Improve logging when files do not pass senting time filter
Silence some known warnings, or reduce them to info messages
Some special cases for NPP-VIIRS data
Add check on fraction of "good" pixels before reading a granule
Add robustness against missing variables
Better capture an exception for corrupt files
Fully disable FutureWarning
Handle incorrect tracking_id attribute in NPP-Cloud productChange time handling in time dependent plottingHandle missing support variables in NPP-CloudHandle row count mismatch for cloud comparison
Fix time-string parsing error
Use the correct method to obtain a reference to the scanline variable
Skip files where the scanline dimension has length zero
Handle zero-length input files
Cleanup, adding Doxygen documentation strings. Treat NISE input files differently when generating the input pointer (they may maintain the original file name.
Catch error when reading string attribute and fall back to using the netCDF4 library to read that attribute.
Allow for on-the-fly renaming of products, for intercomparisons using different versions of a product
Remove debug logging statement.
Close file reference where possible
Handle zero-size files correctly
Print just the file name, not the full path when opening an input file
Fixes for NP_BDx and OMI L2 files combined
Cleanup of the temporary QA4ECV hack
Clarify that a fallback will be used for the missing attribute
HACK: support QA4ECV files (will change)
Extract outline from data when it is missing from the metadata
Capture exception from h5py
Changes for reading OMI L2 data as well, update to version 0.1.3
Improve robustness agains missing or empty attributes.
Also allow for multiple time formats.
switch to h5py
Add input_pointer method
reduce output volume in debug mode; clean up error handling
allow the use of an explicit list of files.
PyCAMA updates, removal of vlen data (as much as possible), others
Report file paths instead of file names
Turn PyCAMA into a real package, including load magic for plugins.