Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,23 +1,17 @@
|
|
| 1 |
import uvicorn
|
| 2 |
from fastapi.staticfiles import StaticFiles
|
| 3 |
import hashlib
|
| 4 |
-
from enum import Enum
|
| 5 |
from fastapi import FastAPI, Header, Query, Depends, HTTPException
|
| 6 |
-
from PIL import Image
|
| 7 |
-
import io
|
| 8 |
-
import fitz # PyMuPDF for PDF handling
|
| 9 |
-
import logging
|
| 10 |
from pymongo import MongoClient
|
| 11 |
-
|
| 12 |
import boto3
|
| 13 |
import openai
|
| 14 |
import os
|
| 15 |
-
import traceback
|
| 16 |
-
import re
|
| 17 |
import json
|
| 18 |
from dotenv import load_dotenv
|
| 19 |
import base64
|
| 20 |
from bson.objectid import ObjectId
|
|
|
|
| 21 |
|
| 22 |
db_client = None
|
| 23 |
load_dotenv()
|
|
@@ -26,26 +20,26 @@ load_dotenv()
|
|
| 26 |
logging.basicConfig(level=logging.INFO)
|
| 27 |
logger = logging.getLogger(__name__)
|
| 28 |
|
| 29 |
-
#
|
| 30 |
MONGODB_URI = os.getenv("MONGODB_URI")
|
| 31 |
DATABASE_NAME = os.getenv("DATABASE_NAME")
|
| 32 |
COLLECTION_NAME = os.getenv("COLLECTION_NAME", "invoice_collection")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
# Check if environment variables are set
|
| 38 |
-
if not MONGODB_URI:
|
| 39 |
-
raise ValueError("MONGODB_URL is not set. Please add it to Hugging Face secrets.")
|
| 40 |
|
| 41 |
# Initialize MongoDB Connection
|
| 42 |
db_client = MongoClient(MONGODB_URI)
|
| 43 |
db = db_client[DATABASE_NAME]
|
| 44 |
invoice_collection = db[COLLECTION_NAME]
|
|
|
|
| 45 |
|
| 46 |
app = FastAPI(docs_url='/')
|
| 47 |
-
use_gpu = False
|
| 48 |
-
output_dir = 'output'
|
| 49 |
|
| 50 |
@app.on_event("startup")
|
| 51 |
def startup_db():
|
|
@@ -55,15 +49,6 @@ def startup_db():
|
|
| 55 |
except Exception as e:
|
| 56 |
logger.error(f"MongoDB connection failed: {str(e)}")
|
| 57 |
|
| 58 |
-
# AWS S3 Configuration
|
| 59 |
-
API_KEY = os.getenv("API_KEY")
|
| 60 |
-
AWS_ACCESS_KEY = os.getenv("AWS_ACCESS_KEY")
|
| 61 |
-
AWS_SECRET_KEY = os.getenv("AWS_SECRET_KEY")
|
| 62 |
-
S3_BUCKET_NAME = os.getenv("S3_BUCKET_NAME")
|
| 63 |
-
|
| 64 |
-
# OpenAI Configuration
|
| 65 |
-
openai.api_key = os.getenv("OPENAI_API_KEY")
|
| 66 |
-
|
| 67 |
# S3 Client
|
| 68 |
s3_client = boto3.client(
|
| 69 |
's3',
|
|
@@ -71,247 +56,59 @@ s3_client = boto3.client(
|
|
| 71 |
aws_secret_access_key=AWS_SECRET_KEY
|
| 72 |
)
|
| 73 |
|
| 74 |
-
# Function to fetch file from S3
|
| 75 |
def fetch_file_from_s3(file_key):
|
| 76 |
try:
|
| 77 |
response = s3_client.get_object(Bucket=S3_BUCKET_NAME, Key=file_key)
|
| 78 |
-
|
| 79 |
-
file_data = response['Body'].read()
|
| 80 |
-
return file_data, content_type # Return file data as BytesIO
|
| 81 |
except Exception as e:
|
| 82 |
-
|
| 83 |
-
|
| 84 |
-
# Function to summarize text using OpenAI GPT
|
| 85 |
-
def extract_invoice_data(file_data, content_type):
|
| 86 |
-
system_prompt = "You are an expert in document data extraction."
