MapperPipeline converting tupes to lists?

[jread]

I have been working on implementing a pipeline which will push data from AppEngine into BigQuery. I am defining my mapper pipeline class in the same manner as in this article https://developers.google.com/bigquery/articles/datastoretobigquery#definepipeline except that I am passing a "filters" param to the input_reader. Specifically, I am passing this:

params={

       "input_reader":{

           "entity_kind": entity_type,

           "filters": [('pid', '=', pid), ('tags', '=', '2000')]

       },

       "output_writer":{

           "filesystem": "gs",

           "gs_bucket_name": settings.GOOGLE_CLOUD_STORAGE_BUCKET_NAME,

           "output_sharding": "input"

       }

}

When I do that, however, lines 532 and 533 of mapreduce.input_readers.py are raising an exception stating that "Filter should be a tuple" and showing me the first filter but represented as a list instead of a tuple. Commenting out these two lines in input_readers.py allows the pipeline to proceed as expected and everything works properly. Am I doing something wrong here or is there a bug that is changing my tuples into lists?

Thanks,

Jeff.