relevanceai.utils.filters#

Module Contents#

class relevanceai.utils.filters.Filter(field, dataset_id, condition, condition_value, **kwargs)#

Filters have been designed to become more pythonic.

old filters = [
    {
        "field": "product_name",
        "filter_type": "exact_match",
        "condition": "==",
        "condition_value": "Durian Leather 2 Seater Sofa"
    }
]

new_filters = dataset["product_name"] == "Durian Leather 2 Seater Sofa"
# Produces the same as above

older_filters = [
    {
        "field": "rank",
        "filter_type": "numeric",
        "condition": ">=",
        "condition_value": 2
    },
    {
        "field": "rank",
        "filter_type": "numeric",
        "condition": "<",
        "condition_value": 3
    }
]

new_filters = (dataset["rank"] >= 2) + (dataset["rank"] < 3)

Exists filtering can be accessed in a simple way.

old_filters = [
    {
        "field": "brand",
        "filter_type": "exists",
        "condition": "==",
        "condition_value": " "
    }
]

new_filters = dataset["brand"].exists()

old_filters = [
    {
        "field": "brand",
        "filter_type": "exists",
        "condition": "!=",
        "condition_value": " "
    }
]

new_filters = dataset["brand"].not_exists()

Same with contains.

old_filters = [
    {
        "field": "description",
        "filter_type": "contains",
        "condition": "==",
        "condition_value": "Durian BID"
    }
]

new_filters = dataset["description"].contains("Durian BID")

Date filtering

old_filters = [
    {
        "field": ""insert_date_"",
        "filter_type": "date",
        "condition": "==",
        "condition_value": "2020-07-01"
    }
]

new_filters = dataset["_insert_date"].date(2020-07-01")
property dtype(self)#
get(self)#