|
| 87 |
-
|
| 88 |
-
# Convert file to Base64
|
| 89 |
-
base64_encoded = base64.b64encode(file_data).decode('utf-8')
|
| 90 |
-
|
| 91 |
-
# Determine the correct MIME type for OpenAI
|
| 92 |
-
if content_type.startswith("image/"):
|
| 93 |
-
mime_type = content_type # e.g., image/png, image/jpeg
|
| 94 |
-
elif content_type == "application/pdf":
|
| 95 |
-
mime_type = "application/pdf"
|
| 96 |
-
else:
|
| 97 |
-
raise ValueError(f"Unsupported content type: {content_type}")
|
| 98 |
-
|
| 99 |
-
try:
|
| 100 |
-
response = openai.ChatCompletion.create(
|
| 101 |
-
model="gpt-4o-mini",
|
| 102 |
-
messages=[
|
| 103 |
-
{"role": "system", "content": system_prompt},
|
| 104 |
-
{
|
| 105 |
-
"role": "user",
|
| 106 |
-
"content": [
|
| 107 |
-
{
|
| 108 |
-
"type": "image_url",
|
| 109 |
-
"image_url": {
|
| 110 |
-
"url": f"data:{mime_type};base64,{base64_encoded}"
|
| 111 |
-
}
|
| 112 |
-
}
|
| 113 |
-
]
|
| 114 |
-
}
|
| 115 |
-
],
|
| 116 |
-
response_format={
|
| 117 |
-
"type": "json_schema",
|
| 118 |
-
"json_schema": {
|
| 119 |
-
"name": "invoice",
|
| 120 |
-
"strict": True,
|
| 121 |
-
"schema": {
|
| 122 |
-
"type": "object",
|
| 123 |
-
"title": "Invoice Information Extractor",
|
| 124 |
-
"$schema": "http://json-schema.org/draft-07/schema#",
|
| 125 |
-
"properties": {
|
| 126 |
-
"LineItems": {
|
| 127 |
-
"type": "array",
|
| 128 |
-
"items": {
|
| 129 |
-
"type": "object",
|
| 130 |
-
"required": [
|
| 131 |
-
"ProductCode",
|
| 132 |
-
"Description",
|
| 133 |
-
"Amount"
|
| 134 |
-
],
|
| 135 |
-
"properties": {
|
| 136 |
-
"ProductCode": {
|
| 137 |
-
"type": "string",
|
| 138 |
-
"title": "Product Code",
|
| 139 |
-
"description": "The code of the product"
|
| 140 |
-
},
|
| 141 |
-
"Description": {
|
| 142 |
-
"type": "string",
|
| 143 |
-
"title": "Description",
|
| 144 |
-
"description": "Description of the product"
|
| 145 |
-
},
|
| 146 |
-
"Amount": {
|
| 147 |
-
"type": "number",
|
| 148 |
-
"title": "Amount",
|
| 149 |
-
"description": "The amount of the product"
|
| 150 |
-
}
|
| 151 |
-
},
|
| 152 |
-
"additionalProperties": False
|
| 153 |
-
},
|
| 154 |
-
"title": "Line Items",
|
| 155 |
-
"description": "List of line items on the invoice"
|
| 156 |
-
},
|
| 157 |
-
"TaxAmount": {
|
| 158 |
-
"type": "number",
|
| 159 |
-
"title": "Tax Amount",
|
| 160 |
-
"description": "The tax amount on the invoice"
|
| 161 |
-
},
|
| 162 |
-
"VendorGST": {
|
| 163 |
-
"type": "string",
|
| 164 |
-
"title": "Vendor GST",
|
| 165 |
-
"description": "The GST number of the vendor"
|
| 166 |
-
},
|
| 167 |
-
"VendorName": {
|
| 168 |
-
"type": "string",
|
| 169 |
-
"title": "Vendor Name",
|
| 170 |
-
"description": "The name of the vendor"
|
| 171 |
-
},
|
| 172 |
-
"InvoiceDate": {
|
| 173 |
-
"type": "string",
|
| 174 |
-
"title": "Invoice Date",
|
| 175 |
-
"description": "The date of the invoice (format: dd-MMM-yyyy)"
|
| 176 |
-
