Getting NameError in unittests for testing a parameter from argparse-CodePudding

I normally use Python for scripts but I am trying to write a unit test and am having a lot of issues. I would like to test a method that creates a parameter --users. The value is how many occurred.

count_users(df, args.metrics)

It is a spark dataframe and the metrics are set like so:

if __name__ == "__main__":
    parser = argparse.ArgumentParser("Processing args")
    parser.add_argument("--metrics", required=True)
    main(parser.parse_args())

The method looks like this:

def count_users(df, metrics):
    users = df.where(df.users > 0).count()
    temp_df = df.withColumn("user_count_values", F.lit(users))
    temp_df.write.json(metrics)

Now I am trying to write my test, and this is where I am not sure about:

def test_count_users(self):
    df = (
        SparkSession.builder.appName("test")
        .getOrCreate()
        .createDataFrame(
            data=[
                (Decimal(0),),
                (Decimal(22),),
            ],
            schema=StructType(
                [
                    StructField("users", DecimalType(38, 4), True),
                ]
            ),
        )
    )

        ap = argparse.ArgumentParser("Test args")
        ap.add_argument("metrics")
  
        args = {_.dest: _ for _ in ap._actions if isinstance(_, _StoreAction)}

        assert args.keys() == {"metrics"}

        count_users(df, args.metrics)

        self.assertTrue(args["metrics"], 1)

Right now I get an error that reads

 count_users(df, args.metrics)
AttributeError: 'dict' object has no attribute 'metrics'

CodePudding user response：

It's unclear what you are trying to achieve with the args = { ... line, or the two asserts. Remove them, use something standard like

import argparse

parser = argparse.ArgumentParser("Test args")
parser.add_argument("--metrics", required=True)

args = parser.parse_args(["--metrics", "output.json"])

count_users(df, args.metrics)

Your args variable won't have the appropriate attribute(s) until you parse the arguments. Of course, normally you'd call

args = parser.parse_args()

and let the user provide the --metrics outputfilename.json arguments to the script. The above is more for example or test use cases.