |
+
# حفظ النتيجة بنفس اسم ملف الصورة وداخل نفس المجلد
|
| 87 |
+
folder = os.path.dirname(image_path)
|
| 88 |
+
base_name = os.path.splitext(os.path.basename(image_path))[0]
|
| 89 |
+
output_file = os.path.join(folder, base_name + ".json")
|
| 90 |
+
|
| 91 |
+
with open(output_file, "w", encoding="utf-8") as f:
|
| 92 |
+
json.dump(result, f, ensure_ascii=False, indent=2)
|
| 93 |
+
|
| 94 |
+
print(f"✅ تم حفظ النتيجة في ملف: {output_file}")
|
| 95 |
+
print(json.dumps(result, indent=2, ensure_ascii=False))
|
| 96 |
+
else:
|
| 97 |
+
print("❌ لم أتمكن من استخراج JSON نظيف من الرد:")
|
| 98 |
+
print(response_text)
|
| 99 |
+
|
| 100 |
+
except Exception as e:
|
| 101 |
+
print(f"❌ حدث خطأ أثناء المعالجة: {e}")
|
| 102 |
+
print("🔴 الرد الكامل من Gemini:")
|
| 103 |
+
print(response.text)
|
gemini_requestcr4.py
ADDED
|
@@ -0,0 +1,135 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
|
| 7 |
+
# Your Gemini API key
|
| 8 |
+
API_KEY = "your_gemini_api_key_here"
|
| 9 |
+
|
| 10 |
+
|
| 11 |
+
#Path to the document image you want to extract data from
|
| 12 |
+
image_path = r"C:\Users\ASUS\Downloads\CR_OCR_Cls.v4-original_images.folder\train\CR4\CS000130_CR_jpg.rf.b2528e60546ec399c3088ac69c6a66e2.jpg"
|
| 13 |
+
|
| 14 |
+
with open(image_path, "rb") as f:
|
| 15 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 16 |
+
|
| 17 |
+
prompt = """
|
| 18 |
+
Extract the following fields from the CR4 document image. Return both Arabic and English text where available:
|
| 19 |
+
|
| 20 |
+
الرقم الموحد
|
| 21 |
+
رقم المنشأة
|
| 22 |
+
التاريخ
|
| 23 |
+
|
| 24 |
+
الاسم التجاري للشركة
|
| 25 |
+
نوعها
|
| 26 |
+
جنسيتها
|
| 27 |
+
مدة الشركة
|
| 28 |
+
تبدأ من
|
| 29 |
+
وتنتهي في
|
| 30 |
+
مركزها الرئيسي
|
| 31 |
+
هاتف
|
| 32 |
+
الرمز البريدي
|
| 33 |
+
|
| 34 |
+
النشاط
|
| 35 |
+
رأس المال
|
| 36 |
+
المديرون
|
| 37 |
+
سلطات المدير/المديرون
|
| 38 |
+
يشهد مكتب السجل التجاري بمدينة
|
| 39 |
+
بأنه تم تسجيل المؤسسة المذكورة أعلاة بمدينة
|
| 40 |
+
وتنتهي صلاحية الشهادات في
|
| 41 |
+
بموجب الإيصال رقم
|
| 42 |
+
وتاريخ
|
| 43 |
+
|
| 44 |
+
Return as JSON with keys:
|
| 45 |
+
{
|
| 46 |
+
"رقم_موحد": ...,
|
| 47 |
+
"رقم_المنشأة": ...,
|
| 48 |
+
"التاريخ": ...,
|
| 49 |
+
"الاسم_التجاري": ...,
|
| 50 |
+
"نوعها": ...,
|
| 51 |
+
"جنسيتها": ...,
|
| 52 |
+
"مدة_الشركة": ...,
|
| 53 |
+
"تبدأ_من": ...,
|
| 54 |
+
"تنتهي_في": ...,
|
| 55 |
+
"مركزها_الرئيسي": ...,
|
| 56 |
+
"هاتف": ...,
|
| 57 |
+
"الرمز_البريدي": ...,
|
| 58 |
+
"النشاط": ...,
|
| 59 |
+
"رأس_المال": ...,
|
| 60 |
+
"المديرون": ...,
|
| 61 |
+
"سلطات_المدير": ...,
|
| 62 |
+
"يشهد_مكتب_السجل": ...,
|
| 63 |
+
"تم_تسجيل_المؤسسة": ...,
|
| 64 |
+
"تنتهي_صلاحية_الشهادة": ...,
|
| 65 |
+
"الإيصال_رقم": ...,
|
| 66 |
+
"الإيصال_تاريخ": ...
|
| 67 |
+
}
|
| 68 |
+
If a field is missing, set it to null.
|
| 69 |
+
"""
|
| 70 |
+
|
| 71 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 72 |
+
|
| 73 |
+
headers = {
|
| 74 |
+
"Content-Type": "application/json"
|
| 75 |
+
}
|
| 76 |
+
|
| 77 |
+
data = {
|
| 78 |
+
"contents": [
|
| 79 |
+
{
|
| 80 |
+
"role": "user",
|
| 81 |
+
"parts": [
|
| 82 |
+
{"text": prompt},
|
| 83 |
+
{
|
| 84 |
+
"inline_data": {
|
| 85 |
+
"mime_type": "image/jpeg",
|
| 86 |
+
"data": image_b64
|
| 87 |
+
}
|
| 88 |
+
}
|
| 89 |
+
]
|
| 90 |
+
}
|
| 91 |
+
]
|
| 92 |
+
}
|
| 93 |
+
|
| 94 |
+
response = requests.post(url, headers=headers, json=data)
|
| 95 |
+
|
| 96 |
+
try:
|
| 97 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 98 |
+
|
| 99 |
+
# استخراج JSON من النص
|
| 100 |
+
match = re.search(r"```json\s*(\{.*?\})\s*```", response_text, re.DOTALL)
|
| 101 |
+
if match:
|
| 102 |
+
json_text = match.group(1)
|
| 103 |
+
result = json.loads(json_text)
|
| 104 |
+
|
| 105 |
+
base_name = os.path.splitext(os.path.basename(image_path))[0]
|
| 106 |
+
folder = os.path.dirname(image_path)
|
| 107 |
+
json_path = os.path.join(folder, f"{base_name}.json")
|
| 108 |
+
|
| 109 |
+
# إذا الملف موجود نقرأه ونضيف عليه
|
| 110 |
+
if os.path.exists(json_path):
|
| 111 |
+
with open(json_path, "r", encoding="utf-8") as f:
|
| 112 |
+
try:
|
| 113 |
+
existing_data = json.load(f)
|
| 114 |
+
if not isinstance(existing_data, list):
|
| 115 |
+
existing_data = [existing_data]
|
| 116 |
+
except json.JSONDecodeError:
|
| 117 |
+
existing_data = []
|
| 118 |
+
else:
|
| 119 |
+
existing_data = []
|
| 120 |
+
|
| 121 |
+
existing_data.append(result)
|
| 122 |
+
|
| 123 |
+
with open(json_path, "w", encoding="utf-8") as f:
|
| 124 |
+
json.dump(existing_data, f, ensure_ascii=False, indent=2)
|
| 125 |
+
|
| 126 |
+
print(f"✅ تم حفظ النتيجة في الملف: {json_path}")
|
| 127 |
+
print(json.dumps(result, indent=2, ensure_ascii=False))
|
| 128 |
+
else:
|
| 129 |
+
print("❌ لم أتمكن من استخراج JSON نظيف من الرد:")
|
| 130 |
+
print(response_text)
|
| 131 |
+
|
| 132 |
+
except Exception as e:
|
| 133 |
+
print(f"❌ حدث خطأ أثناء المعالجة: {e}")
|
| 134 |
+
print("الرد الكامل من Gemini:")
|
| 135 |
+
print(response.text)
|
json-B.py
ADDED
|
@@ -0,0 +1,96 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
|
| 7 |
+
# مفتاح Gemini API
|
| 8 |
+
API_KEY = "AIzaSyBr2-dUqHDZkk20hlWeEcpWnVVdkq9fqyE"
|
| 9 |
+
|
| 10 |
+
# المجلد الذي يحتوي صور cr1
|
| 11 |
+
cr1_images_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1"
|
| 12 |
+
# مجلد إخراج ملفات JSON
|
| 13 |
+
output_json_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1\cr1_json"
|
| 14 |
+
|
| 15 |
+
# Ensure output folder exists
|
| 16 |
+
os.makedirs(output_json_folder, exist_ok=True)
|
| 17 |
+
|
| 18 |
+
# Exact same prompt
|
| 19 |
+
prompt = """
|
| 20 |
+
Please extract the following fields in Arabic and English from the government registration document image:
|
| 21 |
+
- License Number
|
| 22 |
+
- Owner's Name
|
| 23 |
+
- ISIC Classification
|
| 24 |
+
- Detailed Activity
|
| 25 |
+
- Municipality
|
| 26 |
+
- Sub-Municipality
|
| 27 |
+
- District
|
| 28 |
+
- Street
|
| 29 |
+
- Shop's Total Area
|
| 30 |
+
|
| 31 |
+
Return the result in a JSON format with these keys:
|
| 32 |
+
en_license_number, en_owner_name, en_isic_classification, en_detailed_activity, en_municipality, en_sub_municipality, en_district, en_street, en_shop_total_area,
|
| 33 |
+
ar_license_number, ar_owner_name, ar_isic_classification, ar_detailed_activity, ar_municipality, ar_sub_municipality, ar_district, ar_street, ar_shop_total_area
|
| 34 |
+
|
| 35 |
+
If a field is missing, return null.
