Spaces:
Sleeping
Sleeping
Download app.py from m8chaa/re-sheet-develop: direct link, hf CLI and curl.
- Browser
- Download file 40.8 kB
-
https://huggingface.co/spaces/m8chaa/re-sheet-develop/resolve/main/app.py
- Command line
-
hf download hf://spaces/m8chaa/re-sheet-develop/app.py
-
curl -L -o app.py https://huggingface.co/spaces/m8chaa/re-sheet-develop/resolve/main/app.py
40.8 kB
from fastapi import FastAPI, HTTPException, UploadFile, File, Request, BackgroundTasks, Form
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel
from httpx import Timeout
import httpx
import logging
import base64
import firebase_admin
from firebase_admin import credentials
from firebase_admin import firestore
import json
from datetime import datetime, timedelta
import pytz
from typing import List, Optional, Dict
from google.oauth2 import service_account
from googleapiclient.discovery import build
from google.oauth2.credentials import Credentials
from google_auth_oauthlib.flow import InstalledAppFlow
from google.auth.transport.requests import Request as GoogleRequest
from googleapiclient.discovery import build
import os
import json
from io import BytesIO
import google.generativeai as genai
import requests
from google.cloud import vision
import aiohttp
from typing import Tuple
import time
from apscheduler.schedulers.background import BackgroundScheduler
from google.cloud.firestore import AsyncClient
import asyncio
from googleapiclient.errors import HttpError
from typing import Dict, Any
# start application
app = FastAPI()
# allow all origins
app.add_middleware(
CORSMiddleware,
allow_origins=["*"], # Allows all origins
allow_credentials=True, # Allows credentials (such as cookies) to be sent with requests
allow_methods=["*"], # Allows all methods
allow_headers=["*"], # Allows all headers
)
# set up logging
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
logging.info("Application starting up")
request_timeout = Timeout(1000.0, connect=600.0)
# set up firebase
# Load Firebase configuration from a file
with open('/app/secrets/fb_secret.json', 'r') as f:
firebase_configuration = json.load(f)
firebase_credentials = credentials.Certificate(firebase_configuration)
firebase_admin.initialize_app(firebase_credentials, {
'projectId': firebase_configuration['project_id'],
})
firestore_db = firestore.client()
users_collection_ref = firestore_db.collection('users')
transactions_collection_ref = firestore_db.collection('transactions')
SCOPES = ['https://www.googleapis.com/auth/spreadsheets','https://www.googleapis.com/auth/drive']
with open('/app/secrets/openai_key.txt', 'r') as f:
OPENAI_KEY = f.read()
APICredential = service_account.Credentials.from_service_account_file(
'/app/secrets/googleapi_secret.json', scopes=SCOPES)
sheet_service = build('sheets', 'v4', credentials=APICredential)
drive_service = build('drive', 'v3', credentials=APICredential)
with open('/app/secrets/gemini_key.txt', 'r') as f:
api_key = f.read()
genai.configure(api_key=api_key)
# Create the model
generation_config = {
"temperature": 1,
"top_p": 0.95,
"top_k": 64,
"max_output_tokens": 8192,
"response_mime_type": "application/json",
}
model = genai.GenerativeModel(
model_name="gemini-1.5-flash",
generation_config=generation_config,
# safety_settings = Adjust safety settings
# See https://ai.google.dev/gemini-api/docs/safety-settings
)
# first three are "등록일시", "썸네일", "바로가기",
receipt_sheet_headers = ['발행일', '상호', '사업자번호', '합계금액', '내역', '카드번호', '카드종류', '승인번호', '기타']
# at index 1 '등록 일시', at last index '명함보기'
business_card_sheet_headers = ['이름', '회사명', '소속 및 직위', '주소', '전화' ,'휴대폰', '팩스','e-mail','홈페이지']
sms_sheet_headers = ['날짜', '계좌', '입금', '출금', '잔액', '문자']
contact_sheet_headers = ['이름', '전화번호1', '전화번호2', '전화번호3', '이메일주소1', '이메일주소2', '이메일주소3', '그룹']
receipts_ss = 1395967745
business_cards_ss = 1733555840
contacts_ss = 1729750548
sms_ss = 1891574758
kst = pytz.timezone('Asia/Seoul')
class SheetUpdateRequest(BaseModel):
target_sheet_id: int
data: list
def convert_dicts_to_list(dicts: List[Dict], headers: List[str], add_link: bool = False,
image_id: str = '', link_text: str = '',
link_position: int = -1) -> List[List[str]]:
rows = []
# timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
timestamp = datetime.now(kst).strftime("%Y-%m-%d %H:%M:%S")
for dict_item in dicts:
# row = list(dict_item.values())
# row = [str(value) for value in dict_item.values()]
row = [str(dict_item.get(header, '')) for header in headers]
# Always add timestamp to the first index
row.insert(0, timestamp)
if add_link:
# Ensure the list length is at least 9
while len(row) < 9:
row.append('')
if link_position == -1: # Append link at the end
row.append(f'=HYPERLINK("https://drive.google.com/file/d/{image_id}/view?usp=sharing", "{link_text}")')
else:
row.insert(link_position, f'=HYPERLINK("https://drive.google.com/file/d/{image_id}/view?usp=sharing", "{link_text}")')
rows.append(row)
# print(rows)
return rows
async def download_file_from_drive(file_id: str) -> Tuple[bytes, str]:
"""
Downloads a file from Google Drive using the file ID
and returns the file content as bytes and the MIME type.
