GraphQL integrate, data correctness

find_auction_with_lots.py

@@ -0,0 +1,48 @@
+#!/usr/bin/env python3
+"""Find an auction page with lots data"""
+import sqlite3
+import zlib
+import json
+import re
+
+conn = sqlite3.connect('/mnt/okcomputer/output/cache.db')
+
+cursor = conn.execute("""
+    SELECT url, content
+    FROM cache
+    WHERE url LIKE '%/a/%'
+""")
+
+for row in cursor:
+    url, content_blob = row
+    content = zlib.decompress(content_blob).decode('utf-8')
+
+    match = re.search(r'<script[^>]*id="__NEXT_DATA__"[^>]*>(.+?)</script>', content, re.DOTALL)
+    if not match:
+        continue
+
+    data = json.loads(match.group(1))
+    page_props = data.get('props', {}).get('pageProps', {})
+
+    if 'auction' in page_props:
+        auction = page_props['auction']
+        lots = auction.get('lots', [])
+
+        if lots and len(lots) > 0:
+            print(f"Found auction with {len(lots)} lots: {url}\n")
+
+            lot = lots[0]
+            print(f"SAMPLE LOT FROM AUCTION.LOTS[]:")
+            print(f"  displayId: {lot.get('displayId')}")
+            print(f"  title: {lot.get('title', '')[:50]}...")
+            print(f"  urlSlug: {lot.get('urlSlug')}")
+            print(f"\nBIDDING FIELDS:")
+            for key in ['currentBid', 'highestBid', 'startingBid', 'minimumBidAmount', 'bidCount', 'numberOfBids']:
+                print(f"  {key}: {lot.get(key)}")
+            print(f"\nTIMING FIELDS:")
+            for key in ['endDate', 'startDate', 'closingTime']:
+                print(f"  {key}: {lot.get(key)}")
+            print(f"\nALL KEYS: {list(lot.keys())[:30]}...")
+            break
+
+conn.close()