|
| 36 |
+
"""
|
| 37 |
+
|
| 38 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 39 |
+
headers = {"Content-Type": "application/json"}
|
| 40 |
+
|
| 41 |
+
# Iterate over all images
|
| 42 |
+
for image_name in os.listdir(cr1_images_folder):
|
| 43 |
+
if not image_name.lower().endswith(('.jpg', '.jpeg', '.png')):
|
| 44 |
+
continue
|
| 45 |
+
|
| 46 |
+
image_path = os.path.join(cr1_images_folder, image_name)
|
| 47 |
+
base_name = os.path.splitext(image_name)[0]
|
| 48 |
+
output_file = os.path.join(output_json_folder, base_name + ".json")
|
| 49 |
+
|
| 50 |
+
# Skip if JSON already exists
|
| 51 |
+
if os.path.exists(output_file):
|
| 52 |
+
print(f"Skipped {image_name} (JSON file already exists)")
|
| 53 |
+
continue
|
| 54 |
+
|
| 55 |
+
# Read image and convert to base64
|
| 56 |
+
with open(image_path, "rb") as f:
|
| 57 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 58 |
+
|
| 59 |
+
# Send request to Gemini API
|
| 60 |
+
data = {
|
| 61 |
+
"contents": [
|
| 62 |
+
{
|
| 63 |
+
"role": "user",
|
| 64 |
+
"parts": [
|
| 65 |
+
{"text": prompt},
|
| 66 |
+
{
|
| 67 |
+
"inline_data": {
|
| 68 |
+
"mime_type": "image/jpeg",
|
| 69 |
+
"data": image_b64
|
| 70 |
+
}
|
| 71 |
+
}
|
| 72 |
+
]
|
| 73 |
+
}
|
| 74 |
+
]
|
| 75 |
+
}
|
| 76 |
+
|
| 77 |
+
try:
|
| 78 |
+
response = requests.post(url, headers=headers, json=data)
|
| 79 |
+
response.raise_for_status()
|
| 80 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 81 |
+
|
| 82 |
+
match = re.search(r"```json\s*(\{.*\})\s*```", response_text, re.DOTALL)
|
| 83 |
+
if match:
|
| 84 |
+
json_text = match.group(1)
|
| 85 |
+
result = json.loads(json_text)
|
| 86 |
+
|
| 87 |
+
with open(output_file, "w", encoding="utf-8") as f:
|
| 88 |
+
json.dump(result, f, ensure_ascii=False, indent=2)
|
| 89 |
+
|
| 90 |
+
print(f"✅ Processed: {image_name}")
|
| 91 |
+
else:
|
| 92 |
+
print(f"❌ Failed to extract JSON from: {image_name}")
|
| 93 |
+
print(response_text)
|
| 94 |
+
|
| 95 |
+
except Exception as e:
|
| 96 |
+
print(f"❌ Error processing image {image_name}: {e}")
|
json-CR1.py
ADDED
|
@@ -0,0 +1,143 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
|
| 7 |
+
# مفتاح Gemini API
|
| 8 |
+
API_KEY = "AIzaSyBr2-dUqHDZkk20hlWeEcpWnVVdkq9fqyE"
|
| 9 |
+
|
| 10 |
+
# المجلد الذي يحتوي صور cr1
|
| 11 |
+
cr1_images_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1"
|
| 12 |
+
# مجلد إخراج ملفات JSON
|
| 13 |
+
output_json_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1\cr1_json"
|
| 14 |
+
|
| 15 |
+
# Ensure output folder exists
|
| 16 |
+
os.makedirs(output_json_folder, exist_ok=True)
|
| 17 |
+
|
| 18 |
+
# Exact same prompt
|
| 19 |
+
prompt = """
|
| 20 |
+
Please extract the following fields from the CR1 commercial registration document image, all in Arabic only:
|
| 21 |
+
|
| 22 |
+
- Unified number
|
| 23 |
+
- Establishment number
|
| 24 |
+
- Commercial name of the institution
|
| 25 |
+
- Its main center
|
| 26 |
+
- Phone
|
| 27 |
+
- Postal code
|
| 28 |
+
- Trader name
|
| 29 |
+
- Nationality
|
| 30 |
+
- Date of birth
|
| 31 |
+
- Civil registry number - residence
|
| 32 |
+
- Its date
|
| 33 |
+
- Its source
|
| 34 |
+
- Passport number
|
| 35 |
+
- Its date
|
| 36 |
+
- Issuer
|
| 37 |
+
- Activity
|
| 38 |
+
- Capital
|
| 39 |
+
- Name of the manager or authorized agent
|
| 40 |
+
- Nationality
|
| 41 |
+
- Date of birth
|
| 42 |
+
- Civil registry number - residence
|
| 43 |
+
- Its date
|
| 44 |
+
- Its source
|
| 45 |
+
- Certified by the commercial registration office in the city
|
| 46 |
+
- That the above-mentioned institution is registered in the city
|
| 47 |
+
- Certificates expire on
|
| 48 |
+
- Receipt number
|
| 49 |
+
- Date
|
| 50 |
+
|
| 51 |
+
Please return the result as JSON with these exact keys only, and if any field is missing, set its value to null:
|
| 52 |
+
|
| 53 |
+
{
|
| 54 |
+
"رقم الموحد": null,
|
| 55 |
+
"رقم المنشأة": null,
|
| 56 |
+
"الاسم التجاري للمؤسسة": null,
|
| 57 |
+
"مركزها الرئيسي": null,
|
| 58 |
+
"هاتف": null,
|
| 59 |
+
"الرمز البريدي": null,
|
| 60 |
+
"اسم التاجر": null,
|
| 61 |
+
"الجنسية": null,
|
| 62 |
+
"تاريخ الميلاد": null,
|
| 63 |
+
"رقم السجل المدني-الإقامة": null,
|
| 64 |
+
"تاريخه": null,
|
| 65 |
+
"مصدره": null,
|
| 66 |
+
"رقم الحفيظة-الجواز": null,
|
| 67 |
+
"تاريخه_2": null,
|
| 68 |
+
"مصدرة": null,
|
| 69 |
+
"النشاط": null,
|
| 70 |
+
"رأس المال": null,
|
| 71 |
+
"اسم المدير أو الوكيل المفوض": null,
|
| 72 |
+
"الجنسية_2": null,
|
| 73 |
+
"تاريخ الميلاد_2": null,
|
| 74 |
+
"رقم السجل المدني-الإقامة_2": null,
|
| 75 |
+
"تاريخه_3": null,
|
| 76 |
+
"مصدره_2": null,
|
| 77 |
+
"يشهد مكتب السجل التجاري بمدينة": null,
|
| 78 |
+
"بأنه تم تسجيل المؤسسة المذكورة أعلاه بمدينة": null,
|
| 79 |
+
"تنتهي صلاحية الشهادات في": null,
|
| 80 |
+
"بموجب الإيصال رقم": null,
|
| 81 |
+
"تاريخ_الإيصال": null
|
| 82 |
+
}
|
| 83 |
+
"""
|
| 84 |
+
|
| 85 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 86 |
+
headers = {"Content-Type": "application/json"}
|
| 87 |
+
|
| 88 |
+
# Iterate over all images
|
| 89 |
+
for image_name in os.listdir(cr1_images_folder):
|
| 90 |
+
if not image_name.lower().endswith(('.jpg', '.jpeg', '.png')):
|
| 91 |
+
continue
|
| 92 |
+
|
| 93 |
+
image_path = os.path.join(cr1_images_folder, image_name)
|
| 94 |
+
base_name = os.path.splitext(image_name)[0]
|
| 95 |
+
output_file = os.path.join(output_json_folder, base_name + ".json")
|
| 96 |
+
|
| 97 |
+
# Skip if JSON already exists
|
| 98 |
+
if os.path.exists(output_file):
|
| 99 |
+
print(f"Skipped {image_name} (JSON file already exists)")
|
| 100 |
+
continue
|
| 101 |
+
|
| 102 |
+
# Read image and convert to base64
|
| 103 |
+
with open(image_path, "rb") as f:
|
| 104 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 105 |
+
|
| 106 |
+
# Send request to Gemini API
|
| 107 |
+
data = {
|
| 108 |
+
"contents": [
|
| 109 |
+
{
|
| 110 |
+
"role": "user",
|
| 111 |
+
"parts": [
|
| 112 |
+
{"text": prompt},
|
| 113 |
+
{
|
| 114 |
+
"inline_data": {
|
| 115 |
+
"mime_type": "image/jpeg",
|
| 116 |
+
"data": image_b64
|
| 117 |
+
}
|
| 118 |
+
}
|
| 119 |
+
]
|
| 120 |
+
}
|
| 121 |
+
]
|
| 122 |
+
}
|
| 123 |
+
|
| 124 |
+
try:
|
| 125 |
+
response = requests.post(url, headers=headers, json=data)
|
| 126 |
+
response.raise_for_status()
|
| 127 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 128 |
+
|
| 129 |
+
match = re.search(r"```json\s*(\{.*\})\s*```", response_text, re.DOTALL)
|
| 130 |
+
if match:
|
| 131 |
+
json_text = match.group(1)
|
| 132 |
+
result = json.loads(json_text)
|
| 133 |
+
|
| 134 |
+
with open(output_file, "w", encoding="utf-8") as f:
