The image_file attribute was not being correctly set when saving the ECSV file. I corrected this to use the FF name if available, otherwise to construct a name using the same logic as when saving the FTPdetect file.
Tested by manually reducing three fireballs over three days and validating the results.
The bug was causing an inconsistency when processing data with WMPL. Trajectories calculated from ftpdetect files contain the image names in the comment field, but as its missing in ECSV files, fireball data did not contain the filename. I actually use this field on the UKMON website to determine the images to display with a solved fireball. Hence important to me. :)
This fix also includes an improvement to detection of python version when importing TK. Rather than using exception handling, I check the python major version, which feels cleaner, but its a matter of choice :)
The image_file attribute was not being correctly set when saving the ECSV file. I corrected this to use the FF name if available, otherwise to construct a name using the same logic as when saving the FTPdetect file.
Tested by manually reducing three fireballs over three days and validating the results.
The bug was causing an inconsistency when processing data with WMPL. Trajectories calculated from ftpdetect files contain the image names in the comment field, but as its missing in ECSV files, fireball data did not contain the filename. I actually use this field on the UKMON website to determine the images to display with a solved fireball. Hence important to me. :)
This fix also includes an improvement to detection of python version when importing TK. Rather than using exception handling, I check the python major version, which feels cleaner, but its a matter of choice :)
Closes #147