"""
try:
# URL for downloading the file from Google Drive
url = f"https://drive.google.com/uc?export=download&id={file_id}"
async with aiohttp.ClientSession() as session:
async with session.get(url) as response:
if response.status == 200:
content = await response.read()
mime_type = response.headers.get('Content-Type', 'application/octet-stream')
return content, mime_type
else:
raise HTTPException(status_code=500, detail=f"Failed to download file, status code: {response.status}")
except Exception as e:
logging.error(f"Error downloading file from Drive: {e}")
raise HTTPException(status_code=500, detail="Failed to download file from Drive")
async def request_gpt4o_completion(image_id: str, user_credit: int):
try:
# Step 1: Download the image file from Google Drive
file_content, mime_type = await download_file_from_drive(image_id)
# Step 3: Get text data from the image using Google Cloud Vision
client = vision.ImageAnnotatorClient()
image = vision.Image(content=file_content)
response = client.text_detection(image=image)
texts = response.text_annotations
if response.error.message:
logging.error(f"Error in Google Cloud Vision API: {response.error.message}")
raise Exception(
"{}\nFor more info on error messages, check: "
"https://cloud.google.com/apis/design/errors".format(response.error.message)
)
# Extract all detected text
extracted_text = texts[0].description if texts else ""
# Step 4: Prepare the prompt for Gemini
prompt = f'''
#Direction
You are an employee of finance department of a company. For each task well done, you will be rewarded $1,000 bonus.
The boss has come to your office and gave you a task of transfering the content of receipts and business cards onto a google spreadsheet.
He hands you two things.
One, is a paper containing guideline and example.
Two, is an OCR text.
#Guideline
-strictly adhere to the given content of OCR text.
-it could be of receipt(s) or business card(s) but not both
-sort any other info that couldn't be classified as "기타"
-format the extracted information as a JSON object with two main keys: 'receipts' and 'busi_cards'.
-ensure the total number of receipts and business cards combined does not exceed 500.
#Example
##Receipts
Input
은성마켓
사업자번호: 727-27-01455
대표자:허유성
주
소: 경기 시흥시 배곧3로 96
전화 번호: 031-431-1544
판매일: 24-02-23 20:10, 금요일 계산대:001
NO. 상품명
단가 수량 금액
001 느타리버섯
200092
1,000
1. 1,000 #
002 양배추
200071
(#)면세물품:
합
계:
신용카드지불:
★카드신용승인
火
1,800 1 1.800 #
2,800
2,800
2,800
가맹점명 은성마켓
카드번호
4673-09**-****-515*
카드명 : KB국민체크
전표매입: KB국민카드 무서명: 00111521584
승인금액: 2,800
(일시불)
승인번호: 30014507, 전표No:201036
제출 VAN: JTNET-리더기
정상승인
감사합니다.