|
| 135 |
+
json.dump(result, f, ensure_ascii=False, indent=2)
|
| 136 |
+
|
| 137 |
+
print(f"✅ Processed: {image_name}")
|
| 138 |
+
else:
|
| 139 |
+
print(f"❌ Failed to extract JSON from: {image_name}")
|
| 140 |
+
print(response_text)
|
| 141 |
+
|
| 142 |
+
except Exception as e:
|
| 143 |
+
print(f"❌ Error processing image {image_name}: {e}")
|
json-CR2.py
ADDED
|
@@ -0,0 +1,164 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
import time
|
| 7 |
+
|
| 8 |
+
API_KEY = "AIzaSyBr2-dUqHDZkk20hlWeEcpWnVVdkq9fqyE"
|
| 9 |
+
|
| 10 |
+
cr1_images_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR2"
|
| 11 |
+
output_json_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR2\cr2_json"
|
| 12 |
+
|
| 13 |
+
os.makedirs(output_json_folder, exist_ok=True)
|
| 14 |
+
|
| 15 |
+
prompt = """
|
| 16 |
+
يرجى استخراج الحقول التالية من صورة مستند CR2 (السجل التجاري 2)، جميعها بالعربية فقط:
|
| 17 |
+
|
| 18 |
+
- الرقم
|
| 19 |
+
- التاريخ
|
| 20 |
+
- الرقم الموحد للمنشأة
|
| 21 |
+
- الاسم التجاري للمؤسسة
|
| 22 |
+
- مركزها الرئيسي (بصيغة تفصيلية: المبنى، الشارع، الحي، الرقم الإضافي. مثال: "المبنى ٣٠١٧، شارع العباس بن الأحنف، حي الكندرة، الرقم الإضافي ٦٩٠٢")
|
| 23 |
+
- هاتف
|
| 24 |
+
- الرمز البريدي
|
| 25 |
+
- اسم التاجر
|
| 26 |
+
- الجنسية
|
| 27 |
+
- تاريخ الميلاد
|
| 28 |
+
- رقم السجل المدني-الإقامة
|
| 29 |
+
- تاريخه
|
| 30 |
+
- مصدره
|
| 31 |
+
- رقم الحفيظة-الجواز
|
| 32 |
+
- تاريخه
|
| 33 |
+
- مصدرة
|
| 34 |
+
- النشاط
|
| 35 |
+
- رأس المال
|
| 36 |
+
- اسم المدير او الوكيل المفوض
|
| 37 |
+
- الجنسية
|
| 38 |
+
- تاريخ الميلاد
|
| 39 |
+
- رقم السجل المدني-الإقامة
|
| 40 |
+
- تاريخه
|
| 41 |
+
- مصدره
|
| 42 |
+
- سلطات المدير
|
| 43 |
+
- يشهد مكتب السجل التجاري بمدينة
|
| 44 |
+
- بأنه تم تسجيل المؤسسة المذكورة أعلاه بمدينة
|
| 45 |
+
- وتنتهي صلاحية الشهادات في
|
| 46 |
+
- بموجب الإيصال رقم
|
| 47 |
+
- وتاريخ
|
| 48 |
+
- مدير السجل التجاري
|
| 49 |
+
|
| 50 |
+
أرجو إعادة النتيجة بصيغة JSON بهذه المفاتيح فقط، وإذا أي حقل غير موجود فضع قيمته null:
|
| 51 |
+
|
| 52 |
+
{
|
| 53 |
+
"الرقم": null,
|
| 54 |
+
"التاريخ": null,
|
| 55 |
+
"الرقم الموحد للمنشأة": null,
|
| 56 |
+
"الاسم التجاري للمؤسسة": null,
|
| 57 |
+
"مركزها الرئيسي": null,
|
| 58 |
+
"هاتف": null,
|
| 59 |
+
"الرمز البريدي": null,
|
| 60 |
+
"اسم التاجر": null,
|
| 61 |
+
"الجنسية": null,
|
| 62 |
+
"تاريخ الميلاد": null,
|
| 63 |
+
"رقم السجل المدني-الإقامة": null,
|
| 64 |
+
"تاريخه": null,
|
| 65 |
+
"مصدره": null,
|
| 66 |
+
"رقم الحفيظة-الجواز": null,
|
| 67 |
+
"تاريخه_2": null,
|
| 68 |
+
"مصدرة": null,
|
| 69 |
+
"النشاط": null,
|
| 70 |
+
"رأس المال": null,
|
| 71 |
+
"اسم المدير او الوكيل المفوض": null,
|
| 72 |
+
"الجنسية_2": null,
|
| 73 |
+
"تاريخ الميلاد_2": null,
|
| 74 |
+
"رقم السجل المدني-الإقامة_2": null,
|
| 75 |
+
"تاريخه_3": null,
|
| 76 |
+
"مصدره_2": null,
|
| 77 |
+
"سلطات المدير": null,
|
| 78 |
+
"يشهد مكتب السجل التجاري بمدينة": null,
|
| 79 |
+
"بأنه تم تسجيل المؤسسة المذكورة أعلاه بمدينة": null,
|
| 80 |
+
"تنتهي صلاحية الشهادات في": null,
|
| 81 |
+
"بموجب الإيصال رقم": null,
|
| 82 |
+
"تاريخ_الإيصال": null,
|
| 83 |
+
"مدير السجل التجاري": null
|
| 84 |
+
}
|
| 85 |
+
"""
|
| 86 |
+
|
| 87 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 88 |
+
headers = {"Content-Type": "application/json"}
|
| 89 |
+
|
| 90 |
+
def split_address(address):
|
| 91 |
+
parts = [p.strip() for p in address.split("،")]
|
| 92 |
+
while len(parts) < 4:
|
| 93 |
+
parts.append(None)
|
| 94 |
+
return parts
|
| 95 |
+
|
| 96 |
+
for image_name in os.listdir(cr1_images_folder):
|
| 97 |
+
if not image_name.lower().endswith(('.jpg', '.jpeg', '.png')):
|
| 98 |
+
continue
|
| 99 |
+
|
| 100 |
+
image_path = os.path.join(cr1_images_folder, image_name)
|
| 101 |
+
base_name = os.path.splitext(image_name)[0]
|
| 102 |
+
output_file = os.path.join(output_json_folder, base_name + ".json")
|
| 103 |
+
|
| 104 |
+
if os.path.exists(output_file):
|
| 105 |
+
print(f"Skipped {image_name} (JSON file already exists)")
|
| 106 |
+
continue
|
| 107 |
+
|
| 108 |
+
with open(image_path, "rb") as f:
|
| 109 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 110 |
+
|
| 111 |
+
data = {
|
| 112 |
+
"contents": [
|
| 113 |
+
{
|
| 114 |
+
"role": "user",
|
| 115 |
+
"parts": [
|
| 116 |
+
{"text": prompt},
|
| 117 |
+
{
|
| 118 |
+
"inline_data": {
|
| 119 |
+
"mime_type": "image/jpeg",
|
| 120 |
+
"data": image_b64
|
| 121 |
+
}
|
| 122 |
+
}
|
| 123 |
+
]
|
| 124 |
+
}
|
| 125 |
+
]
|
| 126 |
+
}
|
| 127 |
+
|
| 128 |
+
try:
|
| 129 |
+
response = requests.post(url, headers=headers, json=data)
|
| 130 |
+
response.raise_for_status()
|
| 131 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 132 |
+
|
| 133 |
+
# استخدام التعبير المنتظم الصحيح
|
| 134 |
+
match = re.search(r"```json\s*(\{.*?\})\s*```", response_text, re.DOTALL)
|
| 135 |
+
if match:
|
| 136 |
+
json_text = match.group(1)
|
| 137 |
+
result = json.loads(json_text)
|
| 138 |
+
|
| 139 |
+
# تقسيم العنوان إلى أجزاء منفصلة
|
| 140 |
+
if "مركزها الرئيسي" in result and result["مركزها الرئيسي"]:
|
| 141 |
+
parts = split_address(result["مركزها الرئيسي"])
|
| 142 |
+
result["رقم المبنى"] = parts[0]
|
| 143 |
+
result["اسم الشارع"] = parts[1]
|
| 144 |
+
result["اسم الحي"] = parts[2]
|
| 145 |
+
result["الرقم الإضافي"] = parts[3]
|
| 146 |
+
else:
|
| 147 |
+
result["رقم المبنى"] = None
|
| 148 |
+
result["اسم الشارع"] = None
|
| 149 |
+
result["اسم الحي"] = None
|
| 150 |
+
result["الرقم الإضافي"] = None
|
| 151 |
+
|
| 152 |
+
with open(output_file, "w", encoding="utf-8") as f:
|
| 153 |
+
json.dump(result, f, ensure_ascii=False, indent=2)
|
| 154 |
+
|
| 155 |
+
print(f"✅ Processed: {image_name}")
|
| 156 |
+
|
| 157 |
+
else:
|
| 158 |
+
print(f"❌ Failed to extract JSON from: {image_name}")
|
| 159 |
+
print(response_text)
|
| 160 |
+
|
| 161 |
+
time.sleep(3) # ينتظر 3 ثواني قبل إرسال الصورة التالية
|
| 162 |
+
|
| 163 |
+
except Exception as e:
|
| 164 |
+
print(f"❌ Error processing image {image_name}: {e}")
|
json-CR3.py
ADDED
|
@@ -0,0 +1,165 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
import time
|
| 7 |
+
|
| 8 |
+
API_KEY = "AIzaSyBr2-dUqHDZkk20hlWeEcpWnVVdkq9fqyE"
|
| 9 |
+
|
| 10 |
+
cr1_images_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR2"
|
| 11 |
+
output_json_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR2\cr2_json"
|
| 12 |
+
|
| 13 |
+
os.makedirs(output_json_folder, exist_ok=True)
|
| 14 |
+
|
| 15 |
+
prompt = """
|
| 16 |
+
Extract the following fields from the CR3 document image. Return both Arabic and English text where available:
|
| 17 |
+
|
| 18 |
+
الرقم
|
| 19 |
+
التاريخ
|
| 20 |
+
الرقم الموحد للمنشأة
