Perhaps it would be easier to have the code read a table that tells it which manipulations to include in which comparisons? See this PR, which adds a simple CSV table that flags all the manipulations we might want to include in a "water experiments" meta-analysis. The code could then read the table and use it to filter the srdb data. Additional columns could be added for additional analytical groupings.
[ ] Use manipulation_table.csv to filter through SRDB for the manipulation studies we want, which are marked as an "X" in the table
Re: @bpbond's suggestion...
manipulation_table.csvto filter through SRDB for the manipulation studies we want, which are marked as an "X" in the table