거래NO:0223007835 계산원: 관리자(001)
2502230078357
Output
발행일: 2024/02/23 20:10
상호: 은성마켓
사업자번호: 727-27-01455
합계금액: 2,800
내역:
느타리버섯 1,000원 1개
양배추 1,800원 1개
카드번호: 4673-09**-****-515*
카드종류: KB국민체크
승인번호: 30014507
기타: 대표자: 허유성, 주소: 경기 시흥시 배곧3로 96, 전화 번호: 031-431-1544, 면세물품: 2,800, 합계: 2,800, 신용카드지불:2,800, 전표No:201036, 제출 VAN: JTNET-리더기, 거래NO:0223007835 계산원: 관리자(001)
##Business Cards
Input
박장원
아이프리마
팀장/정책사업팀
R&D사업실
Over the MIRACLE
경기도 화성시 정남면 만년로 98번길 55 10층 (우) 18523
Mobile 010-9582-0925
Tel (031) 000-0000
Fax (031) 000-0000
Email iprima@iprima.com
www.iprima.com
Output
이름: 박장원
회사명: 아이프리마
소속 및 직위: 팀장/정책사업팀 R&D사업실
주소: 경기도 화성시 정남면 만년로 98번길 55 (우편번호) 18523
전화: (031) 000-0000
휴대폰: (031) 000-0000
팩스: 010-0000-0000
e-mail: iprima@gmail.com
홈페이지: www.iprima.com
기타: Over the MIRACLE
Image content: {extracted_text}
'''
logging.info("Prompt prepared for Gemini model: " + prompt)
# Generate content using the Gemini model
response = model.generate_content(
contents=[
prompt
]
)
# Check if the response has text content
if response.text:
logging.info("Parsing Gemini model response")
json_response = json.loads(response.text)
logging.info(f"JSON response parsed: {response.text}")
return json.dumps(json_response)
else:
logging.error("Gemini model did not return a text response")
raise Exception("Gemini model did not return a text response")
except Exception as e:
logging.error(f"Error in request_gpt4o_completion: {str(e)}")
raise HTTPException(status_code=500, detail=f"Failure: {str(e)}")
async def move_file_to_folder(file_id, current_parents, new_parents):
try:
# Ensure current_parents is a list of parent IDs
if not isinstance(current_parents, list):
current_parents = [current_parents]
if not isinstance(new_parents, list):
new_parents = [new_parents]
# Retrieve the existing parents to remove
file = drive_service.files().get(fileId=file_id, fields='parents').execute()
existing_parents = file.get('parents')
logging.info(f"Existing parents of the file {file_id}: {existing_parents}")
# Remove only the current parents
parents_to_remove = [parent for parent in current_parents if parent in existing_parents]
parents_to_add = [parent for parent in new_parents if parent not in existing_parents]
previous_parents = ",".join(parents_to_remove)
logging.info(f"Parents to remove: {parents_to_remove}")
logging.info(f"Parents to add: {parents_to_add}")
# Move the file to the new folders while keeping it in yungsoogi folder
if parents_to_add:
file = drive_service.files().update(
fileId=file_id,
addParents=",".join(parents_to_add),
removeParents=previous_parents,
fields='id, parents'
).execute()
logging.info(f"File {file_id} moved to new folders {new_parents}.")
else:
logging.info(f"File {file_id} already in the target folders {new_parents}.")
except Exception as e:
logging.error(f"Failed to move file {file_id} to new folders: {e}")
async def update_user_sheet(spreadsheet_id:str, sheet_name:str, data: List[List[str]], is_reset: Optional[bool] = False):
print(f"Spreadsheet ID: {spreadsheet_id}, Sheet Name: {sheet_name}, Data: {data}")
# Get all the sheets in the spreadsheet
sheet_metadata = sheet_service.spreadsheets().get(spreadsheetId=spreadsheet_id).execute()
sheets = sheet_metadata.get('sheets', '')
# Find the sheet by name
sheet = next((sheet for sheet in sheets if sheet['properties']['title'] == sheet_name), None)
if not sheet:
raise ValueError(f"Sheet '{sheet_name}' not found in the spreadsheet.")
sheet_id = sheet['properties']['sheetId']
# Rest of the function remains the same
no_of_rows = len(data)
no_of_cols = len(data[0])
if is_reset:
new_sheet_name = f"{sheet_name} {datetime.now(kst).strftime('%Y-%m-%d %H:%M')}"
duplicate_requests = [{
"duplicateSheet": {
"sourceSheetId": sheet_id,
"newSheetName": new_sheet_name
}
}]
duplicate_response = sheet_service.spreadsheets().batchUpdate(
spreadsheetId=spreadsheet_id,
body={
'requests': duplicate_requests
}
).execute()
sheet_id = duplicate_response['replies'][0]['duplicateSheet']['properties']['sheetId']
sheet_name = new_sheet_name
# Insert new rows
requests = [
{
"updateSheetProperties": {
"properties": {
"sheetId": sheet_id,
"hidden": False
},
"fields": "hidden"
}
},
{
"insertDimension": {
"range": {
"sheetId": sheet_id,
"dimension": "ROWS",
"startIndex": 1,
"endIndex": no_of_rows + 1
},
"inheritFromBefore": False,
}
}]
response = sheet_service.spreadsheets().batchUpdate(
spreadsheetId=spreadsheet_id,
body={
'requests': requests
}
).execute()
update_body = {
'values': data
}
start_column = 'A'
end_column = chr(ord(start_column) + no_of_cols - 1)