|
| 21 |
+
اسم التاجر
|
| 22 |
+
الجنسية
|
| 23 |
+
تاريخ الميلاد
|
| 24 |
+
رقم السجل المدني-الإقامة
|
| 25 |
+
تاريخه
|
| 26 |
+
مصدرة
|
| 27 |
+
مركزها الرئيسي
|
| 28 |
+
هاتف
|
| 29 |
+
الرمز البريدي
|
| 30 |
+
رقم سجل المركز الرئيسي
|
| 31 |
+
الاسم التجاري للفرع
|
| 32 |
+
العنوان
|
| 33 |
+
الرمز البريدي
|
| 34 |
+
الهاتف
|
| 35 |
+
النشاط
|
| 36 |
+
رأس المال
|
| 37 |
+
إسم المدير او الوكيل المفوض
|
| 38 |
+
الجنسية
|
| 39 |
+
تاريخ الميلاد
|
| 40 |
+
رقم السجل المدني-الإقامة
|
| 41 |
+
تاريخه
|
| 42 |
+
مصدره
|
| 43 |
+
سلطات المدير
|
| 44 |
+
يشهد مكتب السجل التجاري بمدينة
|
| 45 |
+
بأنه تم تسجيل المؤسسة المذكورة أعلاة بمدينة
|
| 46 |
+
وتنتهي صلاحية الشهادات في
|
| 47 |
+
بموجب الإيصال رقم
|
| 48 |
+
وتاريخ
|
| 49 |
+
مدير السجل التجاري
|
| 50 |
+
|
| 51 |
+
Return as JSON with keys:
|
| 52 |
+
{
|
| 53 |
+
"الرقم": ...,
|
| 54 |
+
"التاريخ": ...,
|
| 55 |
+
"الرقم_الموحد_للمنشأة": ...,
|
| 56 |
+
"اسم_التاجر": ...,
|
| 57 |
+
"الجنسية": ...,
|
| 58 |
+
"تاريخ_الميلاد": ...,
|
| 59 |
+
"رقم_السجل_المدني_الإقامة": ...,
|
| 60 |
+
"تاريخه": ...,
|
| 61 |
+
"مصدرة": ...,
|
| 62 |
+
"مركزها_الرئيسي": ...,
|
| 63 |
+
"هاتف": ...,
|
| 64 |
+
"الرمز_البريدي": ...,
|
| 65 |
+
"رقم_سجل_المركز_الرئيسي": ...,
|
| 66 |
+
"الاسم_التجاري_للفرع": ...,
|
| 67 |
+
"العنوان": ...,
|
| 68 |
+
"الرمز_البريدي_الفرع": ...,
|
| 69 |
+
"الهاتف_الفرع": ...,
|
| 70 |
+
"النشاط": ...,
|
| 71 |
+
"رأس_المال": ...,
|
| 72 |
+
"اسم_المدير_او_الوكيل_المفوض": ...,
|
| 73 |
+
"الجنسية_المدير": ...,
|
| 74 |
+
"تاريخ_ميلاد_المدير": ...,
|
| 75 |
+
"رقم_السجل_المدني_الإقامة_المدير": ...,
|
| 76 |
+
"تاريخه_المدير": ...,
|
| 77 |
+
"مصدره_المدير": ...,
|
| 78 |
+
"سلطات_المدير": ...,
|
| 79 |
+
"يشهد_مكتب_السجل": ...,
|
| 80 |
+
"تم_تسجيل_المؤسسة": ...,
|
| 81 |
+
"تنتهي_صلاحية_الشهادة": ...,
|
| 82 |
+
"الإيصال_رقم": ...,
|
| 83 |
+
"الإيصال_تاريخ": ...,
|
| 84 |
+
"مدير_السجل_التجاري": ...
|
| 85 |
+
}
|
| 86 |
+
If a field is missing, set it to null.
|
| 87 |
+
"""
|
| 88 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 89 |
+
headers = {"Content-Type": "application/json"}
|
| 90 |
+
|
| 91 |
+
def split_address(address):
|
| 92 |
+
parts = [p.strip() for p in address.split("،")]
|
| 93 |
+
while len(parts) < 4:
|
| 94 |
+
parts.append(None)
|
| 95 |
+
return parts
|
| 96 |
+
|
| 97 |
+
for image_name in os.listdir(cr1_images_folder):
|
| 98 |
+
if not image_name.lower().endswith(('.jpg', '.jpeg', '.png')):
|
| 99 |
+
continue
|
| 100 |
+
|
| 101 |
+
image_path = os.path.join(cr1_images_folder, image_name)
|
| 102 |
+
base_name = os.path.splitext(image_name)[0]
|
| 103 |
+
output_file = os.path.join(output_json_folder, base_name + ".json")
|
| 104 |
+
|
| 105 |
+
if os.path.exists(output_file):
|
| 106 |
+
print(f"Skipped {image_name} (JSON file already exists)")
|
| 107 |
+
continue
|
| 108 |
+
|
| 109 |
+
with open(image_path, "rb") as f:
|
| 110 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 111 |
+
|
| 112 |
+
data = {
|
| 113 |
+
"contents": [
|
| 114 |
+
{
|
| 115 |
+
"role": "user",
|
| 116 |
+
"parts": [
|
| 117 |
+
{"text": prompt},
|
| 118 |
+
{
|
| 119 |
+
"inline_data": {
|
| 120 |
+
"mime_type": "image/jpeg",
|
| 121 |
+
"data": image_b64
|
| 122 |
+
}
|
| 123 |
+
}
|
| 124 |
+
]
|
| 125 |
+
}
|
| 126 |
+
]
|
| 127 |
+
}
|
| 128 |
+
|
| 129 |
+
try:
|
| 130 |
+
response = requests.post(url, headers=headers, json=data)
|
| 131 |
+
response.raise_for_status()
|
| 132 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 133 |
+
|
| 134 |
+
# استخدام التعبير المنتظم الصحيح
|
| 135 |
+
match = re.search(r"```json\s*(\{.*?\})\s*```", response_text, re.DOTALL)
|
| 136 |
+
if match:
|
| 137 |
+
json_text = match.group(1)
|
| 138 |
+
result = json.loads(json_text)
|
| 139 |
+
|
| 140 |
+
# تقسيم العنوان إلى أجزاء منفصلة
|
| 141 |
+
if "مركزها الرئيسي" in result and result["مركزها الرئيسي"]:
|
| 142 |
+
parts = split_address(result["مركزها الرئيسي"])
|
| 143 |
+
result["رقم المبنى"] = parts[0]
|
| 144 |
+
result["اسم الشارع"] = parts[1]
|
| 145 |
+
result["اسم الحي"] = parts[2]
|
| 146 |
+
result["الرقم الإضافي"] = parts[3]
|
| 147 |
+
else:
|
| 148 |
+
result["رقم المبنى"] = None
|
| 149 |
+
result["اسم الشارع"] = None
|
| 150 |
+
result["اسم الحي"] = None
|
| 151 |
+
result["الرقم الإضافي"] = None
|
| 152 |
+
|
| 153 |
+
with open(output_file, "w", encoding="utf-8") as f:
|
| 154 |
+
json.dump(result, f, ensure_ascii=False, indent=2)
|
| 155 |
+
|
| 156 |
+
print(f"✅ Processed: {image_name}")
|
| 157 |
+
|
| 158 |
+
else:
|
| 159 |
+
print(f"❌ Failed to extract JSON from: {image_name}")
|
| 160 |
+
print(response_text)
|
| 161 |
+
|
| 162 |
+
time.sleep(3) # ينتظر 3 ثواني قبل إرسال الصورة التالية
|
| 163 |
+
|
| 164 |
+
except Exception as e:
|
| 165 |
+
print(f"❌ Error processing image {image_name}: {e}")
|
json-CR4.py
ADDED
|
@@ -0,0 +1,145 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
import time
|
| 7 |
+
|
| 8 |
+
API_KEY = "AIzaSyBr2-dUqHDZkk20hlWeEcpWnVVdkq9fqyE"
|
| 9 |
+
|
| 10 |
+
cr1_images_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR2"
|
| 11 |
+
output_json_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR2\cr2_json"
|
| 12 |
+
|
| 13 |
+
os.makedirs(output_json_folder, exist_ok=True)
|
| 14 |
+
|
| 15 |
+
prompt = """
|
| 16 |
+
Extract the following fields from the CR4 document image. Return both Arabic and English text where available:
|
| 17 |
+
|
| 18 |
+
الرقم الموحد
|
| 19 |
+
رقم المنشأة
|
| 20 |
+
التاريخ
|
| 21 |
+
|
| 22 |
+
الاسم التجاري للشركة
|
| 23 |
+
نوعها
|
| 24 |
+
جنسيتها
|
| 25 |
+
مدة الشركة
|
| 26 |
+
تبدأ من
|
| 27 |
+
وتنتهي في
|
| 28 |
+
مركزها الرئيسي
|
| 29 |
+
هاتف
|
| 30 |
+
الرمز البريدي
|
| 31 |
+
|
| 32 |
+
النشاط
|
| 33 |
+
رأس المال
|
| 34 |
+
المديرون
|
| 35 |
+
سلطات المدير/المديرون
|
| 36 |
+
يشهد مكتب السجل التجاري بمدينة
|
| 37 |
+
بأنه تم تسجيل المؤسسة المذكورة أعلاة بمدينة
|
| 38 |
+
وتنتهي صلاحية الشهادات في
|
| 39 |
+
بموجب الإيصال رقم
|
| 40 |
+
وتاريخ
|
| 41 |
+
|
| 42 |
+
Return as JSON with keys:
|
| 43 |
+
{
|
| 44 |
+
"رقم_موحد": ...,
|
| 45 |
+
"رقم_المنشأة": ...,
|
| 46 |
+
"التاريخ": ...,
|
| 47 |
+
"الاسم_التجاري": ...,
|
| 48 |
+
"نوعها": ...,
|
| 49 |
+
"جنسيتها": ...,
|
| 50 |
+
"مدة_الشركة": ...,
|
| 51 |
+
"تبدأ_من": ...,
|
| 52 |
+
"تنتهي_في": ...,
|
| 53 |
+
"مركزها_الرئيسي": ...,
|
| 54 |
+
"هاتف": ...,
|
| 55 |
+
"الرمز_البريدي": ...,
|
| 56 |
+
"النشاط": ...,
|
| 57 |
+
"رأس_المال": ...,
|
| 58 |
+
"المديرون": ...,
|
| 59 |
+
"سلطات_المدير": ...,
|
| 60 |
+
"يشهد_مكتب_السجل": ...,
|
| 61 |
+
"تم_تسجيل_المؤسسة": ...,
|
| 62 |
+
"تنتهي_صلاحية_الشهادة": ...,
|
| 63 |
+
"الإيصال_رقم": ...,
|
| 64 |
+
"الإيصال_تاريخ": ...
