Som klimaforsker ønsker jeg at de beregnede klimaindeksene skal plasseres rett sted med korrekte metadata for å distribuere til brukere gjennom distribusjonsløsning i Klimakverna WP4
As a climate scientist, I want the option to make my calculated climate indices ~available to end-users automatically~ with the correct metadata.
This test case builds upon test case 1 (producing a climate index, being a statistical aggregation of other datasets) and loosely on test case 5 (producing a climate variable, a derivative dataset that does not aggregate, and which keeps the same time resolution as its parents). Separate subtasks should be created for each of these cases if needed. The aim of the current test case, is to ensure that when netCDF files are written, they are written with the correct metadata as per the spreadsheet above.
Metadata fields can have different origins, including:
Inherited from original input files?
Deduced from performed operation?
Sourced from standardized config file for the variable/index?
The purpose of this test case is to identify this distinction; to conclude what information we should require a "dataset producer" to provide; and to implement the proper setting of metadata fields.
If possible, the metadata filling should be considered an independent step from the creation of new files. This is because we might want to fill or correct metadata in previously generated files without touching the data. Technically, this means any information kept in-memory during file creation, will not be available when filling in the metadata of that same file. We want to find out if this is possible, if the path and basic contents of a file is sufficient for identifying it and fetching or deducing the correct metadata.
Output
Output files are the same as the output files in test cases 1 and 5. But if needed, as a means of testing, a separate file could be produced so that one file exists without the correct metadata and one file exists with the correct metadata.
Related user stories (same as test case 2):
Sources: \ https://github.com/KSSno/Klimakverna-WP1-WP2-WP3b-WP5/issues/11 \ https://docs.google.com/document/d/12joGOmBR4xNjGW45rXxaNbIMv8jDtqAZL0duI7WlwGk/edit?usp=sharing
Input files
This task relies on output netCDF files from test cases 1 and 5.
For list of required metadata attributes, see: https://docs.google.com/spreadsheets/d/1L4dwsB3iH11kxyIvqtnhpgJyaKDWvoGgDCg0klff83c/edit#gid=1886269309
Method
This test case builds upon test case 1 (producing a climate index, being a statistical aggregation of other datasets) and loosely on test case 5 (producing a climate variable, a derivative dataset that does not aggregate, and which keeps the same time resolution as its parents). Separate subtasks should be created for each of these cases if needed. The aim of the current test case, is to ensure that when netCDF files are written, they are written with the correct metadata as per the spreadsheet above.
Metadata fields can have different origins, including:
The purpose of this test case is to identify this distinction; to conclude what information we should require a "dataset producer" to provide; and to implement the proper setting of metadata fields.
If possible, the metadata filling should be considered an independent step from the creation of new files. This is because we might want to fill or correct metadata in previously generated files without touching the data. Technically, this means any information kept in-memory during file creation, will not be available when filling in the metadata of that same file. We want to find out if this is possible, if the path and basic contents of a file is sufficient for identifying it and fetching or deducing the correct metadata.
Output