range_to_update = f'{sheet_name}!{start_column}2:{end_column}{no_of_rows + 1}'
result = sheet_service.spreadsheets().values().update(
spreadsheetId=spreadsheet_id,
range=range_to_update,
valueInputOption='USER_ENTERED',
body=update_body
).execute()
# Format the cells
for row_index, row in enumerate(data):
for col_index, cell_value in enumerate(row):
if cell_value.startswith('=HYPERLINK'):
color = {"red": 0, "green": 0, "blue": 1} # Blue for hyperlinks
else:
color = {"red": 0, "green": 0, "blue": 0} # Black for other text
requests = [
{
"repeatCell": {
"range": {
"sheetId": sheet_id,
"startRowIndex": row_index + 1,
"endRowIndex": row_index + 2,
"startColumnIndex": col_index,
"endColumnIndex": col_index + 1
},
"cell": {
"userEnteredFormat": {
"textFormat": {
"foregroundColor": color
}
}
},
"fields": "userEnteredFormat.textFormat.foregroundColor"
}
}
]
response = sheet_service.spreadsheets().batchUpdate(
spreadsheetId=spreadsheet_id,
body={
'requests': requests
}
).execute()
async def copy_sms_sheet(phone_no: str, spreadsheet_id: str):
requests = [
{
"duplicateSheet": {
"sourceSheetId": sms_ss,
"newSheetName": 'SMS ' + phone_no
}
}
]
response = sheet_service.spreadsheets().batchUpdate(
spreadsheetId=spreadsheet_id,
body={
'requests': requests
}
).execute()
print(response)
new_sheet_id = response['replies'][0]['duplicateSheet']['properties']['sheetId']
unhide_request = [{
"updateSheetProperties": {
"properties": {
"sheetId": new_sheet_id,
"hidden": False
},
"fields": "hidden"
}
}]
sheet_service.spreadsheets().batchUpdate(
spreadsheetId=spreadsheet_id,
body={
'requests': unhide_request
}
).execute()
return new_sheet_id
async def update_fb_transaction(transaction_ref: firestore.DocumentReference, no_sms: Optional[int] = 0, no_contacts: Optional[int] = 0, no_receipts: Optional[int] = 0, no_business_cards: Optional[int] = 0):
try:
# Update the transaction
transaction_ref.update({
"no_sms": firestore.Increment(no_sms),
"no_contacts": firestore.Increment(no_contacts),
"no_receipts": firestore.Increment(no_receipts),
"no_business_cards": firestore.Increment(no_business_cards)
})
return {"status": "Success"}
except Exception as e:
logging.error(f"Error updating transaction: {e}")
return {"status": f"Error: {str(e)}"}
@app.post("/check-userdoc")
async def check_user_doc(user_id: str = Form(...)):
user_ref = users_collection_ref.document(user_id)
user = user_ref.get()
if user.exists:
return {"dbUserExists": True}
transaction_ref = transactions_collection_ref.document(user_id)
transaction = transaction_ref.get()
if not transaction.exists:
transaction_ref.set({
"no_sms": 0,
"no_contacts": 0,
"no_receipts": 0,
"no_business_cards": 0,
"purchased_credit": {
"image_detection": 100,
"sync_data": 500,
},
"uid": user_id,
})
return {"dbUserExists": False}
template_sheet_id = '1i5mrmlTs5sPWtx2mBtc_f-zm1D3x21r1T_77guki8_8'
@app.post("/edit-spreadsheet")
async def edit_spreadsheet(target_sheet_id: str = Form(...)):
try:
# Retrieve the target spreadsheet's metadata
target_spreadsheet = sheet_service.spreadsheets().get(spreadsheetId=target_sheet_id).execute()
# Collect all sheet IDs in the target spreadsheet
target_sheet_ids = [sheet['properties']['sheetId'] for sheet in target_spreadsheet.get('sheets', [])]
# Prepare batch update request to delete existing sheets
requests = []
for sheet_id in target_sheet_ids:
requests.append({
"deleteSheet": {
"sheetId": sheet_id
}
})
# Retrieve the template spreadsheet's metadata
template_spreadsheet = sheet_service.spreadsheets().get(spreadsheetId=template_sheet_id).execute()
# Iterate through each sheet in the template spreadsheet
for sheet in template_spreadsheet.get('sheets', []):
sheet_id = sheet['properties']['sheetId']
# Use the copyTo method to copy the sheet to the target spreadsheet
copy_request = sheet_service.spreadsheets().sheets().copyTo(
spreadsheetId=template_sheet_id,
sheetId=sheet_id,
body={"destinationSpreadsheetId": target_sheet_id}
)
copy_response = copy_request.execute()
# Get the copied sheet ID
copied_sheet_id = copy_response['sheetId']
# Rename the copied sheet to remove "Copy of"
copied_sheet_title = sheet['properties']['title'].replace("Copy of ", "")
rename_request = {
"updateSheetProperties": {
"properties": {
"sheetId": copied_sheet_id,
"title": copied_sheet_title
},
"fields": "title"
}
}
# Execute rename request
sheet_service.spreadsheets().batchUpdate(
spreadsheetId=target_sheet_id,
body={"requests": [rename_request]}
).execute()
logging.info(f"Sheet {sheet['properties']['title']} copied and renamed to {copied_sheet_title} in {target_sheet_id}.")