|
| 65 |
+
}
|
| 66 |
+
If a field is missing, set it to null.
|
| 67 |
+
"""
|
| 68 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 69 |
+
headers = {"Content-Type": "application/json"}
|
| 70 |
+
|
| 71 |
+
def split_address(address):
|
| 72 |
+
parts = [p.strip() for p in address.split("،")]
|
| 73 |
+
while len(parts) < 4:
|
| 74 |
+
parts.append(None)
|
| 75 |
+
return parts
|
| 76 |
+
|
| 77 |
+
for image_name in os.listdir(cr1_images_folder):
|
| 78 |
+
if not image_name.lower().endswith(('.jpg', '.jpeg', '.png')):
|
| 79 |
+
continue
|
| 80 |
+
|
| 81 |
+
image_path = os.path.join(cr1_images_folder, image_name)
|
| 82 |
+
base_name = os.path.splitext(image_name)[0]
|
| 83 |
+
output_file = os.path.join(output_json_folder, base_name + ".json")
|
| 84 |
+
|
| 85 |
+
if os.path.exists(output_file):
|
| 86 |
+
print(f"Skipped {image_name} (JSON file already exists)")
|
| 87 |
+
continue
|
| 88 |
+
|
| 89 |
+
with open(image_path, "rb") as f:
|
| 90 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 91 |
+
|
| 92 |
+
data = {
|
| 93 |
+
"contents": [
|
| 94 |
+
{
|
| 95 |
+
"role": "user",
|
| 96 |
+
"parts": [
|
| 97 |
+
{"text": prompt},
|
| 98 |
+
{
|
| 99 |
+
"inline_data": {
|
| 100 |
+
"mime_type": "image/jpeg",
|
| 101 |
+
"data": image_b64
|
| 102 |
+
}
|
| 103 |
+
}
|
| 104 |
+
]
|
| 105 |
+
}
|
| 106 |
+
]
|
| 107 |
+
}
|
| 108 |
+
|
| 109 |
+
try:
|
| 110 |
+
response = requests.post(url, headers=headers, json=data)
|
| 111 |
+
response.raise_for_status()
|
| 112 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 113 |
+
|
| 114 |
+
# استخدام التعبير المنتظم الصحيح
|
| 115 |
+
match = re.search(r"```json\s*(\{.*?\})\s*```", response_text, re.DOTALL)
|
| 116 |
+
if match:
|
| 117 |
+
json_text = match.group(1)
|
| 118 |
+
result = json.loads(json_text)
|
| 119 |
+
|
| 120 |
+
# تقسيم العنوان إلى أجزاء منفصلة
|
| 121 |
+
if "مركزها الرئيسي" in result and result["مركزها الرئيسي"]:
|
| 122 |
+
parts = split_address(result["مركزها الرئيسي"])
|
| 123 |
+
result["رقم المبنى"] = parts[0]
|
| 124 |
+
result["اسم الشارع"] = parts[1]
|
| 125 |
+
result["اسم الحي"] = parts[2]
|
| 126 |
+
result["الرقم الإضافي"] = parts[3]
|
| 127 |
+
else:
|
| 128 |
+
result["رقم المبنى"] = None
|
| 129 |
+
result["اسم الشارع"] = None
|
| 130 |
+
result["اسم الحي"] = None
|
| 131 |
+
result["الرقم الإضافي"] = None
|
| 132 |
+
|
| 133 |
+
with open(output_file, "w", encoding="utf-8") as f:
|
| 134 |
+
json.dump(result, f, ensure_ascii=False, indent=2)
|
| 135 |
+
|
| 136 |
+
print(f"✅ Processed: {image_name}")
|
| 137 |
+
|
| 138 |
+
else:
|
| 139 |
+
print(f"❌ Failed to extract JSON from: {image_name}")
|
| 140 |
+
print(response_text)
|
| 141 |
+
|
| 142 |
+
time.sleep(3) # ينتظر 3 ثواني قبل إرسال الصورة التالية
|
| 143 |
+
|
| 144 |
+
except Exception as e:
|
| 145 |
+
print(f"❌ Error processing image {image_name}: {e}")
|
json-CR5.py
ADDED
|
File without changes
|
json-CR6.py
ADDED
|
@@ -0,0 +1,138 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
|
| 7 |
+
# مفتاح Gemini API
|
| 8 |
+
API_KEY = "AIzaSyBr2-dUqHDZkk20hlWeEcpWnVVdkq9fqyE"
|
| 9 |
+
|
| 10 |
+
# المجلد الذي يحتوي صور cr1
|
| 11 |
+
cr1_images_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR6"
|
| 12 |
+
|
| 13 |
+
# مجلد إخراج ملفات JSON
|
| 14 |
+
output_json_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR6\cr6_json"
|
| 15 |
+
|
| 16 |
+
# Ensure output folder exists
|
| 17 |
+
os.makedirs(output_json_folder, exist_ok=True)
|
| 18 |
+
|
| 19 |
+
# Exact same prompt
|
| 20 |
+
prompt = """
|
| 21 |
+
يرجى استخراج الحقول التالية من مستند السجل التجاري (CR6) بالصورة، باللغة العربية فقط:
|
| 22 |
+
|
| 23 |
+
- الكيان التجاري
|
| 24 |
+
- حالة السجل
|
| 25 |
+
- مدة المنشأة
|
| 26 |
+
- الرقم الوطني الموحد للمنشأة
|
| 27 |
+
- رابط المتجر الإكتروني
|
| 28 |
+
- رأس المال
|
| 29 |
+
- المدينة
|
| 30 |
+
- صندوق البريد
|
| 31 |
+
- الرمز البريدي
|
| 32 |
+
- هاتف
|
| 33 |
+
- تاريخ اصدار السجل
|
| 34 |
+
- تاريخ انتهاء السجل
|
| 35 |
+
- الموقع الاكتروني
|
| 36 |
+
- العنوان
|
| 37 |
+
- النشاط التجاري
|
| 38 |
+
|
| 39 |
+
أرجو إعادة النتيجة بصيغة JSON بهذه المفاتيح فقط، وإذا أي حقل غير موجود فضع قيمته null:
|
| 40 |
+
|
| 41 |
+
{
|
| 42 |
+
"الكيان التجاري": null,
|
| 43 |
+
"حالة السجل": null,
|
| 44 |
+
"مدة المنشأة": null,
|
| 45 |
+
"الرقم الوطني الموحد للمنشأة": null,
|
| 46 |
+
"رابط المتجر الإكتروني": null,
|
| 47 |
+
"رأس المال": null,
|
| 48 |
+
"المدينة": null,
|
| 49 |
+
"صندوق البريد": null,
|
| 50 |
+
"الرمز البريدي": null,
|
| 51 |
+
"هاتف": null,
|
| 52 |
+
"تاريخ اصدار السجل": null,
|
| 53 |
+
"تاريخ انتهاء السجل": null,
|
| 54 |
+
"الموقع الاكتروني": null,
|
| 55 |
+
"العنوان": null,
|
| 56 |
+
"النشاط التجاري": null,
|
| 57 |
+
"رقم المبنى": null,
|
| 58 |
+
"الرقم الإضافي": null,
|
| 59 |
+
"رقم الوحدة": null
|
| 60 |
+
}
|
| 61 |
+
"""
|
| 62 |
+
|
| 63 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 64 |
+
headers = {"Content-Type": "application/json"}
|
| 65 |
+
|
| 66 |
+
# Iterate over all images
|
| 67 |
+
for image_name in os.listdir(cr1_images_folder):
|
| 68 |
+
if not image_name.lower().endswith(('.jpg', '.jpeg', '.png')):
|
| 69 |
+
continue
|
| 70 |
+
|
| 71 |
+
image_path = os.path.join(cr1_images_folder, image_name)
|
| 72 |
+
base_name = os.path.splitext(image_name)[0]
|
| 73 |
+
output_file = os.path.join(output_json_folder, base_name + ".json")
|
| 74 |
+
|
| 75 |
+
# Skip if JSON already exists
|
| 76 |
+
if os.path.exists(output_file):
|
| 77 |
+
print(f"⏩ Skipped {image_name} (JSON file already exists)")
|
| 78 |
+
continue
|
| 79 |
+
|
| 80 |
+
# Read image and convert to base64
|
| 81 |
+
with open(image_path, "rb") as f:
|
| 82 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 83 |
+
|
| 84 |
+
# Send request to Gemini API
|
| 85 |
+
data = {
|
| 86 |
+
"contents": [
|
| 87 |
+
{
|
| 88 |
+
"role": "user",
|
| 89 |
+
"parts": [
|
| 90 |
+
{"text": prompt},
|
| 91 |
+
{
|
| 92 |
+
"inline_data": {
|
| 93 |
+
"mime_type": "image/jpeg",
|
| 94 |
+
"data": image_b64
|
| 95 |
+
}
|
| 96 |
+
}
|
| 97 |
+
]
|
| 98 |
+
}
|
| 99 |
+
]
|
| 100 |
+
}
|
| 101 |
+
|
| 102 |
+
try:
|
| 103 |
+
response = requests.post(url, headers=headers, json=data)
|
| 104 |
+
response.raise_for_status()
|
| 105 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 106 |
+
|
| 107 |
+
match = re.search(r"```json\s*(\{.*\})\s*```", response_text, re.DOTALL)
