Approximate Results Count Search
The "Approximate Results Count" feature enables users to perform a search that returns an estimate of the total number of products within a specific similarity score range. This feature is useful for filtering and narrowing down search results based on how closely they match the query.
How It Works:
The search method accepts a total_counts
object, which defines the similarity thresholds for the search, and a text
field, which contains the search query. The similarity score for each product is measured on a scale from 0 to 1, where 1 represents a perfect match. Users can specify a range by setting the min and max thresholds to filter the products within the desired similarity range.
total_counts
: An object containing two float values:
min_score_threshold
: The minimum similarity score threshold (a float value between 0 and 1).max_score_threshold
: The maximum similarity score threshold (a float value between 0 and 1).
text
: A string representing the search query.
Once the request is sent, the system returns a single number, total_count
, which represents the number of products that have a similarity score between the defined min_score_threshold
and max_score_threshold
values.
Example:
Request:
{
"text": "summer shoes",
"total_counts": {
"min": 0.7,
"max": 0.9
}
}
Response:
{
"total_count": 42
}
In this example, a request is made to find products related to "summer shoes" with a similarity score between 0.7 and 0.9. The response returns a total_count of 42, indicating that 42 products fall within this similarity range.
Updated 16 days ago