You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It has only been tested on a stitched stack, e.g. /home/oku/Developments/MobyLe/data/downloaded/stacks/nl_veenweiden_s1_dsc_t037/stack.
Current unit testing runs on an example metadata of the reference img of the above stack (20200328). We should consider add more testing case.
Comparing with the original version (readMetadata function), the part of determine the stack type based on path names is taken out.
# type of image: stitched: TRUE / burst: FALSEimgType=str(resFile).split(os.sep)[-2].isdigit()
The above code breaks the unittest since it requires the metadata file to be strictly located at a path.We need to have a way to tell the stack type based on metadata info itself.
There are two many regular expression searches. We may consider build a searching dict with {attr_key: reg_patterns}. This ideas follows Meiert's comment in PR Reconstruct slc from inteferograms #9:
There is a huge number of the regular expression searches here and in the following. I'd suggest moving this to a separate module and making a wrapper which calls them by name, allowing the correct implementation to be retrieved. This also makes one central place for keeping track of all these things and easily supports adding additional methods as needed.
Happy to discuss @rogerkuou
The following part is missing and need more investigations in the original implementation:
# ++++ 24 and 25 - azFmRateArray and dcPolyArray# Are skipped because the io.datetimeToMJD function is missing
There are many hard-coded metadata, for example the azimuth_resolution get several constant value from S1 report, and is selected based on the swath number. It would be better if we can put these constant into a central place, e.g. constants.py and import them here. This would be easier to maintain.
Check compatability to Doris v4 metadata files: /project/caroline/Share/stacks/tsx/nl_pijnacker_tsx_dsc_t048/process
The text was updated successfully, but these errors were encountered:
Currently the metadata reading functions are implemented in io.py. This is a direct copy from: Paolo's implementation based on Gecoris
There following imperfections:
It has only been tested on a stitched stack, e.g.
/home/oku/Developments/MobyLe/data/downloaded/stacks/nl_veenweiden_s1_dsc_t037/stack
.Current unit testing runs on an example metadata of the reference img of the above stack (20200328). We should consider add more testing case.
Comparing with the original version (readMetadata function), the part of determine the stack type based on path names is taken out.
The above code breaks the unittest since it requires the metadata file to be strictly located at a path.We need to have a way to tell the stack type based on metadata info itself.
There are many hard-coded metadata, for example the
azimuth_resolution
get several constant value from S1 report, and is selected based on the swath number. It would be better if we can put these constant into a central place, e.g.constants.py
and import them here. This would be easier to maintain.Check compatability to Doris v4 metadata files:
/project/caroline/Share/stacks/tsx/nl_pijnacker_tsx_dsc_t048/process
The text was updated successfully, but these errors were encountered: