Presets, stopwords and stemming

Three small, independent configuration resources a search can reference by name — none of them change ranking the way search rules do; they change what a request is allowed to omit and how the tokenizer reads a word. Every operation on this page needs a full key.

Presets

A preset stores a full search-parameters (or multi-search) payload under a name, so a client sends preset=<name> instead of repeating query_by, filter_by and the rest on every call:

await client.presets.upsertPreset({
  presetId: "storefront-default",
  body: { value: { query_by: "name,brand", filter_by: "in_stock:true", per_page: 24 } },
});

curl -G https://api.aacsearch.com/v1/collections/products/documents/search \
  -H "X-NOVA-API-KEY: $NOVA_SEARCH_KEY" \
  --data-urlencode "q=drill" --data-urlencode "preset=storefront-default"

Unlike a search profile, a preset is a flat replay of engine parameters with no release history, no facet/sort allowlist and no browser-safe view — reach for a search profile when a browser key needs to be restricted to it, and a preset for a server-side shortcut you control yourself.

Stopwords

A stopwords set is a list of words a search should ignore — brand boilerplate, connector words the catalogue's language doesn't need matched on:

await client.stopwords.upsertStopwordsSet({
  setId: "countries",
  body: { stopwords: ["Germany", "France", "Italy"], locale: "en" },
});

Reference the set's id from the stopwords search parameter on any search of the collection — the words in the set are removed from the query before it runs. A set does nothing until a search names it.

Stemming dictionaries

A stemming dictionary maps irregular word forms to their root (people → person) for languages or vocabularies the built-in stemmer gets wrong:

cat > irregular-plurals.jsonl <<'EOF'
{"word":"people","root":"person"}
{"word":"children","root":"child"}
EOF

curl -X POST "https://api.aacsearch.com/v1/stemming/dictionaries/import?id=irregular-plurals" \
  -H "X-NOVA-API-KEY: $NOVA_API_KEY" -H "Content-Type: application/json" \
  --data-binary @irregular-plurals.jsonl

Reference the dictionary's id from a field's stem_dictionary schema option and set stem: true on that field when creating a collection. There is no update or delete endpoint yet — re-import under the same id to replace a dictionary's contents.