Saved Searches
With saved searches you can capture your search results at a given point in time, and record the inputs, annotations, search parameters and models that you are running through the platform. Saved searches help you to:
Save a search so that you can return to those same results later
Share a search with a collaborator in your app
Carve out data that you want to evaluate your models on as a "golden set"
Split your data into training sets and test tests
Create a saved search
curl --location --request POST 'https://api.clarifai.com/v2/searches/' \
--header 'Content-Type: application/json' \
--header 'Authorization: Key {{key}}' \
--data-raw '{
"searches": [{
"query": {
"ands": [{
"annotation":{
"annotation_info": {
"asset_set": ["set1", "set2"]
}
}
}]
},
"save": true,
"id": "dataset-1589318146",
"name": "Dataset #1589318146"
}]
}
'
List saved searches in an app
curl --location --request GET 'https://api.clarifai.com/v2/searches/' \
--header 'Content-Type: application/json' \
--header 'Authorization: Key {{key}}'
Search by search ID
curl --location --request POST 'https://api.clarifai.com/v2/searches/{{search_id}}' \
--header 'Content-Type: application/json' \
--header 'Authorization: Key {{key}}' \
--data-raw '{
"pagination": {
"page": 1,
"per_page": 5
}
}'
Saved searches are extremely helpful when testing out multiple versions of a model, or collaborating with a team. Just save your search, and refer back to it any time.
Last updated
Was this helpful?