if requests:
# Execute batch update to delete sheets
batch_update_request = {"requests": requests}
sheet_service.spreadsheets().batchUpdate(
spreadsheetId=target_sheet_id,
body=batch_update_request
).execute()
logging.info(f"Deleted existing sheets in target spreadsheet {target_sheet_id}.")
logging.info(f"All sheets from template {template_sheet_id} copied to {target_sheet_id}.")
return {"message": "Spreadsheet copied successfully"}
except Exception as e:
logging.error(f"Failed to copy sheets: {e}")
return {"error": str(e)}
async def create_shortcut(file_id, parent_folder_id, name):
try:
file_metadata = {
'name': name,
'mimeType': 'application/vnd.google-apps.shortcut',
'parents': [parent_folder_id],
'shortcutDetails': {
'targetId': file_id
}
}
shortcut = drive_service.files().create(body=file_metadata, fields='id').execute()
return shortcut.get('id')
except Exception as e:
logging.error(f"Failed to create shortcut: {e}")
return None
async def update_user_sheets_and_folders(user_data, transaction_ref, transaction_data, image_id):
try:
parent_folders = [user_data['gDrive_metadata']['yungsoogi'], user_data['gDrive_metadata']['uploaded']]
dataJson = await request_gpt4o_completion(image_id, transaction_data['purchased_credit']['image_detection'])
if dataJson is None:
return {"error": "An error occurred while processing the image"}
data = json.loads(dataJson)
# Check if 'receipts' key exists and the length
found_receipt_no = len(data['receipts']) if 'receipts' in data else 0
found_business_cards_no = len(data['busi_cards']) if 'busi_cards' in data else 0
new_folder = None
secondary_folder = None
# timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
timestamp = datetime.now(kst).strftime("%Y-%m-%d %H:%M:%S")
# https://drive.google.com/file/d/1cFIA09PBqFjM8YvIz1D60p7HMGk1uBvo/view?usp=sharing
download_image_url = f"https://drive.google.com/uc?id={image_id}&export=download"
image_url = f"https://drive.google.com/file/d/{image_id}/view?usp=sharing"
# 이미지를 보여주어야 할때 =image({image_url})를 사용하고 갈 수 있는 경로를 만들때는 =HYPERLINK({image_url}, '바로가기')를 사용하세요.
# Determine the primary folder to move the image to and secondary folder for shortcut
if found_receipt_no > 0:
new_folder = user_data['gDrive_metadata']['receipts']
receipts_data = convert_dicts_to_list(
data['receipts'],
receipt_sheet_headers,
add_link=True,
image_id=image_id,
link_text="바로가기",
link_position=1 # Link third
)
print(receipts_data)
await update_user_sheet(user_data['gDrive_metadata']['spreadsheet'], "영수증", receipts_data)
if found_business_cards_no > 0:
if new_folder is None:
new_folder = user_data['gDrive_metadata']['business_cards']
else:
secondary_folder = user_data['gDrive_metadata']['business_cards']
business_cards_data = convert_dicts_to_list(
data['busi_cards'],
business_card_sheet_headers,
add_link=True,
image_id=image_id,
link_text="명함보기",
link_position=-1 # Link at the end
)
await update_user_sheet(user_data['gDrive_metadata']['spreadsheet'], "명함", business_cards_data)
print(f"{found_receipt_no}, {found_business_cards_no}")
status = await update_fb_transaction(
transaction_ref,
no_receipts=found_receipt_no,
no_business_cards=found_business_cards_no
)
if status['status'] == "Success":
if new_folder:
await move_file_to_folder(image_id, parent_folders, [new_folder])
if secondary_folder:
await create_shortcut(image_id, secondary_folder, f"Shortcut to {image_id}")
return {"success": True}
else:
logging.error(f"Transaction update failed with status: {status['status']}")
return {"error": "An error has occurred while updating the transaction"}
except Exception as e:
logging.error(f"An error occurred: {e}")
# Move the image file to the error folder
await move_file_to_folder(image_id, parent_folders, [user_data['gDrive_metadata']['error']])
return {"error": str(e)}
@app.post("/process-image")
async def process_photo(background_tasks: BackgroundTasks, image_id: str = Form(...), user_id: str = Form(...)):
transaction_ref = transactions_collection_ref.document(user_id)
transaction = transaction_ref.get()
if not transaction.exists:
transaction_ref.set({
"no_sms": 0,
"no_contacts": 0,
"no_receipts": 0,
"no_business_cards": 0,
"purchased_credit": {
"image_detection": 100,
"sync_data": 500,
}
})
transaction = transaction_ref.get()
transaction_data = transaction.to_dict()
if transaction_data['no_receipts'] + transaction_data['no_business_cards'] >= transaction_data['purchased_credit']['image_detection']:
