search implement
This commit is contained in:
Priec
@@ -42,6 +42,8 @@ mod m20260621_000004_add_business_sale_price_to_products;
|
||||
mod m20260622_000001_audience_discount_profiles;
|
||||
mod m20260622_000002_product_variants;
|
||||
mod m20260622_000003_variant_stock_nullable;
|
||||
mod m20260622_000004_product_search;
|
||||
mod m20260622_000005_product_search_aggregate;
|
||||
pub struct Migrator;
|
||||
|
||||
#[async_trait::async_trait]
|
||||
@@ -88,6 +90,8 @@ impl MigratorTrait for Migrator {
|
||||
Box::new(m20260622_000001_audience_discount_profiles::Migration),
|
||||
Box::new(m20260622_000002_product_variants::Migration),
|
||||
Box::new(m20260622_000003_variant_stock_nullable::Migration),
|
||||
Box::new(m20260622_000004_product_search::Migration),
|
||||
Box::new(m20260622_000005_product_search_aggregate::Migration),
|
||||
// inject-above (do not remove this comment)
|
||||
]
|
||||
}
|
||||
|
||||
92
migration/src/m20260622_000004_product_search.rs
Normal file
92
migration/src/m20260622_000004_product_search.rs
Normal file
@@ -0,0 +1,92 @@
|
||||
//! Full-text + fuzzy search over the product catalog.
|
||||
//!
|
||||
//! Storefront search has to cope with Slovak text (diacritics, ad-hoc spelling)
|
||||
//! and customer typos, while staying entirely inside Postgres — the catalog is
|
||||
//! small (hundreds of products), so a separate search engine would be pure
|
||||
//! operational overhead. This migration sets up:
|
||||
//!
|
||||
//! 1. `unaccent` + `pg_trgm` extensions, and an IMMUTABLE `f_unaccent` wrapper
|
||||
//! (the stock `unaccent` is only STABLE, so it can't be used in an index
|
||||
//! expression without wrapping it).
|
||||
//! 2. a `sk_unaccent` text-search configuration: the `simple` dictionary
|
||||
//! (no English stemming, which would mangle Slovak) folded through
|
||||
//! `unaccent` so "kompresor" and "kompresór" tokenize identically.
|
||||
//! 3. a STORED generated `products.search_vector`, weighting the name above
|
||||
//! the description, with a GIN index for `@@` matching.
|
||||
//! 4. a trigram GIN index on the (unaccented) name for fuzzy matching.
|
||||
//!
|
||||
//! The matching query itself lives in `products::Entity::search`.
|
||||
|
||||
use sea_orm_migration::prelude::*;
|
||||
|
||||
#[derive(DeriveMigrationName)]
|
||||
pub struct Migration;
|
||||
|
||||
#[async_trait::async_trait]
|
||||
impl MigrationTrait for Migration {
|
||||
async fn up(&self, m: &SchemaManager) -> Result<(), DbErr> {
|
||||
let db = m.get_connection();
|
||||
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
CREATE EXTENSION IF NOT EXISTS unaccent;
|
||||
CREATE EXTENSION IF NOT EXISTS pg_trgm;
|
||||
|
||||
-- IMMUTABLE wrapper so unaccent() can be used in generated columns
|
||||
-- and index expressions (the extension's own unaccent() is STABLE).
|
||||
CREATE OR REPLACE FUNCTION f_unaccent(text)
|
||||
RETURNS text
|
||||
LANGUAGE sql IMMUTABLE PARALLEL SAFE STRICT AS
|
||||
$func$ SELECT public.unaccent('public.unaccent', $1) $func$;
|
||||
|
||||
-- 'simple' (no stemming) + unaccent: a good fit for Slovak, where
|
||||
-- English stemming is wrong and accents are typed inconsistently.
