# load.py against the chunks.jsonl made by build_chunks.py
304 chunks, model=nomic-embed-text, dim=768
chunks with an empty heading: 4
max batch size: 5461
embedding function on collection: None
upserted in 0.23s, batch size 128
count: 304
19 sections, total 304
  api                  129
  sdk                  59
  structured-outputs   18
  guides               17
  errors               16
  crawling-agent       10
  feeds                10
  async-requests       8
  getting-started      6
  markdown-cleaning    5
  pricing              5
  crawling-types       4
  job                  4
  root                 3
  scrape               3
  access-key           2
  integrations         2
  rate-limits          2
  sdks                 1
.venv/bin/python load.py  0.49s user 0.10s system 24% cpu 2.389 total

# query.py
vector norms: min 0.9999994 max 1.0000006

Q: How do I get notified when a crawl job finishes?
  0.255  api             /docs/api/cancel  [Cancel job]
  0.268  sdk             /docs/sdk/go  [Go > Response Objects > Job]
  0.271  async-requests  /docs/async-requests  [Async and Webhooks > Using API calls]

Q: How long are scrape results cached and how do I disable the cache?
  0.186  api             /docs/api/caching  [Caching]
  0.213  api             /docs/api/caching  [Caching > Cache Parameter > Values]
  0.220  api             /docs/api/caching  [Caching > How Caching Works]

smoke test, 6 questions, expected URL in top 5:
  rank 1  d=0.244  How do I limit how many pages a crawl fetches?
  rank 3  d=0.255  How do I get notified when a crawl job finishes?
  rank 1  d=0.186  How long are scrape results cached and how do I disable the cache?
  rank 1  d=0.186  How do I cancel a running crawl job?
  rank 1  d=0.174  How do I check my organization's remaining balance through the API?
  rank 1  d=0.110  How do I use WebCrawlerAPI from an MCP client like Claude Desktop?
ranks: [1, 3, 1, 1, 1, 1] | hit in top 5: 6 / 6