|
| 108 |
+
if match:
|
| 109 |
+
json_text = match.group(1)
|
| 110 |
+
result = json.loads(json_text)
|
| 111 |
+
|
| 112 |
+
# تقسيم حقل العنوان إذا كان موجود
|
| 113 |
+
address = result.get("العنوان")
|
| 114 |
+
if address:
|
| 115 |
+
parts = address.strip().split()
|
| 116 |
+
if len(parts) == 5:
|
| 117 |
+
result["رقم المبنى"] = parts[0]
|
| 118 |
+
result["المدينة"] = parts[1]
|
| 119 |
+
result["الرمز البريدي"] = parts[2]
|
| 120 |
+
result["الرقم الإضافي"] = parts[3]
|
| 121 |
+
result["رقم الوحدة"] = parts[4]
|
| 122 |
+
else:
|
| 123 |
+
result["رقم المبنى"] = None
|
| 124 |
+
result["المدينة"] = None
|
| 125 |
+
result["الرمز البريدي"] = None
|
| 126 |
+
result["الرقم الإضافي"] = None
|
| 127 |
+
result["رقم الوحدة"] = None
|
| 128 |
+
|
| 129 |
+
with open(output_file, "w", encoding="utf-8") as f:
|
| 130 |
+
json.dump(result, f, ensure_ascii=False, indent=2)
|
| 131 |
+
|
| 132 |
+
print(f"✅ Processed: {image_name}")
|
| 133 |
+
else:
|
| 134 |
+
print(f"❌ Failed to extract JSON from: {image_name}")
|
| 135 |
+
print(response_text)
|
| 136 |
+
|
| 137 |
+
except Exception as e:
|
| 138 |
+
print(f"❌ Error processing image {image_name}: {e}")
|
json-CR7.py
ADDED
|
@@ -0,0 +1,112 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
|
| 7 |
+
# مفتاح Gemini API
|
| 8 |
+
API_KEY = "AIzaSyBr2-dUqHDZkk20hlWeEcpWnVVdkq9fqyE"
|
| 9 |
+
|
| 10 |
+
# المجلد الذي يحتوي صور cr1
|
| 11 |
+
cr1_images_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1"
|
| 12 |
+
# مجلد إخراج ملفات JSON
|
| 13 |
+
output_json_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1\cr1_json"
|
| 14 |
+
|
| 15 |
+
# Ensure output folder exists
|
| 16 |
+
os.makedirs(output_json_folder, exist_ok=True)
|
| 17 |
+
|
| 18 |
+
# Exact same prompt
|
| 19 |
+
prompt = """
|
| 20 |
+
يرجى استخراج الحقول التالية من مستند السجل التجاري (CR7) بالصورة، باللغة العربية فقط:
|
| 21 |
+
|
| 22 |
+
- اسم المنشأة
|
| 23 |
+
- نوع السجل
|
| 24 |
+
- حالة السجل
|
| 25 |
+
- الرقم الموحد للمنشأة
|
| 26 |
+
- رقم السجل التجاري
|
| 27 |
+
- اسم المالك
|
| 28 |
+
- نوع الكيان
|
| 29 |
+
- تاريخ الاصدار
|
| 30 |
+
- تاريخ الانتهاء
|
| 31 |
+
- قائمة المدراء
|
| 32 |
+
- المدينة
|
| 33 |
+
- الموقع الإلكتروني
|
| 34 |
+
- الانشطة التجارية
|
| 35 |
+
|
| 36 |
+
أرجو إعادة النتيجة بصيغة JSON بهذه المفاتيح فقط، وإذا أي حقل غير موجود فضع قيمته null:
|
| 37 |
+
|
| 38 |
+
{
|
| 39 |
+
"اسم المنشأة": null,
|
| 40 |
+
"نوع السجل": null,
|
| 41 |
+
"حالة السجل": null,
|
| 42 |
+
"الرقم الموحد للمنشأة": null,
|
| 43 |
+
"رقم السجل التجاري": null,
|
| 44 |
+
"اسم المالك": null,
|
| 45 |
+
"نوع الكيان": null,
|
| 46 |
+
"تاريخ الاصدار": null,
|
| 47 |
+
"تاريخ الانتهاء": null,
|
| 48 |
+
"قائمة المدراء": null,
|
| 49 |
+
"المدينة": null,
|
| 50 |
+
"الموقع الإلكتروني": null,
|
| 51 |
+
"الانشطة التجارية": null
|
| 52 |
+
}
|
| 53 |
+
"""
|
| 54 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 55 |
+
headers = {"Content-Type": "application/json"}
|
| 56 |
+
|
| 57 |
+
# Iterate over all images
|
| 58 |
+
for image_name in os.listdir(cr1_images_folder):
|
| 59 |
+
if not image_name.lower().endswith(('.jpg', '.jpeg', '.png')):
|
| 60 |
+
continue
|
| 61 |
+
|
| 62 |
+
image_path = os.path.join(cr1_images_folder, image_name)
|
| 63 |
+
base_name = os.path.splitext(image_name)[0]
|
| 64 |
+
output_file = os.path.join(output_json_folder, base_name + ".json")
|
| 65 |
+
|
| 66 |
+
# Skip if JSON already exists
|
| 67 |
+
if os.path.exists(output_file):
|
| 68 |
+
print(f"Skipped {image_name} (JSON file already exists)")
|
| 69 |
+
continue
|
| 70 |
+
|
| 71 |
+
# Read image and convert to base64
|
| 72 |
+
with open(image_path, "rb") as f:
|
| 73 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 74 |
+
|
| 75 |
+
# Send request to Gemini API
|
| 76 |
+
data = {
|
| 77 |
+
"contents": [
|
| 78 |
+
{
|
| 79 |
+
"role": "user",
|
| 80 |
+
"parts": [
|
| 81 |
+
{"text": prompt},
|
| 82 |
+
{
|
| 83 |
+
"inline_data": {
|
| 84 |
+
"mime_type": "image/jpeg",
|
| 85 |
+
"data": image_b64
|
| 86 |
+
}
|
| 87 |
+
}
|
| 88 |
+
]
|
| 89 |
+
}
|
| 90 |
+
]
|
| 91 |
+
}
|
| 92 |
+
|
| 93 |
+
try:
|
| 94 |
+
response = requests.post(url, headers=headers, json=data)
|
| 95 |
+
response.raise_for_status()
|
| 96 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 97 |
+
|
| 98 |
+
match = re.search(r"```json\s*(\{.*\})\s*```", response_text, re.DOTALL)
|
| 99 |
+
if match:
|
| 100 |
+
json_text = match.group(1)
|
| 101 |
+
result = json.loads(json_text)
|
| 102 |
+
|
| 103 |
+
with open(output_file, "w", encoding="utf-8") as f:
|
| 104 |
+
json.dump(result, f, ensure_ascii=False, indent=2)
|
| 105 |
+
|
| 106 |
+
print(f"✅ Processed: {image_name}")
|
| 107 |
+
else:
|
| 108 |
+
print(f"❌ Failed to extract JSON from: {image_name}")
|
| 109 |
+
print(response_text)
|
| 110 |
+
|
| 111 |
+
except Exception as e:
|
| 112 |
+
print(f"❌ Error processing image {image_name}: {e}")
|
json-V2.py
ADDED
|
@@ -0,0 +1,97 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
|
| 7 |
+
# مفتاح Gemini API
|
| 8 |
+
API_KEY = "AIzaSyBr2-dUqHDZkk20hlWeEcpWnVVdkq9fqyE"
|
| 9 |
+
|
| 10 |
+
# المجلد الذي يحتوي صور cr1
|
| 11 |
+
cr1_images_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1"
|
| 12 |
+
# مجلد إخراج ملفات JSON
|
| 13 |
+
output_json_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1\cr1_json"
|
| 14 |
+
|
| 15 |
+
# Ensure output folder exists
|
| 16 |
+
os.makedirs(output_json_folder, exist_ok=True)
|
| 17 |
+
|
| 18 |
+
# Exact same prompt
|
| 19 |
+
prompt = """
|
| 20 |
+
Please extract the following fields in Arabic and English from the tax document image:
|
| 21 |
+
Taxpayer Name
|
| 22 |
+
VAT Registration Number
|
| 23 |
+
Effective Registration Date
|
| 24 |
+
Taxpayer Address
|
| 25 |
+
CR/License
|
| 26 |
+
Contract/ID No
|
| 27 |
+
Tax period
|
| 28 |
+
First Filing due date
|
| 29 |
+
|
| 30 |
+
Return the result in a JSON format with these keys:
|
| 31 |
+
en_taxpayer_name, en_vat_registration_number, en_effective_registration_date, en_taxpayer_address,
|
| 32 |
+
en_cr_license, en_contract_id_no, en_tax_period, en_first_filing_due_date,
|
| 33 |
+
ar_taxpayer_name, ar_vat_registration_number, ar_effective_registration_date, ar_taxpayer_address,
|
| 34 |
+
ar_cr_license, ar_contract_id_no, ar_tax_period, ar_first_filing_due_date
|
| 35 |
+
|
| 36 |
+
If a field is missing, return null.