return {"error": "You have reached the limit of the number of items you can process. Please upgrade your plan."}
user_ref = users_collection_ref.document(user_id)
user = user_ref.get()
user_data = user.to_dict()
background_tasks.add_task(update_user_sheets_and_folders, user_data, transaction_ref, transaction_data, image_id)
return {"success": True}
@app.post("/sync-data")
async def process_photo(data_json: str = Form(...), user_id: str = Form(...)):
# print(data_json)
data: Dict[str, Any] = json.loads(data_json)
transaction_ref = transactions_collection_ref.document(user_id)
transaction = transaction_ref.get()
if not transaction.exists:
transaction_ref.set({
"no_sms": 0,
"no_contacts": 0,
"no_receipts": 0,
"no_business_cards": 0,
"purchased_credit": {
"image_detection": 100,
"sync_data": 500,
}
})
transaction = transaction_ref.get()
transaction_data = transaction.to_dict()
if transaction_data['no_contacts'] + transaction_data['no_sms'] >= transaction_data['purchased_credit']['sync_data']:
return {"error": "You have reached the limit of the number of items you can process. Please upgrade your plan."}
user_ref = users_collection_ref.document(user_id)
user = user_ref.get()
user_data = user.to_dict()
try:
if data['isContacts']:
await update_user_sheet(user_data['gDrive_metadata']['spreadsheet'], contacts_ss, "연락처", data['data'], True)
transaction_data['no_contacts'] += len(data['data'])
transaction_ref.update({'no_contacts': transaction_data['no_contacts']})
else:
if 'sms_ids' not in user_data:
print('No sms_ids')
user_data['sms_ids'] = {}
existing_phones = set(user_data.get('sms_ids', {}).keys())
new_phones = [phone['address'] for phone in data['data'] if phone['address'] not in existing_phones]
print(f"Existing phones: {existing_phones}")
for phone_address in new_phones:
new_sheet_id = await copy_sms_sheet(phone_address, user_data['gDrive_metadata']['spreadsheet'])
print(f"New sheet id: {new_sheet_id}")
user_data['sms_ids'][phone_address] = {'sheet_id': new_sheet_id, 'last_updated': "1970-01-01T00:00:00.000"}
print(user_data['sms_ids'])
updated_sms_count = 0
for phone in data['data']:
last_updated = user_data['sms_ids'][phone['address']]['last_updated']
print(f"Last updated: {last_updated}")
# Convert last_updated to datetime for comparison
last_updated_dt = datetime.fromisoformat(last_updated)
update_message = [sms for sms in phone['sms'] if datetime.fromisoformat(sms[0]) > last_updated_dt]
print(f"Update message: {update_message}")
if not update_message:
continue
sheet_id = user_data['sms_ids'][phone['address']]['sheet_id']
print(f"Sheet ID: {sheet_id}")
await update_user_sheet(user_data['gDrive_metadata']['spreadsheet'], sheet_id, 'SMS ' + phone['address'], update_message, False)
user_data['sms_ids'][phone['address']]['last_updated'] = phone['sms'][0][0]
updated_sms_count += len(update_message)
user_ref.update({'sms_ids': user_data['sms_ids']})
transaction_data['no_sms'] += updated_sms_count
transaction_ref.update({'no_sms': transaction_data['no_sms']})
return {"success": True}
except Exception as e:
logging.error(f"An error occurred: {e}")
return {"error": str(e)}
class DriveWatcher:
def __init__(self):
self.watch_list: Dict[str, str] = {}
self.channels: Dict[str, Dict[str, Any]] = {}
# def create_channel(self, folder_id: str) -> Dict[str, Any]:
# try:
# body = {
# 'id': f'channel-{folder_id}-{int(time.time())}',
# 'type': 'web_hook',
# 'address': 'https://m8chaa-re-sheet-develop.hf.space/notifications',
# }
# # 'expiration': int((datetime.now() + timedelta(days=7)).timestamp() * 1000) # Set expiration to 7 days
# logging.info(body)
# response = app.state.drive_service.files().watch(fileId=folder_id, body=body).execute()
# logging.info(f"Created channel for folder {folder_id}: {response['id']}")
# return response
# except HttpError as e:
# logging.error(f"HttpError creating channel for folder {folder_id}: {e}")
# raise
# except Exception as e:
# logging.error(f"Unexpected error creating channel for folder {folder_id}: {e}")
# raise
def create_channel(self, folder_id: str) -> Dict[str, Any]:
try:
body = {
'id': f'channel-{folder_id}-{int(time.time())}',
'type': 'web_hook',
'address': 'https://m8chaa-re-sheet-develop.hf.space/notifications',
}
logging.info(f"Attempting to create channel for folder {folder_id} with body: {body}")
response = app.state.drive_service.files().watch(fileId=folder_id, body=body).execute()