|
||||
DROP TEXT SEARCH CONFIGURATION IF EXISTS sk_unaccent;
|
||||
CREATE TEXT SEARCH CONFIGURATION sk_unaccent ( COPY = simple );
|
||||
ALTER TEXT SEARCH CONFIGURATION sk_unaccent
|
||||
ALTER MAPPING FOR hword, hword_part, word
|
||||
WITH unaccent, simple;
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
ALTER TABLE products
|
||||
ADD COLUMN search_vector tsvector
|
||||
GENERATED ALWAYS AS (
|
||||
setweight(to_tsvector('sk_unaccent', COALESCE(name, '')), 'A') ||
|
||||
setweight(to_tsvector('sk_unaccent', COALESCE(description, '')), 'B')
|
||||
) STORED;
|
||||
|
||||
CREATE INDEX idx_products_search_vector
|
||||
ON products USING GIN (search_vector);
|
||||
CREATE INDEX idx_products_name_trgm
|
||||
ON products USING GIN (f_unaccent(name) gin_trgm_ops);
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
Ok(())
|
||||
}
|
||||
|
||||
async fn down(&self, m: &SchemaManager) -> Result<(), DbErr> {
|
||||
let db = m.get_connection();
|
||||
|
||||
// Drop the trigram index (it depends on f_unaccent) before the function;
|
||||
// dropping the column takes its own GIN index with it.
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
DROP INDEX IF EXISTS idx_products_name_trgm;
|
||||
ALTER TABLE products DROP COLUMN IF EXISTS search_vector;
|
||||
DROP TEXT SEARCH CONFIGURATION IF EXISTS sk_unaccent;
|
||||
DROP FUNCTION IF EXISTS f_unaccent(text);
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
// The unaccent / pg_trgm extensions are left installed: other objects may
|
||||
// rely on them and they are harmless on their own.
|
||||
Ok(())
|
||||
}
|
||||
}
|
||||
232
migration/src/m20260622_000005_product_search_aggregate.rs
Normal file
232
migration/src/m20260622_000005_product_search_aggregate.rs
Normal file
@@ -0,0 +1,232 @@
|
||||
//! Broaden product search to the whole purchasable surface.
|
||||
//!
|
||||
//! The `product_search` migration could only index columns living on `products`
|
||||
//! itself (name, description), because a STORED generated column may not read
|
||||
//! other tables. To also match by tag, variant label and SKU, `search_vector`
|
||||
//! becomes a plain column maintained by triggers:
|
||||
//!
|
||||
//! * `kompress_build_product_search(name, description, id)` builds the weighted
|
||||
//! vector for one product, pulling tags + variant labels + SKUs by id
|
||||
//! (name = A, tags + labels = B, description + SKU = C).
|
||||
//! * a BEFORE trigger on `products` keeps a product's own row in sync, and
|
||||
//! * AFTER triggers on `product_variants`, `product_product_tags` and tag
|
||||
//! renames refresh the affected product(s).
|
||||
//!
|
||||
//! The result is one `products.search_vector` that every search query can reuse,
|
||||
//! always consistent with the catalog.
|
||||
|
||||
use sea_orm_migration::prelude::*;
|
||||
|
||||
#[derive(DeriveMigrationName)]
|
||||
pub struct Migration;
|
||||
|
||||
#[async_trait::async_trait]
|
||||
impl MigrationTrait for Migration {
|
||||
async fn up(&self, m: &SchemaManager) -> Result<(), DbErr> {
|
||||
let db = m.get_connection();
|
||||
|
||||
// Swap the generated column (name + description only) for a plain column
|
||||
// the triggers can own. Dropping it takes its GIN index with it.
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
ALTER TABLE products DROP COLUMN IF EXISTS search_vector;
|
||||
ALTER TABLE products ADD COLUMN search_vector tsvector;
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
// Single source of truth for a product's search document.
