Translate a plain-English ecological question into a ranked list of candidate NEON data product IDs, with DOIs, release, theme, and per-site/per-year availability. Use whenever the agent needs to turn a research question into concrete product IDs before fetching anything.
The Librarian's primary data-finding skill. You do not guess product IDs — you look them up against the live NEON API or the bundled product index.
neon-ingestion-stacker.scripts/search_products.py to hit the NEON products endpoint and return matches with DOIs and site availability.references/product_index.md (curated shortlist of the most-requested products).scripts/rank_candidates.py to order matches by relevance, temporal coverage, and site overlap.productID (e.g. DP1.10022.001)theme (e.g. "Organisms, Populations, Communities")doi (from the current release)release (default RELEASE-2026)site_coverage: list of site codes with (start, end) tuplescadence (bout / subhourly / annual / campaign)required_docs: protocol + user guide + ATBD IDs to hand to documentation-interpreterscripts/search_products.py — API-first search with markdown fallback.scripts/rank_candidates.py — deterministic ranking by overlap + recency.references/product_index.md — curated shortlist of high-traffic NEON products.references/themes.md — the NEON theme taxonomy.open_questions entry asking the user to clarify. Do not hallucinate a product ID.→ documentation-interpreter (to read the protocols for the top candidates)
→ metadata-interpretation (to decode variables file once downloaded)
→ coverage-planner (to confirm site × time overlap before fetching)