Spaces:
Running
Running
File size: 7,998 Bytes
e69be74 4a10a29 eba303d e69be74 eba303d 4a10a29 eba303d 4a10a29 e69be74 4a10a29 eba303d 4a10a29 e69be74 4a10a29 e69be74 4a10a29 e69be74 4a10a29 eba303d 4a10a29 e69be74 4a10a29 eba303d e69be74 4a10a29 eba303d 4a10a29 eba303d 4a10a29 eba303d 4a10a29 eba303d 4a10a29 e69be74 4a10a29 e69be74 4a10a29 eba303d 4a10a29 e69be74 4a10a29 e69be74 4a10a29 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 |
import json
import os
import time
from loguru import logger
from openai import OpenAI
from common.enum.ai_service_error import AiServiceError
from common.exceptions import AiServiceException
from common.utils import encode_image_to_webp_base64
from image_processing_interface import ImageProcessingInterface
class OpenAIService(ImageProcessingInterface):
_instance = None
TOOLS = [
{
"type": "function",
"function": {
"name": "parse_image",
"description": "Parses receipt data from image into a structured JSON format.",
"parameters": {
"type": "object",
"properties": {
"store_name": {"type": "string"},
"country": {"type": "string"},
"receipt_type": {"type": "string"},
"address": {"type": "string"},
"datetime": {"type": "string"},
"currency": {"type": "string"},
"sub_total_amount": {"type": "number"},
"total_price": {"type": "number"},
"total_discount": {"type": "number"},
"all_items_price_with_tax": {"type": "boolean"},
"payment_method": {
"type": "string",
"enum": ["card", "cash", "unknown"]
},
"rounding": {"type": "number"},
"tax": {"type": "number"},
"taxes_not_included_sum": {"type": "number"},
"tips": {"type": "number"},
"items": {
"type": "array",
"items": {
"type": "object",
"properties": {
"name": {"type": "string"},
"unit_price": {"type": "number"},
"quantity": {"type": "number"},
"measurement_unit": {"type": "string"},
"total_price_without_discount": {"type": "number"},
"total_price_with_discount": {"type": "number"},
"discount": {"type": "number"},
"category": {"type": "string"},
"item_price_with_tax": {"type": "boolean"},
},
"required": ["name", "unit_price", "quantity", "total_price_without_discount"]
}
},
"taxs_items": {
"type": "array",
"items": {
"type": "object",
"properties": {
"tax_name": {"type": "string"},
"percentage": {"type": "number"},
"tax_from_amount": {"type": "number"},
"tax": {"type": "number"},
"total": {"type": "number"},
"tax_included": {"type": "boolean"},
},
}
}
},
"required": ["total_price", "items"]
}
}
}
]
def __new__(cls, *args, **kwargs):
if cls._instance is None:
cls._instance = super(OpenAIService, cls).__new__(cls)
return cls._instance
def __init__(self, api_key=None):
if not hasattr(self, "_initialized"):
self.api_key = api_key or os.environ.get("OPENAI_API_KEY")
if self.api_key:
logger.info("OPENAI_API_KEY was found.")
else:
raise ValueError("OPENAI_API_KEY not found.")
self.client = OpenAI(api_key=self.api_key)
self._initialized = True
def process_image(self, input_image64, model_name, prompt, system="You are a receipt recognizer!", temperature=0.0):
if not input_image64:
raise ValueError("No image provided.")
try:
start_time = time.time()
response = self.client.chat.completions.create(
model=model_name,
messages=[
{"role": "system", "content": system},
{"role": "user", "content": [
{"type": "text", "text": prompt},
{"type": "image_url", "image_url": {
"url": f"data:image/webp;base64,{input_image64}"}
}
]}
],
tools=self.TOOLS,
temperature=temperature,
response_format={"type": "json_object"}
)
end_time = time.time()
logger.info(f"Recognition spent {end_time - start_time:.2f} seconds.")
# Extract the function call result
if not response.choices:
raise ValueError("The API response does not contain valid choices.")
choice = response.choices[0]
if not choice.message.tool_calls:
raise ValueError(choice.message.content or "No tool calls found in the API response.")
function_call = choice.message.tool_calls[0]
if not (function_call.function and function_call.function.arguments):
raise ValueError("No valid function call data found in the API response.")
logger.debug(f"Raw API Response: {function_call.function.arguments}")
try:
json_content = json.loads(function_call.function.arguments)
except json.JSONDecodeError:
error_message = f"The receipt could not be recognized. Please retake the photo."
logger.error(error_message)
raise AiServiceException(AiServiceError.RETAKE_PHOTO, error_message)
if not self._validate_receipt_data(json_content):
error_message = f"The receipt is empty or contains no valid items. Please ensure the receipt is correctly scanned and try again"
logger.error(error_message)
raise AiServiceException(AiServiceError.RETAKE_PHOTO, error_message)
json_content = self._add_total_price(json_content)
json_content = self._add_discount_item(json_content)
json_content = self._add_rounding_item(json_content)
# Add token usage information
json_content['input_tokens'] = response.usage.prompt_tokens
json_content['output_tokens'] = response.usage.completion_tokens
json_content['total_tokens'] = response.usage.total_tokens
json_content['time'] = end_time - start_time
model_input = {
"system": system,
"prompt": prompt
}
return json.dumps(json_content, indent=4), model_input
except Exception as e:
raise RuntimeError(f"Failed to process the image: {str(e)}")
if __name__ == "__main__":
try:
processor = OpenAIService()
# Image processing
image_path = "./examples/fatlouis.webp"
input_image64 = encode_image_to_webp_base64(image_path)
system = "You are a receipt recognizer."
with open('common/prompt_v1.txt', 'r', encoding='utf-8') as file:
prompt = file.read()
result = processor.process_image(input_image64, "gpt-4o-mini", prompt, system, 0.0)
print(f'Image processing result: {result}')
except Exception as e:
print(f"Error: {e}")
|