enb (experiment notebook) is a python library/framework to design, run and analyze computer-based experiments. Persistence, parallelization and plotting can be automatically handled by enb, to help you develop efficient and reproducible science.
Thank you! Still a couple of things to be addressed:
The idea is for to inherit from FAPEC_Abstract and call the initializer appropriately. Please take a look at the other classes and update FAPEC_FITS accordingly. Note that the current code does not even pass the created param_dict to the WrapperInitializer - please make sure to test the codec before the next pull request.
For Zstandard, the idea would be to have 2 different classes: the first one that applies it normally (the class that already existed) and another one that generates a specific dictionary. However, the current PR simply overwrites the previously existing behavior.
For the second Zstandard class, we need to consider the size of the side information (and right now it is not). To do this, you can overwrite the compress() and decompress() methods, i.e., so that they use the tarlite module we discussed this morning. If needed, you can check out the JPEG-LS and M-CALIC codec sources for examples on how to employ this tarlite module. Please verify manually that the reported compressed size is as expected, i.e., only a few bytes larger than the sum of sizes of the compressed file and the side information file.
FAPEC comments have been addressed. Plus, an update on zstandard codec to add dictionary training has been made.