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
CREATE OR REPLACE FUNCTION kompress_build_product_search(
|
||||
p_name text, p_description text, p_id integer
|
||||
) RETURNS tsvector
|
||||
LANGUAGE sql STABLE AS $func$
|
||||
SELECT
|
||||
setweight(to_tsvector('sk_unaccent', COALESCE(p_name, '')), 'A')
|
||||
|| setweight(to_tsvector('sk_unaccent', COALESCE((
|
||||
SELECT string_agg(t.name, ' ')
|
||||
FROM product_product_tags ppt
|
||||
JOIN product_tags t ON t.id = ppt.product_tag_id
|
||||
WHERE ppt.product_id = p_id
|
||||
), '')), 'B')
|
||||
|| setweight(to_tsvector('sk_unaccent', COALESCE((
|
||||
SELECT string_agg(v.label, ' ')
|
||||
FROM product_variants v
|
||||
WHERE v.product_id = p_id
|
||||
), '')), 'B')
|
||||
|| setweight(to_tsvector('sk_unaccent', COALESCE(p_description, '')), 'C')
|
||||
|| setweight(to_tsvector('sk_unaccent', COALESCE((
|
||||
SELECT string_agg(v.sku, ' ')
|
||||
FROM product_variants v
|
||||
WHERE v.product_id = p_id AND v.sku IS NOT NULL
|
||||
), '')), 'C');
|
||||
$func$;
|
||||
|
||||
-- Refresh one product's stored vector (used by the satellite triggers).
|
||||
CREATE OR REPLACE FUNCTION kompress_refresh_product_search(p_id integer)
|
||||
RETURNS void LANGUAGE sql AS $func$
|
||||
UPDATE products
|
||||
SET search_vector = kompress_build_product_search(name, description, id)
|
||||
WHERE id = p_id;
|
||||
$func$;
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
// BEFORE trigger on products: recompute on its own writes. When a refresh
|
||||
// only touches search_vector (name + description unchanged) it skips the
|
||||
// recompute and keeps the supplied value — which also breaks recursion.
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
CREATE OR REPLACE FUNCTION kompress_products_search_tg() RETURNS trigger
|
||||
LANGUAGE plpgsql AS $func$
|
||||
BEGIN
|
||||
IF TG_OP = 'UPDATE'
|
||||
AND NEW.name IS NOT DISTINCT FROM OLD.name
|
||||
AND NEW.description IS NOT DISTINCT FROM OLD.description
|
||||
AND NEW.search_vector IS DISTINCT FROM OLD.search_vector THEN
|
||||
RETURN NEW;
|
||||
END IF;
|
||||
NEW.search_vector :=
|
||||
kompress_build_product_search(NEW.name, NEW.description, NEW.id);
|
||||
RETURN NEW;
|
||||
END;
|
||||
$func$;
|
||||
|
||||
DROP TRIGGER IF EXISTS products_search_tg ON products;
|
||||
CREATE TRIGGER products_search_tg
|
||||
BEFORE INSERT OR UPDATE ON products
|
||||
FOR EACH ROW EXECUTE FUNCTION kompress_products_search_tg();
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
// Variants: any change refreshes the owning product (both, on reparent).
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
CREATE OR REPLACE FUNCTION kompress_variants_search_tg() RETURNS trigger
|
||||
LANGUAGE plpgsql AS $func$
|
||||
BEGIN
|
||||
IF TG_OP = 'DELETE' THEN
|
||||
PERFORM kompress_refresh_product_search(OLD.product_id);
|
||||
RETURN OLD;
|
||||
END IF;
|
||||
PERFORM kompress_refresh_product_search(NEW.product_id);
|
||||
IF TG_OP = 'UPDATE' AND NEW.product_id IS DISTINCT FROM OLD.product_id THEN
|
||||
PERFORM kompress_refresh_product_search(OLD.product_id);
|
||||
END IF;
|
||||
RETURN NEW;
|
||||
END;
|
||||
$func$;
|
||||
|
||||
DROP TRIGGER IF EXISTS product_variants_search_tg ON product_variants;
|
||||
CREATE TRIGGER product_variants_search_tg
|
||||
AFTER INSERT OR UPDATE OR DELETE ON product_variants
|
||||
FOR EACH ROW EXECUTE FUNCTION kompress_variants_search_tg();
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
// Tag links: attaching/detaching a tag refreshes the product.