},
|
| 177 |
-
"TotalAmount": {
|
| 178 |
-
"type": "number",
|
| 179 |
-
"title": "Total Amount",
|
| 180 |
-
"description": "The total amount on the invoice"
|
| 181 |
-
},
|
| 182 |
-
"InvoiceNumber": {
|
| 183 |
-
"type": "string",
|
| 184 |
-
"title": "Invoice Number",
|
| 185 |
-
"description": "The number of the invoice"
|
| 186 |
-
},
|
| 187 |
-
"VendorAddress": {
|
| 188 |
-
"type": "string",
|
| 189 |
-
"title": "Vendor Address",
|
| 190 |
-
"description": "The address of the vendor"
|
| 191 |
-
},
|
| 192 |
-
"InvoiceCurrency": {
|
| 193 |
-
"type": "string",
|
| 194 |
-
"title": "Invoice Currency",
|
| 195 |
-
"description": "The currency used in the invoice (e.g., USD, INR, AUD)"
|
| 196 |
-
}
|
| 197 |
-
},
|
| 198 |
-
"required": [
|
| 199 |
-
"LineItems",
|
| 200 |
-
"TaxAmount",
|
| 201 |
-
"VendorGST",
|
| 202 |
-
"VendorName",
|
| 203 |
-
"InvoiceDate",
|
| 204 |
-
"TotalAmount",
|
| 205 |
-
"InvoiceNumber",
|
| 206 |
-
"VendorAddress",
|
| 207 |
-
"InvoiceCurrency"
|
| 208 |
-
],
|
| 209 |
-
"additionalProperties": False,
|
| 210 |
-
"description": "Schema for extracting structured invoice data"
|
| 211 |
-
}
|
| 212 |
-
}
|
| 213 |
-
},
|
| 214 |
-
temperature=0.5,
|
| 215 |
-
max_tokens=16384
|
| 216 |
-
)
|
| 217 |
|
| 218 |
-
|
| 219 |
-
|
| 220 |
-
cleaned_content = content.strip().strip('```json').strip('```')
|
| 221 |
|
| 222 |
-
|
| 223 |
-
|
| 224 |
-
return parsed_content
|
| 225 |
-
except json.JSONDecodeError as e:
|
| 226 |
-
logger.error(f"JSON Parse Error: {e}")
|
| 227 |
-
return None
|
| 228 |
|
| 229 |
-
|
| 230 |
-
|
| 231 |
-
return {"error": str(e)}
|
| 232 |
|
| 233 |
-
# Dependency to check API Key
|
| 234 |
def verify_api_key(api_key: str = Header(...)):
|
| 235 |
if api_key != API_KEY:
|
| 236 |
raise HTTPException(status_code=401, detail="Invalid API Key")
|
| 237 |
|
| 238 |
-
@app.get("/")
|
| 239 |
-
def read_root():
|
| 240 |
-
return {"message": "Welcome to the Invoice Summarization API!"}
|
| 241 |
-
|
| 242 |
-
def get_content_type_from_s3(file_key):
|
| 243 |
-
"""Fetch the content type (MIME type) of a file stored in S3."""
|
| 244 |
-
try:
|
| 245 |
-
response = s3_client.head_object(Bucket=S3_BUCKET_NAME, Key=file_key)
|
| 246 |
-
return response.get('ContentType', 'application/octet-stream') # Default to binary if not found
|
| 247 |
-
except Exception as e:
|
| 248 |
-
raise Exception(f"Failed to get content type from S3: {str(e)}")
|
| 249 |
-
|
| 250 |
-
|
| 251 |
@app.get("/ocr/extraction")
|
| 252 |
def extract_text_from_file(
|
| 253 |
api_key: str = Depends(verify_api_key),
|
| 254 |
-
file_key: str = Query(..., description="S3 file key
|
| 255 |
-
document_type: str = Query(..., description="
|
| 256 |
entity_ref_key: str = Query(..., description="Entity Reference Key")
|
| 257 |
):
|
| 258 |
-
"""Extract text from a PDF or Image stored in S3 and process it based on document size."""
