def aggregate_addons()

in mozetl/addon_aggregates/addon_aggregates.py [0:0]

• 26 lines of code
• 1 McCabe index (conditional complexity)

def aggregate_addons(df):
    """
    Aggregates add-on indicators by client, channel, version and locale.
    The result is a DataFrame with the additional aggregate columns:

    n_self_installed_addons (int)
    n_shield_addons (int)
    n_foreign_installed_addons (int)
    n_system_addons (int)
    n_web_extensions (int)
    first_addon_install_date (str %Y%m%d)
    profile_creation_date (str %Y%m%d)

    for each of the above facets.

    :param df: an expoded instance of main_summary by active_addons
               with various additional indicator columns
    :return SparkDF: an aggregated dataset with each of the above columns
    """
    addon_aggregates = (
        df.distinct()
        .groupBy("client_id", "normalized_channel", "app_version", "locale")
        .agg(
            fun.sum("is_self_install").alias("n_self_installed_addons"),
            fun.sum("is_shield_addon").alias("n_shield_addons"),
            fun.sum("is_foreign_install").alias("n_foreign_installed_addons"),
            fun.sum("is_system").alias("n_system_addons"),
            fun.sum("is_web_extension").alias("n_web_extensions"),
            fun.min(
                fun.when(
                    df.is_self_install == 1,
                    fun.date_format(
                        fun.from_unixtime(fun.col("install_day") * 60 * 60 * 24),
                        "yyyyMMdd",
                    ),
                ).otherwise(None)
            ).alias("first_addon_install_date"),
            fun.date_format(
                fun.from_unixtime(fun.min("profile_creation_date") * 60 * 60 * 24),
                "yyyyMMdd",
            ).alias("profile_creation_date"),
        )
    )
    return addon_aggregates