logging.info(f"Created channel for folder {folder_id}: {response['id']}")
return response
except HttpError as e:
logging.error(f"HttpError creating channel for folder {folder_id}: {e.content}")
raise
except Exception as e:
logging.error(f"Unexpected error creating channel for folder {folder_id}: {str(e)}")
raise
def close_channel(self, channel_id: str, resource_id: str):
try:
app.state.drive_service.channels().stop(body={
'id': channel_id,
'resourceId': resource_id,
}).execute()
logging.info(f"Closed channel: {channel_id}")
except HttpError as e:
logging.error(f"HttpError closing channel {channel_id}: {e}")
except Exception as e:
logging.error(f"Unexpected error closing channel {channel_id}: {e}")
def update_watch_list(self):
current_time = int(time.time())
for channel_id, channel_info in list(self.channels.items()):
if current_time >= channel_info['expiration'] // 1000:
self.close_channel(channel_id, channel_info['resourceId'])
del self.channels[channel_id]
for user_id, folder_id in self.watch_list.items():
if not any(info['folder_id'] == folder_id for info in self.channels.values()):
try:
response = self.create_channel(folder_id)
self.channels[response['id']] = {
'resourceId': response['resourceId'],
'user_id': user_id,
'folder_id': folder_id,
'expiration': int(response['expiration'])
}
logging.info(f"Created new channel for user {user_id}, folder {folder_id}")
except Exception as e:
logging.error(f"Error creating channel for user {user_id}: {str(e)}")
def on_user_change(self, snapshot, changes, read_time):
update_required = False
for change in changes:
user_id = change.document.id
user_data = change.document.to_dict()
if change.type.name in ['ADDED', 'MODIFIED']:
if 'gDrive_metadata' in user_data and 'uploaded' in user_data['gDrive_metadata']:
folder_id = user_data['gDrive_metadata']['uploaded']
if user_id not in self.watch_list or self.watch_list[user_id] != folder_id:
self.watch_list[user_id] = folder_id
update_required = True
elif change.type.name == 'REMOVED':
if user_id in self.watch_list:
del self.watch_list[user_id]
update_required = True
if update_required:
self.update_watch_list()
# @app.on_event("startup")
# async def startup_event():
# app.state.drive_watcher = DriveWatcher()
# app.state.drive_service = build('drive', 'v3', credentials=APICredential)
# # Fetch all users and initialize the watch list
# # users = users_collection_ref.get()
# # for user in users:
# # user_data = user.to_dict()
# # if 'gDrive_metadata' in user_data and 'uploaded' in user_data['gDrive_metadata']:
# # app.state.drive_watcher.watch_list[user.id] = user_data['gDrive_metadata']['uploaded']
# app.state.drive_watcher.watch_list["xJs8nDUdbVWvq1QIKY3VYmFkidn1"] = "1paIWJDmlLyWaGsB_eXrpP7ILZiw3Xb70"
# # Initial update of watch list channels
# # app.state.drive_watcher.update_watch_list()
# # Set up the Firebase listener
# # app.state.users_watch = users_collection_ref.on_snapshot(app.state.drive_watcher.on_user_change)
# logging.info("Application startup completed")
@app.on_event("startup")
async def startup_event():
app.state.drive_watcher = DriveWatcher()
app.state.drive_service = build('drive', 'v3', credentials=APICredential)
# Fetch all users and initialize the watch list
app.state.drive_watcher.watch_list["xJs8nDUdbVWvq1QIKY3VYmFkidn1"] = "1paIWJDmlLyWaGsB_eXrpP7ILZiw3Xb70"
# Initial update of watch list channels
app.state.drive_watcher.update_watch_list()
logging.info("Application startup completed")
@app.on_event("shutdown")
async def shutdown_event():
# Close all channels
for channel_id, channel_info in app.state.drive_watcher.channels.items():
await app.state.drive_watcher.close_channel(channel_id, channel_info['resourceId'])
# Stop the Firebase listener
app.state.users_watch.unsubscribe()
logging.info("Application shutdown completed")
@app.post("/notifications")
async def notifications(request: Request, background_tasks: BackgroundTasks):
try:
channel_id = request.headers.get('X-Goog-Channel-ID')
resource_id = request.headers.get('X-Goog-Resource-ID')
resource_state = request.headers.get('X-Goog-Resource-State')
resource_uri = request.headers.get('X-Goog-Resource-URI')
changed = request.headers.get('X-Goog-Changed')
logging.info(f"Notification received: Channel ID: {channel_id}, Resource ID: {resource_id}, State: {resource_state}, URI: {resource_uri}, Changed: {changed}")
