Add a full-text search index to a Convex table and run a ranked search query with filter fields

Documented steps

1. In convex/schema.ts, chain .searchIndex("search_body", { searchField: "body", filterFields: ["channel"], staged: false }) onto the target table's defineTable(...) call.
2. Ensure searchField is a single field of type v.string(); filterFields is optional, supports any type, and accepts up to 16 fields.
3. Set staged: true to backfill the index asynchronously without blocking the deploy, or false to build it as part of the deploy.
4. Deploy the schema so the search index is built.
5. In a query function run ctx.db.query("messages").withSearchIndex("search_body", (q) => q.search("body", "hello hi").eq("channel", "#general")).take(10).
6. A search filter expression must contain exactly one .search(fieldName, queryString) call, optionally followed by one or more .eq(fieldName, value) calls on declared filterFields.
7. Retrieve results with .collect(), .take(n), .first(), .unique(), or .paginate(paginationOpts); results always come back in relevance order.
8. Add a post-index .filter(...) call only for extra conditions beyond the indexed filter fields, accepting that it does not reduce the scanned set.
9. Official docs: https://docs.convex.dev/search/text-search