|
| 37 |
+
"""
|
| 38 |
+
|
| 39 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 40 |
+
headers = {"Content-Type": "application/json"}
|
| 41 |
+
|
| 42 |
+
# Iterate over all images
|
| 43 |
+
for image_name in os.listdir(cr1_images_folder):
|
| 44 |
+
if not image_name.lower().endswith(('.jpg', '.jpeg', '.png')):
|
| 45 |
+
continue
|
| 46 |
+
|
| 47 |
+
image_path = os.path.join(cr1_images_folder, image_name)
|
| 48 |
+
base_name = os.path.splitext(image_name)[0]
|
| 49 |
+
output_file = os.path.join(output_json_folder, base_name + ".json")
|
| 50 |
+
|
| 51 |
+
# Skip if JSON already exists
|
| 52 |
+
if os.path.exists(output_file):
|
| 53 |
+
print(f"Skipped {image_name} (JSON file already exists)")
|
| 54 |
+
continue
|
| 55 |
+
|
| 56 |
+
# Read image and convert to base64
|
| 57 |
+
with open(image_path, "rb") as f:
|
| 58 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 59 |
+
|
| 60 |
+
# Send request to Gemini API
|
| 61 |
+
data = {
|
| 62 |
+
"contents": [
|
| 63 |
+
{
|
| 64 |
+
"role": "user",
|
| 65 |
+
"parts": [
|
| 66 |
+
{"text": prompt},
|
| 67 |
+
{
|
| 68 |
+
"inline_data": {
|
| 69 |
+
"mime_type": "image/jpeg",
|
| 70 |
+
"data": image_b64
|
| 71 |
+
}
|
| 72 |
+
}
|
| 73 |
+
]
|
| 74 |
+
}
|
| 75 |
+
]
|
| 76 |
+
}
|
| 77 |
+
|
| 78 |
+
try:
|
| 79 |
+
response = requests.post(url, headers=headers, json=data)
|
| 80 |
+
response.raise_for_status()
|
| 81 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 82 |
+
|
| 83 |
+
match = re.search(r"```json\s*(\{.*\})\s*```", response_text, re.DOTALL)
|
| 84 |
+
if match:
|
| 85 |
+
json_text = match.group(1)
|
| 86 |
+
result = json.loads(json_text)
|
| 87 |
+
|
| 88 |
+
with open(output_file, "w", encoding="utf-8") as f:
|
| 89 |
+
json.dump(result, f, ensure_ascii=False, indent=2)
|
| 90 |
+
|
| 91 |
+
print(f"✅ Processed: {image_name}")
|
| 92 |
+
else:
|
| 93 |
+
print(f"❌ Failed to extract JSON from: {image_name}")
|
| 94 |
+
print(response_text)
|
| 95 |
+
|
| 96 |
+
except Exception as e:
|
| 97 |
+
print(f"❌ Error processing image {image_name}: {e}")
|
json-V3.py
ADDED
|
@@ -0,0 +1,104 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
|
| 7 |
+
# مفتاح Gemini API
|
| 8 |
+
API_KEY = "AIzaSyBr2-dUqHDZkk20hlWeEcpWnVVdkq9fqyE"
|
| 9 |
+
|
| 10 |
+
# المجلد الذي يحتوي صور cr1
|
| 11 |
+
cr1_images_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1"
|
| 12 |
+
# مجلد إخراج ملفات JSON
|
| 13 |
+
output_json_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1\cr1_json"
|
| 14 |
+
|
| 15 |
+
# Ensure output folder exists
|
| 16 |
+
os.makedirs(output_json_folder, exist_ok=True)
|
| 17 |
+
|
| 18 |
+
# Exact same prompt
|
| 19 |
+
prompt = """
|
| 20 |
+
استخرج الحقول التالية من مستند باللغة العربية فقط، وأرجعها بصيغة JSON:
|
| 21 |
+
|
| 22 |
+
1. اسم المكلف
|
| 23 |
+
2. عنوان المركز الرئيسي
|
| 24 |
+
3. المدينة
|
| 25 |
+
4. الحي
|
| 26 |
+
5. صندوق البريد
|
| 27 |
+
6. الرمز البريدي
|
| 28 |
+
7. الهاتف
|
| 29 |
+
|
| 30 |
+
الصيغة المطلوبة:
|
| 31 |
+
|
| 32 |
+
{
|
| 33 |
+
"اسم المكلف": "...",
|
| 34 |
+
"عنوان المركز الرئيسي": "...",
|
| 35 |
+
"المدينة": "...",
|
| 36 |
+
"الحي": "...",
|
| 37 |
+
"صندوق البريد": "...",
|
| 38 |
+
"الرمز البريدي": "...",
|
| 39 |
+
"الهاتف": "..."
|
| 40 |
+
}
|
| 41 |
+
|
| 42 |
+
إذا لم يوجد حقل، أرجعه كـ null.
|
| 43 |
+
"""
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 47 |
+
headers = {"Content-Type": "application/json"}
|
| 48 |
+
|
| 49 |
+
# Iterate over all images
|
| 50 |
+
for image_name in os.listdir(cr1_images_folder):
|
| 51 |
+
if not image_name.lower().endswith(('.jpg', '.jpeg', '.png')):
|
| 52 |
+
continue
|
| 53 |
+
|
| 54 |
+
image_path = os.path.join(cr1_images_folder, image_name)
|
| 55 |
+
base_name = os.path.splitext(image_name)[0]
|
| 56 |
+
output_file = os.path.join(output_json_folder, base_name + ".json")
|
| 57 |
+
|
| 58 |
+
# Skip if JSON already exists
|
| 59 |
+
if os.path.exists(output_file):
|
| 60 |
+
print(f"Skipped {image_name} (JSON file already exists)")
|
| 61 |
+
continue
|
| 62 |
+
|
| 63 |
+
# Read image and convert to base64
|
| 64 |
+
with open(image_path, "rb") as f:
|
| 65 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 66 |
+
|
| 67 |
+
# Send request to Gemini API
|
| 68 |
+
data = {
|
| 69 |
+
"contents": [
|
| 70 |
+
{
|
| 71 |
+
"role": "user",
|
| 72 |
+
"parts": [
|
| 73 |
+
{"text": prompt},
|
| 74 |
+
{
|
| 75 |
+
"inline_data": {
|
| 76 |
+
"mime_type": "image/jpeg",
|
| 77 |
+
"data": image_b64
|
| 78 |
+
}
|
| 79 |
+
}
|
| 80 |
+
]
|
| 81 |
+
}
|
| 82 |
+
]
|
| 83 |
+
}
|
| 84 |
+
|
| 85 |
+
try:
|
| 86 |
+
response = requests.post(url, headers=headers, json=data)
|
| 87 |
+
response.raise_for_status()
|
| 88 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 89 |
+
|
| 90 |
+
match = re.search(r"```json\s*(\{.*\})\s*```", response_text, re.DOTALL)
|
| 91 |
+
if match:
|
| 92 |
+
json_text = match.group(1)
|
| 93 |
+
result = json.loads(json_text)
|
| 94 |
+
|
| 95 |
+
with open(output_file, "w", encoding="utf-8") as f:
|
| 96 |
+
json.dump(result, f, ensure_ascii=False, indent=2)
|
| 97 |
+
|
| 98 |
+
print(f"✅ Processed: {image_name}")
|
| 99 |
+
else:
|
| 100 |
+
print(f"❌ Failed to extract JSON from: {image_name}")
|
| 101 |
+
print(response_text)
|
| 102 |
+
|
| 103 |
+
except Exception as e:
|
| 104 |
+
print(f"❌ Error processing image {image_name}: {e}")
|
json-v1.py
ADDED
|
@@ -0,0 +1,92 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import base64
|
| 2 |
+
import json
|
| 3 |
+
import re
|
| 4 |
+
import requests
|
| 5 |
+
import os
|
| 6 |
+
|
| 7 |
+
# مفتاح Gemini API
|
| 8 |
+
API_KEY = "AIzaSyBr2-dUqHDZkk20hlWeEcpWnVVdkq9fqyE"
|
| 9 |
+
|
| 10 |
+
# المجلد الذي يحتوي صور cr1
|
| 11 |
+
cr1_images_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1"
|
| 12 |
+
# مجلد إخراج ملفات JSON
|
| 13 |
+
output_json_folder = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch\classified\CR1\cr1_json"
|
| 14 |
+
|
| 15 |
+
# Ensure output folder exists
|
| 16 |
+
os.makedirs(output_json_folder, exist_ok=True)
|
| 17 |
+
|
| 18 |
+
# Exact same prompt
|
| 19 |
+
prompt = """
|
| 20 |
+
Please extract the following fields in Arabic and English from the tax registration document image:
|
| 21 |
+
Taxpayer Name
|
| 22 |
+
VAT Registration Number
|
| 23 |
+
Effective Registration Date
|
| 24 |
+
Taxpayer Address
|
| 25 |
+
|
| 26 |
+
Return the result in a JSON format with these keys:
|
| 27 |
+
en_taxpayer_name, en_vat_registration_number, en_effective_registration_date, en_taxpayer_address,
|
| 28 |
+
ar_taxpayer_name, ar_vat_registration_number, ar_effective_registration_date, ar_taxpayer_address
|
| 29 |
+
|
| 30 |
+
If a field is missing, return null.
