Blacklists and whitelists

更新时间:
复制 MD 格式

Use blacklists and whitelists to control which search queries appear as drop-down suggestions, top searches, and hints. Add unwanted queries to a blacklist to block them, or add high-quality queries to a whitelist to promote them to the top of results.

How it works

  • Blocking: A blacklist uses inclusion matching (equivalent to substring matching). If any segmented term in a candidate search query matches a blacklisted keyword, the entire query is blocked. For example, if "garbage" is in a blacklist, any query whose terms include "garbage" is excluded from results.

  • Promoting: A whitelist query that meets the recommendation condition is prioritized in the candidate search query results.

  • Conflict resolution: If the same query matches both a blacklist and a whitelist, the blacklist takes precedence.

  • Regular expression support: Blacklists support regular expressions, which lets you cover many similar patterns with a single entry—useful for staying within entry limits.

  • Immediate effect: Changes take effect immediately. Retraining the model is not required.

Behavior by list type

List type Behavior when active
Drop-down suggestion blacklist Matching queries do not appear in drop-down suggestions
Drop-down suggestion whitelist Matching queries are prioritized in drop-down suggestions
Top search blacklist Matching queries are excluded from top searches
Top search whitelist Queries appear in the specified positions of top searches. If no queries are configured for a position, results follow the top search model ranking
Hint blacklist Matching queries are excluded from hints
Hint whitelist Queries rotate as recommended search queries in hints

Create and apply an intervention dictionary

Prerequisites

Before you begin, ensure that you have:

  • Access to the OpenSearch console

  • An existing drop-down suggestion model to attach the dictionary to

Create a blacklist or whitelist

  1. Log on to the OpenSearch console. In the left-side navigation pane, choose Search Algorithm Center > Search Guidance.

  2. On the Drop-down Suggestions page, click Blacklists and Whitelists in the left-side pane.

  3. On the Blacklists and Whitelists page, click Create.

    image

  4. In the Create Whitelist or Blacklist panel, configure the following:

    • Name: Enter a name for the intervention dictionary.

    • Type: Select the list type. Choose Drop-down Suggestions Whitelist to promote queries, or Drop-down Suggestions Blacklist to block them.

    Click Save.

    image

  5. On the Blacklists and Whitelists page, find the dictionary you created and click Manage Entries in the Actions column.

    image

  6. Add intervention entries to the dictionary.

    image

Attach the dictionary to a model

After adding entries, attach the intervention dictionary to a drop-down suggestion model to activate it.

imageimageimage

Limits

List type Max entries
Drop-down suggestion blacklist 500
Drop-down suggestion whitelist 500
Top search blacklist 500
Top search whitelist 10
Hint blacklist 500
Hint whitelist 3
Note: Blacklists support regular expressions. Use a single regex pattern to match multiple similar queries and stay within the entry limit. Blacklists and whitelists referenced by a model cannot be deleted—detach the dictionary from all models before deleting it.