[BUG] found 3 issues

[espiritocz]

hi, i got to some issues, hope those notes will help improve the development of RAiDER:

Issue 1: RAiDER expects non-existing function of dem_stitcher:

in dem.py:
from dem_stitcher.stitcher import download_dem as download_stitched_dem
there is no such function in currently available dem_stitcher (installed from conda-forge, but should equal github repository codes)

Issue 2: some bug in use of 'statevectors'

having S1 EOF file and a station list with heights (i.e. csv with Lat,Lon,Hgt_m columns), the following command:

raiderDelay.py --date 20211210 --time 17:00:00 --model ERA5 --statevectors ../S1A_OPER_AUX_POEORB_OPOD_20211230T121611_V20211209T225942_20211211T005942.EOF --station_file raider_station.list

ends up in errors:

2022-02-16 15:51:29,209 DEBUG Beginning DEM calculation
2022-02-16 15:51:29,233 DEBUG DEM height range for the queried region is 10.00-10.00 m
2022-02-16 15:51:29,234 DEBUG Lats shape is (1,)
2022-02-16 15:51:29,234 DEBUG lat/lon box is -20.544445/-20.544445/-175.390417/-175.390417 (SNWE)
2022-02-16 15:51:29,234 DEBUG Beginning line-of-sight calculation
Traceback (most recent call last):
  File "/gws/smf/j04/nceo_geohazards/software/mambalics/lib/python3.8/site-packages/RAiDER/losreader.py", line 148, in getLookVectors
    svs = np.stack(get_sv(los_type, time, pad), axis=-1)
  File "/gws/smf/j04/nceo_geohazards/software/mambalics/lib/python3.8/site-packages/RAiDER/losreader.py", line 213, in get_sv
    svs = read_txt_file(los_file)
  File "/gws/smf/j04/nceo_geohazards/software/mambalics/lib/python3.8/site-packages/RAiDER/losreader.py", line 395, in read_txt_file
    with open(filename, 'r') as f:
FileNotFoundError: [Errno 2] No such file or directory: 'sv'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/gws/smf/j04/nceo_geohazards/software/mambalics/bin/raiderDelay.py", line 12, in <module>
    parseCMD()
  File "/gws/smf/j04/nceo_geohazards/software/mambalics/lib/python3.8/site-packages/RAiDER/runProgram.py", line 198, in parseCMD
    _tropo_delay(new_args)
  File "/gws/smf/j04/nceo_geohazards/software/mambalics/lib/python3.8/site-packages/RAiDER/runProgram.py", line 210, in _tropo_delay
    (_, _) = tropo_delay(args_copy)
  File "/gws/smf/j04/nceo_geohazards/software/mambalics/lib/python3.8/site-packages/RAiDER/delay.py", line 165, in tropo_delay
    los, lengths = getLookVectors(los, lats, lons, hgts, zref=zref, time=time)
  File "/gws/smf/j04/nceo_geohazards/software/mambalics/lib/python3.8/site-packages/RAiDER/losreader.py", line 164, in getLookVectors
    'getLookVectors: I cannot parse the file {}'.format(look_vecs)
UnboundLocalError: local variable 'look_vecs' referenced before assignment

Issue 3: the request for weather models must be improved, especially in case of CDSAPI:

ok, that's not really a 'bug', but i think you may want to improve things, i.e.
first of all, it now contains the HRES, while ecmwf_api is sometimes less preferred source (as in my case).
second,
the requests should be done using whole set of requested epochs, not one-by-one, as the waiting time can be e.g. 1 request processed in hours (!)

basically, for cdsapi, one may just include parameter 'date' instead of the 'months, years,....', so this can be enough:

params = {
            "product_type": dataset_name,
            "{}".format(levType): 'all',
            "levtype": "pl", 
            "param": var,
            "stream": "oper",
            "type": "an",
            "step": "0",
            "area": nwse,
            "format": "netcdf"}
params['date'] = list(dates.strftime('%Y-%m-%d %H:%M')) \
        if isinstance(dates, pd.core.indexes.datetimes.DatetimeIndex) \
        else list(dates)

i also recommend disabling warnings using:

urllib3.disable_warnings(urllib3.exceptions.InsecureRequestWarning)

(after import urllib3 )

OS Linux CentOS 7

[jlmaurer]

Hi @espiritocz thanks for these comments! We are working on them; #339 should fix issue 1, #332 should fix issue 2. issue 3 is one we haven't addressed yet, so thanks for the suggestion! Hopefully these will get integrated soon.