Closed wmakGIT closed 1 year ago
That does not sound right, but I am not able to reproduce this.
This doesn't seem to be an isolated case. That data must have been generated using code from another branch, most likely from https://github.com/AI4OPT/ACOPFGenerator/pull/15 (which was closed without merging).
According to the diff from that PR, nodal primal/dual solution was only extracted for buses that contain a load. This matches what I see in the JSON files: there are 201 loads in ieee300, and the solutions have primal/dual solution exactly for the corresponding buses.
We'll have to re-generate those datasets.
I have reformatted some of the sampling code in #18, #19, #20. I am in the process of re-generating the datasets; will keep this issue open in the meantime
The bug has been fixed and the data has been re-generated.
Symptom: After parsing one of the solution for IEEE 300 (pglib_opf_case300_ieee_9.json.gz) as a dictionary structure (dict):