Error in GB diet composition data - only 1 prey per input line

[gavinfay]

@sgaichas There are multiple lines per predator size bin for the diet data in the GB data set. Looking at the -ts.dat file, it appears each row only contains information for one prey item, rather than a composition across all prey.
I'll have a look at the generation for this but flagging in case you can think why this might be happening. I note that the error does not appear in the NOBA-ts.dat files, they are OK.

(can't do some rowsumming because the other prey proportions are based on the sum of the rest of the vector)

[sgaichas]

ugh, sorry--let me check. It uses the same code to do simulated and real data but I probably missed something

[gavinfay]

Thanks. I'm just not familiar with how the hydraDataList is populated, and couldn't figure it out from some initial sweeps of code, otherwise would try to figure it out myself.

[sgaichas]

I had a line out of order--can't pivot wider before removing the extra named columns! Also the species that is never prey was out of column order. All should be fixed now in thefaylab/hydradata@7b3deb5

And new data files in 91cc736

[gavinfay]

Ah, great - thanks!

[gavinfay]

Diagnostics look less like DNA tests now. Thank you!
What are surveys 3 & 4?

[sgaichas]

uh oh. let me check, may have accidentally left in the vessel separation. or I made a different mistake.

[sgaichas]

Surveys 3 and 4 are summer and winter diet data which are included in the diet inputs and therefore come out as a different survey, with fall and spring being surveys 1 and 2. You can ignore them, and I will remove them from the model inputs. Sorry about that!

[sgaichas]

fixed in thefaylab/hydradata@6931356

new file generated in 91cc736