|
| 31 |
+
"""
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
url = f"https://generativelanguage.googleapis.com/v1/models/gemini-1.5-flash:generateContent?key={API_KEY}"
|
| 35 |
+
headers = {"Content-Type": "application/json"}
|
| 36 |
+
|
| 37 |
+
# Iterate over all images
|
| 38 |
+
for image_name in os.listdir(cr1_images_folder):
|
| 39 |
+
if not image_name.lower().endswith(('.jpg', '.jpeg', '.png')):
|
| 40 |
+
continue
|
| 41 |
+
|
| 42 |
+
image_path = os.path.join(cr1_images_folder, image_name)
|
| 43 |
+
base_name = os.path.splitext(image_name)[0]
|
| 44 |
+
output_file = os.path.join(output_json_folder, base_name + ".json")
|
| 45 |
+
|
| 46 |
+
# Skip if JSON already exists
|
| 47 |
+
if os.path.exists(output_file):
|
| 48 |
+
print(f"Skipped {image_name} (JSON file already exists)")
|
| 49 |
+
continue
|
| 50 |
+
|
| 51 |
+
# Read image and convert to base64
|
| 52 |
+
with open(image_path, "rb") as f:
|
| 53 |
+
image_b64 = base64.b64encode(f.read()).decode()
|
| 54 |
+
|
| 55 |
+
# Send request to Gemini API
|
| 56 |
+
data = {
|
| 57 |
+
"contents": [
|
| 58 |
+
{
|
| 59 |
+
"role": "user",
|
| 60 |
+
"parts": [
|
| 61 |
+
{"text": prompt},
|
| 62 |
+
{
|
| 63 |
+
"inline_data": {
|
| 64 |
+
"mime_type": "image/jpeg",
|
| 65 |
+
"data": image_b64
|
| 66 |
+
}
|
| 67 |
+
}
|
| 68 |
+
]
|
| 69 |
+
}
|
| 70 |
+
]
|
| 71 |
+
}
|
| 72 |
+
|
| 73 |
+
try:
|
| 74 |
+
response = requests.post(url, headers=headers, json=data)
|
| 75 |
+
response.raise_for_status()
|
| 76 |
+
response_text = response.json()['candidates'][0]['content']['parts'][0]['text']
|
| 77 |
+
|
| 78 |
+
match = re.search(r"```json\s*(\{.*\})\s*```", response_text, re.DOTALL)
|
| 79 |
+
if match:
|
| 80 |
+
json_text = match.group(1)
|
| 81 |
+
result = json.loads(json_text)
|
| 82 |
+
|
| 83 |
+
with open(output_file, "w", encoding="utf-8") as f:
|
| 84 |
+
json.dump(result, f, ensure_ascii=False, indent=2)
|
| 85 |
+
|
| 86 |
+
print(f"✅ Processed: {image_name}")
|
| 87 |
+
else:
|
| 88 |
+
print(f"❌ Failed to extract JSON from: {image_name}")
|
| 89 |
+
print(response_text)
|
| 90 |
+
|
| 91 |
+
except Exception as e:
|
| 92 |
+
print(f"❌ Error processing image {image_name}: {e}")
|
orch.py
ADDED
|
@@ -0,0 +1,73 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import shutil
|
| 3 |
+
from ultralytics import YOLO
|
| 4 |
+
|
| 5 |
+
# =======================================
|
| 6 |
+
# 🟡🔧 User-configurable settings
|
| 7 |
+
# =======================================
|
| 8 |
+
|
| 9 |
+
# 📁 Folder containing images to classify
|
| 10 |
+
INPUT_FOLDER = r"C:\Users\ASUS\OneDrive - Binder\Desktop\test.orch" # ← Modify this path
|
| 11 |
+
|
| 12 |
+
# 📁 Folder to save classified images
|
| 13 |
+
OUTPUT_BASE = os.path.join(INPUT_FOLDER, "classified") # ← Automatically inside input folder
|
| 14 |
+
|
| 15 |
+
# 📦 Path to YOLOv8 classification model
|
| 16 |
+
MODEL_PATH = r"C:\Users\ASUS\Downloads\best.pt" # ← Modify this path
|
| 17 |
+
|
| 18 |
+
# ✅ Allowed classes only
|
| 19 |
+
ALLOWED_CLASSES = {"CR1", "CR2", "CR3", "CR4", "CR5", "CR6", "CR7", "b1", "b2", "b3", "b4", "v1", "v2", "v3"}
|
| 20 |
+
|
| 21 |
+
# =======================================
|
| 22 |
+
# 🚀 Load the model
|
| 23 |
+
# =======================================
|
| 24 |
+
model = YOLO(MODEL_PATH)
|
| 25 |
+
|
| 26 |
+
# =======================================
|
| 27 |
+
# 🧠 Classify a single image
|
| 28 |
+
# =======================================
|
| 29 |
+
def classify_image_yolo(image_path):
|
| 30 |
+
try:
|
| 31 |
+
results = model(image_path, verbose=False)[0]
|
| 32 |
+
class_id = int(results.probs.top1)
|
| 33 |
+
class_name = model.names.get(class_id, "others")
|
| 34 |
+
return class_name
|
| 35 |
+
except Exception as e:
|
| 36 |
+
print(f"❌ Error classifying image {image_path}: {e}")
|
| 37 |
+
return "others"
|
| 38 |
+
|
| 39 |
+
# =======================================
|
| 40 |
+
# 📂 Classify and move images
|
| 41 |
+
# =======================================
|
| 42 |
+
processed_files = set()
|
| 43 |
+
|
| 44 |
+
for filename in os.listdir(INPUT_FOLDER):
|
| 45 |
+
file_path = os.path.join(INPUT_FOLDER, filename)
|
| 46 |
+
|
| 47 |
+
if not os.path.isfile(file_path):
|
| 48 |
+
continue
|
| 49 |
+
|
| 50 |
+
if filename in processed_files:
|
| 51 |
+
continue
|
| 52 |
+
|
| 53 |
+
class_name = classify_image_yolo(file_path)
|
| 54 |
+
|
| 55 |
+
# ✅ Check if class is allowed, otherwise assign to "others"
|
| 56 |
+
if class_name not in ALLOWED_CLASSES:
|
| 57 |
+
class_name = "others"
|
| 58 |
+
|
| 59 |
+
# Create class folder if not exists
|
| 60 |
+
output_folder = os.path.join(OUTPUT_BASE, class_name)
|
| 61 |
+
os.makedirs(output_folder, exist_ok=True)
|
| 62 |
+
|
| 63 |
+
destination_path = os.path.join(output_folder, filename)
|
| 64 |
+
|
| 65 |
+
# Avoid duplicate images
|
| 66 |
+
if not os.path.exists(destination_path):
|
| 67 |
+
shutil.move(file_path, destination_path)
|
| 68 |
+
print(f"✅ Moved image {filename} to [{class_name}]")
|
| 69 |
+
processed_files.add(filename)
|
| 70 |
+
else:
|
| 71 |
+
print(f"⚠️ Image {filename} already exists in [{class_name}], skipped")
|
| 72 |
+
|
| 73 |
+
print("🎉 Finished classifying images using YOLOv8 Classification!")
|