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
CREATE OR REPLACE FUNCTION kompress_product_tags_link_search_tg() RETURNS trigger
|
||||
LANGUAGE plpgsql AS $func$
|
||||
BEGIN
|
||||
IF TG_OP = 'DELETE' THEN
|
||||
PERFORM kompress_refresh_product_search(OLD.product_id);
|
||||
RETURN OLD;
|
||||
END IF;
|
||||
PERFORM kompress_refresh_product_search(NEW.product_id);
|
||||
RETURN NEW;
|
||||
END;
|
||||
$func$;
|
||||
|
||||
DROP TRIGGER IF EXISTS product_product_tags_search_tg ON product_product_tags;
|
||||
CREATE TRIGGER product_product_tags_search_tg
|
||||
AFTER INSERT OR UPDATE OR DELETE ON product_product_tags
|
||||
FOR EACH ROW EXECUTE FUNCTION kompress_product_tags_link_search_tg();
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
// Renaming a tag refreshes every product carrying it.
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
CREATE OR REPLACE FUNCTION kompress_tag_rename_search_tg() RETURNS trigger
|
||||
LANGUAGE plpgsql AS $func$
|
||||
BEGIN
|
||||
UPDATE products p
|
||||
SET search_vector =
|
||||
kompress_build_product_search(p.name, p.description, p.id)
|
||||
WHERE p.id IN (
|
||||
SELECT ppt.product_id FROM product_product_tags ppt
|
||||
WHERE ppt.product_tag_id = NEW.id
|
||||
);
|
||||
RETURN NEW;
|
||||
END;
|
||||
$func$;
|
||||
|
||||
DROP TRIGGER IF EXISTS product_tags_rename_search_tg ON product_tags;
|
||||
CREATE TRIGGER product_tags_rename_search_tg
|
||||
AFTER UPDATE OF name ON product_tags
|
||||
FOR EACH ROW EXECUTE FUNCTION kompress_tag_rename_search_tg();
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
// Backfill existing rows, then (re)create the GIN index for `@@`.
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
UPDATE products
|
||||
SET search_vector = kompress_build_product_search(name, description, id);
|
||||
|
||||
CREATE INDEX idx_products_search_vector
|
||||
ON products USING GIN (search_vector);
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
Ok(())
|
||||
}
|
||||
|
||||
async fn down(&self, m: &SchemaManager) -> Result<(), DbErr> {
|
||||
let db = m.get_connection();
|
||||
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
DROP TRIGGER IF EXISTS product_tags_rename_search_tg ON product_tags;
|
||||
DROP TRIGGER IF EXISTS product_product_tags_search_tg ON product_product_tags;
|
||||
DROP TRIGGER IF EXISTS product_variants_search_tg ON product_variants;
|
||||
DROP TRIGGER IF EXISTS products_search_tg ON products;
|
||||
DROP FUNCTION IF EXISTS kompress_tag_rename_search_tg();
|
||||
DROP FUNCTION IF EXISTS kompress_product_tags_link_search_tg();
|
||||
DROP FUNCTION IF EXISTS kompress_variants_search_tg();
|
||||
DROP FUNCTION IF EXISTS kompress_products_search_tg();
|
||||
DROP FUNCTION IF EXISTS kompress_refresh_product_search(integer);
|
||||
DROP FUNCTION IF EXISTS kompress_build_product_search(text, text, integer);
|
||||
ALTER TABLE products DROP COLUMN IF EXISTS search_vector;
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
// Restore the name + description generated column from the prior migration.
|
||||
db.execute_unprepared(
|
||||
r#"
|
||||
ALTER TABLE products
|
||||
ADD COLUMN search_vector tsvector
|
||||
GENERATED ALWAYS AS (
|
||||
setweight(to_tsvector('sk_unaccent', COALESCE(name, '')), 'A') ||
|
||||
setweight(to_tsvector('sk_unaccent', COALESCE(description, '')), 'B')
|
||||
) STORED;
|
||||
|
||||
CREATE INDEX idx_products_search_vector
|
||||
ON products USING GIN (search_vector);
|
||||
"#,
|
||||
)
|
||||
.await?;
|
||||
|
||||
Ok(())
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user