|
| 259 |
try:
|
| 260 |
existing_document = invoice_collection.find_one({"entityrefkey": entity_ref_key})
|
| 261 |
-
|
| 262 |
if existing_document:
|
| 263 |
existing_document["_id"] = str(existing_document["_id"])
|
| 264 |
-
return {
|
| 265 |
-
|
| 266 |
-
|
| 267 |
-
}
|
| 268 |
-
|
| 269 |
-
# Retrieve file from S3 and determine content type (Ensure this step is implemented)
|
| 270 |
-
content_type = get_content_type_from_s3(file_key) # Implement this function
|
| 271 |
-
|
| 272 |
-
# Extract text (Ensure Extraction function is implemented)
|
| 273 |
extracted_text, num_pages = extract_text_from_s3(file_key, content_type)
|
| 274 |
-
|
| 275 |
-
# Define values for small/large files
|
| 276 |
base64DataResp = None
|
| 277 |
summary = None
|
| 278 |
if num_pages <= 2:
|
| 279 |
-
base64DataResp = convert_to_base64(file_key)
|
| 280 |
-
summary = generate_summary(extracted_text)
|
| 281 |
-
|
| 282 |
-
# Store extracted data in MongoDB
|
| 283 |
document = {
|
| 284 |
"file_key": file_key,
|
| 285 |
"file_type": content_type,
|
| 286 |
"document_type": document_type,
|
| 287 |
"entityrefkey": entity_ref_key,
|
| 288 |
"num_pages": num_pages,
|
| 289 |
-
"base64DataResp": base64DataResp,
|
| 290 |
-
"extracted_text": extracted_text,
|
| 291 |
-
"summary": summary,
|
| 292 |
}
|
| 293 |
-
|
| 294 |
inserted_doc = invoice_collection.insert_one(document)
|
| 295 |
-
document_id
|
| 296 |
-
|
| 297 |
-
return {
|
| 298 |
-
"message": "Document successfully stored in MongoDB",
|
| 299 |
-
"document_id": document_id,
|
| 300 |
-
"file_key": file_key,
|
| 301 |
-
"num_pages": num_pages,
|
| 302 |
-
"summary": summary if summary else "Skipped for large documents"
|
| 303 |
-
}
|
| 304 |
-
|
| 305 |
except Exception as e:
|
| 306 |
-
|
| 307 |
-
"error_type": type(e).__name__,
|
| 308 |
-
"error_message": str(e),
|
| 309 |
-
"traceback": traceback.format_exc()
|
| 310 |
-
}
|
| 311 |
-
return {"error": error_details}
|
| 312 |
-
|
| 313 |
-
# Serve the output folder as static files
|
| 314 |
-
app.mount("/output", StaticFiles(directory="output", follow_symlink=True, html=True), name="output")
|
| 315 |
-
|
| 316 |
-
if __name__ == '__main__':
|
| 317 |
-
uvicorn.run(app=app)
|
|
|
|
| 1 |
import uvicorn
|
| 2 |
from fastapi.staticfiles import StaticFiles
|
| 3 |
import hashlib
|
|
|
|
| 4 |
from fastapi import FastAPI, Header, Query, Depends, HTTPException
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
from pymongo import MongoClient
|
|
|
|
| 6 |
import boto3
|
| 7 |
import openai
|
| 8 |
import os
|
| 9 |
+
import traceback
|
|
|
|
| 10 |
import json
|
| 11 |
from dotenv import load_dotenv
|
| 12 |
import base64
|
| 13 |
from bson.objectid import ObjectId
|
| 14 |
+
import logging
|
| 15 |
|
| 16 |
db_client = None
|
| 17 |
load_dotenv()
|
|
|
|
| 20 |
logging.basicConfig(level=logging.INFO)
|
| 21 |
logger = logging.getLogger(__name__)
|
| 22 |
|
| 23 |
+
# Validate and Load Environment Variables
|
| 24 |
MONGODB_URI = os.getenv("MONGODB_URI")
|
| 25 |
DATABASE_NAME = os.getenv("DATABASE_NAME")
|
| 26 |
COLLECTION_NAME = os.getenv("COLLECTION_NAME", "invoice_collection")
|
| 27 |
+
AWS_ACCESS_KEY = os.getenv("AWS_ACCESS_KEY")
|
| 28 |
+
AWS_SECRET_KEY = os.getenv("AWS_SECRET_KEY")
|
| 29 |
+
S3_BUCKET_NAME = os.getenv("S3_BUCKET_NAME")
|
| 30 |
+
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
|
| 31 |
+
API_KEY = os.getenv("API_KEY")
|
| 32 |
|
| 33 |
+
if not all([MONGODB_URI, DATABASE_NAME, AWS_ACCESS_KEY, AWS_SECRET_KEY, S3_BUCKET_NAME, OPENAI_API_KEY, API_KEY]):
|
| 34 |
+
raise ValueError("One or more required environment variables are missing.")
