Use species names in column headers when generating simulated data #51

Activity

1. dmitrisvetlov

   self-assigned this

   on Jul 25, 2026
2. dmitrisvetlov

   added the

   Feature

   issue type on Jul 25, 2026
3. dmitrisvetlov commented on Jul 25, 2026

   

   dmitrisvetlov

   on Jul 25, 2026

   Last edited by dmitrisvetlov

   CollaboratorAuthor

   More actions

   @Munsky, I will need your help in diagnosing the failures. Here are the following changes I made:

   • Generation of simulated data now creates columns containing the parameter names (e.g., instead of exp1_s1, we will write exp1_rna)
   • Data loading therefore changes:
     • linkedSpecies is now optional
     • The procedure changes:
       1. First, we attempt to automatically link species based on the presence of columns named exp1_[speciesName].
       2. Second, we link species based on the linkedSpecies parameter if it is not empty.
       3. Third, now we apply conditions and parameter math. The reason is that this allows the user to write these in terms of the names of the species in the model rather than in terms of exp1_s#, because by this point, all the latter columns should have been converted to actual model species names.
     • I have had to update several places in loadData to reflect the fact that linkedSpecies may now be empty.
       • One problem is that downstream code relies on dataSet.linkedSpecies being set. When linkedSpecies is empty, I am setting dataSet.linkedSpecies = species, and I think that this should work, but there are still failures.
       • More generally, I don't understand why the data set should have any concept of linked species. All data stored with a model should only be in terms of the names of the species in the model. Why should there be any record or use of what the corresponding names were in the original data files?
   • I updated a few test cases that were directly affected by the changes to loadData.
   • Throughout various places involved in generating propensity functions, I changed the code to use strings and string methods rather than character arrays wherever possible.
     • I also refactored to use filepaths and filesep for clarity.
     • I discovered the above while trying to test the code by creating SSIT("ToggleSwitch") and encountering problems. Users should not need to know that prefixes need to be single-quoted. The code should work equivalently whether strings are single- or double-quoted. So the affected methods now cast all text inputs to strings and handle them accordingly.
4. dmitrisvetlov

   assigned

   Munsky

   on Jul 25, 2026
5. dmitrisvetlov commented on Jul 28, 2026

   

   dmitrisvetlov

   on Jul 28, 2026

   Last edited by dmitrisvetlov

   CollaboratorAuthor

   More actions

   @Munsky, while the GitHub testing may yield false positives until #47 is merged, there are only two remaining failures on my machine (2026a, Windows), and these might be faster for you to debug. They will probably fail on Mac as well.

   • poisson2Dtest/likelihoodFunctions
   • miscelaneousTests/testCommandLine
6. dmitrisvetlov

   linked a pull request that will close this issue51 use species names in column headers when generating simulated data #53on Jul 31, 2026
7. dmitrisvetlov

   added a commit that references this issue on Aug 14, 2026

   Merge branch for #51 to use updated data loading

   be31867