if channel_id in app.state.drive_watcher.channels:
logging.info("channel id is in the drive watcher")
channel_info = app.state.drive_watcher.channels[channel_id]
user_id = channel_info['user_id']
folder_id = channel_info['folder_id']
logging.info(f"Notification is for user: {user_id}, folder: {folder_id}")
if resource_state == 'update' and changed == 'children':
logging.info("appropriate resource changed")
try:
# List files in the folder that have been recently modified
results = app.state.drive_service.files().list(
q=f"'{folder_id}' in parents and mimeType contains 'image/' and trashed=false",
fields="files(id, name, mimeType, modifiedTime)",
orderBy="modifiedTime desc"
).execute()
files = results.get('files', [])
for file in files:
if file['mimeType'].startswith('image/'):
logging.info(f"Processing new image: {file['id']}")
await process_photo(user_id, file['id'], background_tasks)
except HttpError as e:
logging.error(f"HttpError processing folder changes: {e}")
except Exception as e:
logging.error(f"Unexpected error processing folder changes: {e}")
else:
logging.warning(f"Received notification for unknown channel: {channel_id}")
background_tasks.add_task(stop_channel, channel_id, resource_id)
return {"message": "Notification processed successfully"}
except Exception as e:
logging.error(f"Error processing notification: {str(e)}")
raise HTTPException(status_code=500, detail=str(e))
def stop_channel(channel_id, resource_id):
request_body = {
'id': channel_id,
'resourceId': resource_id
}
try:
app.state.drive_service.channels().stop(body=request_body).execute()
print(f"Channel {channel_id} stopped successfully.")
except Exception as e:
print(f"An error occurred: {e}")
from fastapi import FastAPI, HTTPException, UploadFile, File, Request, BackgroundTasks, Form
from fastapi.middleware.cors import CORSMiddleware
from pydantic import BaseModel
from httpx import Timeout
import httpx
import logging
import base64
import firebase_admin
from firebase_admin import credentials
from firebase_admin import firestore
import json
from datetime import datetime, timedelta
import pytz
from typing import List, Optional, Dict
from google.oauth2 import service_account
from googleapiclient.discovery import build
from google.oauth2.credentials import Credentials
from google_auth_oauthlib.flow import InstalledAppFlow
from google.auth.transport.requests import Request as GoogleRequest
from googleapiclient.discovery import build
import os
import json
from io import BytesIO
import google.generativeai as genai
import requests
from google.cloud import vision
import aiohttp
from typing import Tuple
import time
from apscheduler.schedulers.background import BackgroundScheduler
from google.cloud.firestore import AsyncClient
import asyncio
from googleapiclient.errors import HttpError
from typing import Dict, Any
# start application
app = FastAPI()
# allow all origins
app.add_middleware(
CORSMiddleware,
allow_origins=["*"], # Allows all origins
allow_credentials=True, # Allows credentials (such as cookies) to be sent with requests
allow_methods=["*"], # Allows all methods
allow_headers=["*"], # Allows all headers
)
# set up logging
logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
logging.info("Application starting up")
request_timeout = Timeout(1000.0, connect=600.0)
# set up firebase
# Load Firebase configuration from a file
with open('/app/secrets/fb_secret.json', 'r') as f:
firebase_configuration = json.load(f)
firebase_credentials = credentials.Certificate(firebase_configuration)
firebase_admin.initialize_app(firebase_credentials, {
'projectId': firebase_configuration['project_id'],
})
firestore_db = firestore.client()
users_collection_ref = firestore_db.collection('users')
transactions_collection_ref = firestore_db.collection('transactions')
SCOPES = ['https://www.googleapis.com/auth/spreadsheets','https://www.googleapis.com/auth/drive']
with open('/app/secrets/openai_key.txt', 'r') as f:
OPENAI_KEY = f.read()
APICredential = service_account.Credentials.from_service_account_file(
'/app/secrets/googleapi_secret.json', scopes=SCOPES)
sheet_service = build('sheets', 'v4', credentials=APICredential)
drive_service = build('drive', 'v3', credentials=APICredential)
with open('/app/secrets/gemini_key.txt', 'r') as f:
api_key = f.read()
genai.configure(api_key=api_key)
# Create the model
generation_config = {
"temperature": 1,
"top_p": 0.95,
"top_k": 64,