Python Analytics SDK

For engineers Other libraries and apps
1

Hi,

I’m working with the Python Analytics API trying to read data from Google Cloud Storage (written by Storage Loader).
However the event parsing fails for me on:

return jsonify_good_event(line.split('\\t'), known_fields, add_geolocation_data)
AttributeError: 'dict' object has no attribute 'split'

I’m not quite sure how to handle this or what the expected input format would be as I’m passing on the TSV row string. Do I have to take care of transformations beforehand?

This is my code:

def snowplowTsv2Json(content):
    """Use the Snowplow SDK to generate a JSON object from TSV files
    Args:
         content: File retrieved from GCS
    """
    snowplowEvents = []
    print(type(content))

    with open(content, encoding='utf-8') as tsvfile:
        reader = [line.rstrip('\n') for line in tsvfile]
        #reader = csv.DictReader(tsvfile, dialect='excel-tab')
        print('First row of reader: {}, Type of reader var {}'.format(reader[0], type(reader)))
        for row in reader:
            print('Iterator row: {}'.format(str(row)))
            try:
                jsonRow = snowplow_analytics_sdk.event_transformer.transform(
                    str(row))
                print('Json row: {}'.format(jsonRow))
                snowplowEvents.append(
                    snowplow_analytics_sdk.event_transformer.transform(jsonRow))
            except snowplow_analytics_sdk.snowplow_event_transformation_exception.SnowplowEventTransformationException as e:
                for error_message in e.error_messages:
                    print('Error in snowplowTsv2Json: {}'.format(error_message))
        return snowplowEvents

Do you have any idea on how to progress? I’ve tried various options of feeding the TSV rows, but none worked…

2

It looks like you’re calling this twice? As far as I’m aware, you should only need to transform the row once and then you’ll have the JSON object that you’re after.

3

Oh my! Thank you @PaulBoocock! I was trying to fix the input while the problem was somewhere else… Desperately waiting for the weekend!

1 Like