Output internal model to JSON/yaml/something

[rsds143]

the normal flow of sperf is:

• individual file parsers are told to look for a number of matching lines through regex
• the regex also specifies what values to capture
• the captured values are then aggregated on and the report is generated

The parsing is generally the same on all tools for.a given file (but some tools don't parse all files), this was a design decision we made to speed development, it does mean slower parsing for some individual tools but it is useful for this issue. It would be very easy to save to disk these parsed rows. But why?

• eventually we could start to reparse the data so that reports are faster to run after the first one (including different reports would benefit)
• other tooling (graphing, spreadsheets etc) could use the saved off data as a baseline for generating custom reports on the fly without having to implement their own parsing logic or rules some of which were surprisingly complex to account for.

Problems with this

• not all tools read all the same files...do we now parse all files? do we have a tool that does all parsers?
• duplicate data is handled differently by different tooling in sperf. Would need to figure out how to handle this and not break other tools or surprise third parties consuming the data
• where to store it? what if there isn't enough space? Is this easy for people to find?