Create dictionary of statutes from UCCS schema

[newswim]

Today, when we normalize statute data we send a CSV containing the unprocessed charge data to the University of Michigan and they send us back a processed CSV. Since we have a record of the inputs and outputs, we should be able to create a dictionary to map unprocessed strings to their processed strings, based on the case number and charge.

If this step throws an error, like "Statute not found", we'll add that string to a new data set and send it to U of Mich for processing.

@emilyrbartha should have access to uccs_schema.csv.

[tpadmanabhan]

Review with Nick and Nidhi

[tpadmanabhan]

@newswim review with @nicolassaw

[nicolassaw]

@newswim How do we send and receive this data with umichigan? Is there an API we can send jobs to? I have pulled the last week and ran it against the big umich database (7-8k unique charge pairs) and found a lot of the charges on the new cases are missing. This poses some issues with normalizing charge type. Perhaps some formatting changes (like removing count values from "(COUNT ONE) ARSON") might help.

[nicolassaw]

Uploaded a JSON of UMich charge data to the resources folder.