Postgres full-text search over counselling reports, ranked by relevance. Filter by school, grade, or report kind.
Kabir Khan:** Honestly, more pressure now. Dad keeps saying CA and Commerce is the safe path. Mom is more
Kabir Khan. Top RIASEC dimensions: Realistic (44), Investigative (88), Artistic (93). The pattern suggests strong fit for analytical
Kabir Khan based on assessment battery + 3 counselling sessions. Top pathway: architecture (consistent across RIASEC, DAT V, and self
Kabir Khan based on assessment battery + 3 counselling sessions. Top pathway: medicine (consistent across RIASEC, DAT V, and self
Kabir Khan was engaged today. We explored their interest in creative writing, discussing pathway options at the intersection
Every reports.text and counsellor_note is indexed by Postgres tsvector (english config). One GIN index, refreshed on insert.
Search box runs plainto_tsquery + ts_rank_cd. Filter chips (school, grade, RIASEC, kind) compose into the same SQL — no client-side filtering.
ts_headline wraps matches in <mark> tags so counsellors see the evidence in context.
Top 30 sorted by rank × recency. The aim is < 200 ms server-side on the full corpus, no Elasticsearch needed.