Improve TIMModel implementation by updating the timeseries data type.

[priscavdsluis]

What is the need for this task.
In issue #348 support of the tim file format was introduced.
However the timeseries attribute on the TimModel has a dictionary as type.
This was decided because the times in the time series data should be unique and with each a time a set of values is associated.

But the PolyFile, XYNModel, XYZModel and ForcingBase all do it differently than the tim file.

What is the task?

• Discuss with @veenstrajelmer and team which type is suited
• Update the type of the TimModel.timeseries to e.g. List[TimRecord] where TimRecord has two attributes: time (float) and data (List[float])
• Add validation for unique times in TimModel.

[priscavdsluis]

I have not made this a feature/bug because it is a loose tail from #348

[veenstrajelmer]

Copied from https://github.com/Deltares/dfm_tools/blob/main/tests/examples_workinprogress/workinprogress_meshkernel_creategrid.py

While doing a quick test I noticed the TimModel is not quite in line with other similar hydrolib-core objects. Since the timfile is similar than the files for XYNModel/XYZModel etc, I would expect more similarities in the object formatting. Below is some example code that shows how the objects are built up. In line with this, I would expect TimModel to have a points property and TimModel.points to be a list of Points (or similar object). Not set in stone of course, but it is an expectation which can be discussed.

For completeness, the Timeseries object (from a bc-file) is also included in the example code, but since that object is way more complex it might not be relevant.

import hydrolib.core.dflowfm as hcdfm

#xyz model
file_xyz = r'p:\archivedprojects\11205258-006-kpp2020_rmm-g6\C_Work\08_RMM_FMmodel\geometry_j19_NL_6-v2\rmm_vzm_v1p1_initial_water_level_onlyriv.xyz'
data_xyz = hcdfm.XYZModel(file_xyz)
print('\nXYZModel.points is list of XYZPoint, which has x/y/z/comment keys (first 5 are printed):') #TODO: comment can be dropped?
print(type(data_xyz.points))
print(data_xyz.points[:5])

#load xyn model
file_xyn = r'p:\archivedprojects\11206813-006-kpp2021_rmm-2d\C_Work\31_RMM_FMmodel\geometry_j19_6-v2\output_locations\rmm_vzm-j19_6-v2b_3_measurement_obs.xyn'
data_xyn = hcdfm.XYNModel(file_xyn)
print('\nXYNModel.points is list of XYNPoint, which has x/y/n keys (first 5 are printed):')
print(type(data_xyn.points))
print(data_xyn.points[:5])

#load pol/tek/pli/ldb file
file_pli = r'p:\11208054-004-dcsm-fm\models\model_input\bnd_cond\pli\DCSM-FM_OB_all_20181108.pli'
polyfile_object = hcdfm.PolyFile(file_pli)
poly_object_first = polyfile_object.objects[0]
print('\nPolyObject.points is list of Point, which has x/y/z/data keys (first 5 are printed):')
print(type(poly_object_first.points))
print(poly_object_first.points[:5])

# #bcfile
# file_bc_ts = r'c:\DATA\dfm_tools_testdata\hydrolib_bc\DCSM\steric_DCSM-FM_OB_all_20181108_firstpart.bc'
# m = hcdfm.ForcingModel(file_bc_ts)
# timeseries_object = m.forcing[0]
# print('\nTimeseries.datablock is list of lists with [time, val] (first 5 are printed):')
# print(type(timeseries_object.datablock))
# print(timeseries_object.datablock[:5])

#timfile
file_tim = r'p:\archivedprojects\11205258-006-kpp2020_rmm-g6\C_Work\08_RMM_FMmodel\structures_toRTC\Algerakering_GateLowerEdgeLevel.tim'
data_tim = hcdfm.TimModel(file_tim)
print('\nTimModel.timeseries is dict with times as keys and rest of columns as list of values:')
print(type(data_tim.timeseries))
print(data_tim.timeseries)

[veenstrajelmer]

Alternatively, we could consider to add the refdate from the mdu to the timfile (as a property) if it is being read via the FMModel tree