Spaces:
Running
on
Zero
Running
on
Zero
File size: 1,645 Bytes
821310d |
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 |
import os
import openai
import firebase_admin
from firebase_admin import firestore
from firebase_setup import init_firebase
from utils.coin_tools import get_coin_data, save_coin_data
from utils.image_tools import compare_images
# Init Firebase
init_firebase()
db = firestore.client()
# Init OpenAI
openai.api_key = os.getenv("OPENAI_API_KEY")
# === USER INPUT (Simulated for now) ===
user_image_path = "images/user_coin.jpg"
suspected_coin_name = "1955 Lincoln Cent DDO-001"
# === Step 1: Check if coin already exists in Firebase ===
coin_doc = db.collection("coins").document(suspected_coin_name.replace(" ", "_")).get()
if coin_doc.exists:
print(f"β
Found {suspected_coin_name} in the database. Comparing now...")
ref_images = coin_doc.to_dict().get("reference_images", [])
result = compare_images(user_image_path, ref_images)
print(f"Match Confidence: {result['confidence']*100:.1f}%")
else:
print(f"π No match found in database. Performing web search...")
# === Step 2: Use web_search_preview tool to find data ===
response = openai.responses.create(
model="gpt-4.1",
tools=[{"type": "web_search_preview"}],
input=f"Find die marker info, image links, and estimated value for {suspected_coin_name}"
)
output_text = response.output[1]['content'][0]['text']
print("π AI Search Result:\n", output_text)
# === Step 3: Save AI result to Firebase ===
coin_data = get_coin_data(suspected_coin_name, output_text, user_image_path)
save_coin_data(suspected_coin_name, coin_data)
print(f"πΎ Saved new coin info to Firebase under {suspected_coin_name}.")
|