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
|
| 36 |
# Initialize MongoDB Connection
|
| 37 |
db_client = MongoClient(MONGODB_URI)
|
| 38 |
db = db_client[DATABASE_NAME]
|
| 39 |
invoice_collection = db[COLLECTION_NAME]
|
| 40 |
+
openai.api_key = OPENAI_API_KEY
|
| 41 |
|
| 42 |
app = FastAPI(docs_url='/')
|
|
|
|
|
|
|
| 43 |
|
| 44 |
@app.on_event("startup")
|
| 45 |
def startup_db():
|
|
|
|
| 49 |
except Exception as e:
|
| 50 |
logger.error(f"MongoDB connection failed: {str(e)}")
|
| 51 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
# S3 Client
|
| 53 |
s3_client = boto3.client(
|
| 54 |
's3',
|
|
|
|
| 56 |
aws_secret_access_key=AWS_SECRET_KEY
|
| 57 |
)
|
| 58 |
|
|
|
|
| 59 |
def fetch_file_from_s3(file_key):
|
| 60 |
try:
|
| 61 |
response = s3_client.get_object(Bucket=S3_BUCKET_NAME, Key=file_key)
|
| 62 |
+
return response['Body'].read(), response['ContentType']
|
|
|
|
|
|
|
| 63 |
except Exception as e:
|
| 64 |
+
logger.error(f"Failed to fetch file from S3: {e}")
|
| 65 |
+
raise HTTPException(status_code=500, detail=f"S3 fetch error: {str(e)}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 66 |
|
| 67 |
+
def extract_text_from_s3(file_key, content_type):
|
| 68 |
+
return "Extracted text from file", 1 # Placeholder for real extraction logic
|
|
|
|
| 69 |
|
| 70 |
+
def convert_to_base64(file_key):
|
| 71 |
+
return "Base64 encoded data" # Placeholder
|
|
|
|
|
|
|
|
|
|
|
|
|
| 72 |
|
| 73 |
+
def generate_summary(extracted_text):
|
| 74 |
+
return "Summarized text" # Placeholder
|
|
|
|
| 75 |
|
|
|
|
| 76 |
def verify_api_key(api_key: str = Header(...)):
|
| 77 |
if api_key != API_KEY:
|
| 78 |
raise HTTPException(status_code=401, detail="Invalid API Key")
|
| 79 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
@app.get("/ocr/extraction")
|
| 81 |
def extract_text_from_file(
|
| 82 |
api_key: str = Depends(verify_api_key),
|
| 83 |
+
file_key: str = Query(..., description="S3 file key"),
|
| 84 |
+
document_type: str = Query(..., description="Document type"),
|
| 85 |
entity_ref_key: str = Query(..., description="Entity Reference Key")
|
| 86 |
):
|
|
|
|
| 87 |
try:
|
| 88 |
existing_document = invoice_collection.find_one({"entityrefkey": entity_ref_key})
|
|
|
|
| 89 |
if existing_document:
|
| 90 |
existing_document["_id"] = str(existing_document["_id"])
|
| 91 |
+
return {"message": "Document Retrieved from MongoDB.", "document": existing_document}
|
| 92 |
+
|
| 93 |
+
file_data, content_type = fetch_file_from_s3(file_key)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 94 |
extracted_text, num_pages = extract_text_from_s3(file_key, content_type)
|
|
|
|
|
|
|
| 95 |
base64DataResp = None
|
| 96 |
summary = None
|
| 97 |
if num_pages <= 2:
|
| 98 |
+
base64DataResp = convert_to_base64(file_key)
|
| 99 |
+
summary = generate_summary(extracted_text)
|
| 100 |
+
|
|
|
|
| 101 |
document = {
|
| 102 |
"file_key": file_key,
|
| 103 |
"file_type": content_type,
|
| 104 |
"document_type": document_type,
|
| 105 |
"entityrefkey": entity_ref_key,
|
| 106 |
"num_pages": num_pages,
|
| 107 |
+
"base64DataResp": base64DataResp,
|
| 108 |
+
"extracted_text": extracted_text if num_pages <= 2 else "Skipped",
|
| 109 |
+
"summary": summary,
|
| 110 |
}
|
|
|
|
| 111 |
inserted_doc = invoice_collection.insert_one(document)
|
| 112 |
+
return {"message": "Document stored in MongoDB", "document_id": str(inserted_doc.inserted_id)}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 113 |
except Exception as e:
|
| 114 |
+
return {"error": {"type": type(e).__name__, "message": str(e), "traceback": traceback.format_exc()}}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|