Rewrite code to use new streaming data model

[smithjessk]

Each tuple should be of the following format:

Name	Type	Description
chip_type	string	the type of this chip (e.g. 'pd.huex.1.0.st.v2')
file_name	string	...
sample_id	integer	the id of this sample
intensity	float	the recorded intensity of the sample

[smithjessk]

The current process in Oligo needs to be somewhat emulated.

[smithjessk]

• Grab the chip type from the first file in the list and load that SQL file. Use that for all files in the reading collection. Implemented via template arguments that carry over.
• Intermediate tuples are then produced of the form (file_name, chip_type, fid, intensity) where chip_number is determined by a template argument. fid is the position of this measurement if rows were put end-to-end. So named because this is how it is called in oligo. See the corresponding milestone
• Need to collect the set of chip_types so that no repeat work is done in bringing the relationship files into memory.
• Note that these relationship files contain, for a given chip_type, the mapping of position to sample_id (see Extract information from SQL files  #21). Once this information is extracted, it will be applied to the tuples with a Join on fid in both the tuples and the database files.