Closed BrianSchiller closed 2 weeks ago
Somehow these changes break the compute_features command
So far I understand that cfh.computing_features gets the variable "list_feature_computation_job" from the SparkleFeatureDataCSV, and the entries in the variable contain the names of the extractors that should be run to compute the feature data
The SparkleFeatureDataCSV, based on FeatureData/sparkle_feature_data.csv, now contains a lot more columns that are all converted to extractors, even though there is only one extractor available.
(This still works in development, only breaks in the version with the new extractor wrapper)
I found the error.
The problem was that the special_sparkle_string that you asked me to remove was used as a seperator in the compute_feature code. I brought the special_sparkle_string back and rewrote the rest of the wrapper based on your suggestions and now everything should work.
Wrote sparkle_extractor_wrapper, based on the existing implemenation of spakle_run_default_wrapper.