Skip to content

Documentation for Suggesting Usage and Management API


/get-top-k

Functionality

Returns top-k phrase suggestions based on a partial phrase and optional domain.

Request Parameters

  • phrase (str): The partial input text to complete.
  • domain (str, optional): Contextual domain to narrow down suggestions.
  • top_k (int, default=3): Number of top suggestions to return (max: 100).

Request JSON Example

{
  "phrase": "deep l",
  "domain": "ml",
  "top_k": 5
}
- phrase: The text input for which to generate suggestions. - domain: A string specifying the contextual domain for the input. - top_k: An integer indicating the number of top suggestions to return.

Response JSON Example

{
  "suggestions": [
    {
      "prefix": "deep ",
      "postfix": "learning",
      "matching_span": { "start": 0, "end": 5 },
      "match_type": "prefix",
      "prob": 0.92,
      "labels": ["neural-nets", "ml"],
      "domains": ["ml"]
    }
  ]
}
- prefix: The portion of the phrase matched by user input. - postfix: The suggested text completion. - matching_span: Object indicating which part of the input matched. - match_type: Matching strategy type (exact, prefix, fuzzy, etc.). - prob: Probability/confidence score for suggestion relevance. - labels: Related labels or tags for the suggestion. - domains: Associated application domains.


/phrases/add

Functionality

Adds multiple phrases to the suggestion database.

Request JSON Example

{
  "phrases": [
    {
      "phrase": "deep learning",
      "labels": ["ai", "ml"],
      "domains": ["ml"],
      "prob": 0.95
    }
  ]
}
- phrase: Text of the suggesting phrase. - labels: Tags or categories assigned to the phrase. - domains: Contexts where the phrase applies. - prob: Numeric weighting for phrase ranking (higher = more likely to appear).

Response JSON Example

{
  "phrase_ids": ["a1b2c3d4"]
}
- phrase_ids: List of internal system-assigned IDs for each added phrase.


/phrases/add-labels

Functionality

Adds new labels to a specific phrase.

Request JSON Example

{
  "phrase_id": "abc123",
  "labels": ["new-tag", "trending"]
}
- phrase_id: Unique identifier of the phrase to modify. - labels: New labels to associate with the phrase.


/phrases/remove-labels

Functionality

Removes specific labels from a phrase.

Request JSON Example

{
  "phrase_id": "abc123",
  "labels": ["old-tag"]
}
- phrase_id: Unique phrase identifier. - labels: List of labels to be removed.


/phrases/add-domains

Functionality

Associates domains with a phrase.

Request JSON Example

{
  "phrase_id": "abc123",
  "domains": ["science", "edu"]
}
- phrase_id: ID of the phrase to update. - domains: New domains to associate with the phrase.


/phrases/remove-domains

Functionality

Removes domains from a phrase.

Request JSON Example

{
  "phrase_id": "abc123",
  "domains": ["edu"]
}
- phrase_id: Phrase identifier. - domains: Domains to disassociate.


/phrases/remove-all-domains

Functionality

Globally removes specified domains from all phrases.

Request JSON Example

{
  "domains": ["deprecated"]
}
- domains: List of domain names to remove globally.


/phrases/update-probability

Functionality

Updates the probability score of a phrase.

Request JSON Example

{
  "phrase_id": "abc123",
  "new_prob": 0.85
}
- phrase_id: ID of the phrase to modify. - new_prob: New probability value (0–1) used for ranking.


/phrases/get-info

Functionality

Returns full details for a single phrase.

Response JSON Example

{
  "phrase": "deep learning",
  "labels": ["ai", "ml"],
  "domains": ["ml"],
  "prob": 0.92
}
- phrase: Suggesting phrase text. - labels: Phrase categories or tags. - domains: Relevant contextual scopes. - prob: Phrase ranking score.


/phrases/list

Functionality

Returns paginated list of all phrases.

Request JSON Example

{
  "limit": 20,
  "offset": 0
}
- limit: Number of items to return. - offset: Pagination start index.

Response JSON Example

{
  "items": [
    {
      "phrase": "machine learning",
      "labels": ["ml"],
      "domains": ["ml"],
      "prob": 1.0
    }
  ]
}
- items: List of full suggesting phrase records.