Delete gemini_requestcr4.py
Browse files- gemini_requestcr4.py +0 -135
gemini_requestcr4.py
DELETED
|
@@ -1,135 +0,0 @@
|
|
| 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)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|