provide read/write interface to all INP sections

[aerispaha]

Summary

All INP sections that follow the common tabular structure with unique indices as the first column can be read via the swmmio.utils.create_dataframeINP function. However, only a limited number of these sections have been added as properties in the higher-level swmmio.core.inp object.

Add coverage for read/writing all INP sections from the swmmio.core.inp object. INP sections are defined currently as properties with setter/getters in the swmmio.core module here. List of INP sections described here in the OWA SWMM repo.

INP Sections to Cover

• TITLE
• OPTIONS
• FILES
• RAINGAGES
• TEMPERATURE
• EVAPORATION
• SUBCATCHMENT
• SUBAREAS
• INFILTRATION
• AQUIFER
• GROUNDWATER
• SNOWPACK
• JUNCTIONS
• OUTFALLS
• STORAGES
• DIVIDERS
• CONDUITS
• PUMPS
• ORIFICES
• WEIRS
• OUTLETS
• XSECTIONS
• TRANSECTS
• LOSSES
• CONTROLS
• POLLUTION
• LANDUSE
• BUILDUP
• WASHOFF
• COVERAGE
• INFLOW
• DWF
• PATTERN
• RDII
• HYDROGRAPH
• LOADING
• TREATMENT
• CURVES
• TIMESERIES
• REPORT
• MAP
• COORDINATES
• VERTICES
• POLYGONS
• SYMBOLS
• LABELS
• BACKDROP
• TAGS
• PROFILES
• LID_CONTROLS
• LID_USAGE
• GWF
• ADJUSTMENT
• EVENT
• STREETS (new in SWMM 5.2)
• INLETS (new in SWMM 5.2)
• INLET_USAGE (new in SWMM 5.2)

[BuczynskiRafal]

Hi,
I think I am using swmmio incorrectly or there was some issue in the code at the "Polygons" section.
The event occurs when using the "replace_inp_section" function.
The code I am executing:

def _add_coords(self, subcatchment_id):
    exist = [
        (self.model.inp.polygons["X"][-1], self.model.inp.polygons["Y"][-1]),
        (self.model.inp.polygons["X"][-2], self.model.inp.polygons["Y"][-2]),
        (self.model.inp.polygons["X"][-3], self.model.inp.polygons["Y"][-3]),
        (self.model.inp.polygons["X"][-4], self.model.inp.polygons["Y"][-4]),
    ]
    coords = pd.DataFrame(
        data={
            "X": [exist[0][0], exist[1][0], exist[2][0], exist[3][0]],
            "Y": [exist[0][1] - 5, exist[1][1] - 5, exist[2][1] - 5, exist[3][1] - 5],
        },
        index=[subcatchment_id for _ in range(4)]
    )
    coords.index.names = ['Name']
    self.model.inp.polygons = pd.concat([self.model.inp.polygons, coords])
    replace_inp_section(self.model.inp.path, "[Polygons]", self.model.inp.polygons)

I see that SWMM writes literally "Polygons" in the *inp file, swmmio modifies this name by writing "POLYGONS".

Here is the section I have in the file before executing the code

[Polygons]
;;Subcatchment   X-Coord            Y-Coord           
;;-------------- ------------------ ------------------
S1               777180.000         592590.000        
S1               777180.000         592585.000        
S1               777175.000         592585.000        
S1               777175.000         592590.000        

This is what the section looks like after the first execution of "replace_inp_section"

[POLYGONS]
;;   X          Y         
S1   777180.0   592590.0
S1   777180.0   592585.0
S1   777175.0   592585.0
S1   777175.0   592590.0
S2   777175.0   592585.0
S2   777175.0   592580.0
S2   777180.0   592580.0
S2   777180.0   592585.0

This is what the sections looks like after the second execution of "replace_inp_section"

[POLYGONS]
;;   X          Y         
S1   777180.0   592590.0
S1   777180.0   592585.0
S1   777175.0   592585.0
S1   777175.0   592590.0
S2   777175.0   592585.0
S2   777175.0   592580.0
S2   777180.0   592580.0
S2   777180.0   592585.0

[POLYGONS]
;;   X          Y         
S1   777180.0   592590.0
S1   777180.0   592585.0
S1   777175.0   592585.0
S1   777175.0   592590.0
S2   777175.0   592585.0
S2   777175.0   592580.0
S2   777180.0   592580.0
S2   777180.0   592585.0
S3   777180.0   592580.0
S3   777180.0   592575.0
S3   777175.0   592575.0
S3   777175.0   592580.0

In the code I execute I always use "[Polygons]", after the second execution of "replace_inp_section" the sections are duplicated.
The problem doesn't occur when I change the header to "[POLYGONS]" before executing the code, and in my function I also use "[POLYGONS]"

If I should post this thread in discussions I apologize and please delete.

[aerispaha]

@BuczynskiRafal this is very interesting - I suspect that you've uncovered a bug. That said, I think we should continue this conversation in a new issue. Do you mind creating an issue and restating what you've described above?

If you've found a solution already (e.g. by changing the header to [POLYGONS]), please feel free to submit a pull request with your proposed changes.

Thanks for finding and documenting this issue so clearly!

[BuczynskiRafal]

I have created a separate issue, at my spare time I will prepare a PR in which I will propose a solution.

[aerispaha]

One hurdle in providing coverage for the rest of the INP sections is a lack of example models that can be used for testing.

Others are probably aware of this, but I recently realized we can use some of GitHub's fancier search features to find examples of SWMM model INP files with some of these less-commonly-used sections. For example, we can find some INPs that use the TREATMENT section with a search query like this:

path:.inp [TREATMENT]

[aerispaha]

The NCIMM-Black-White-Box repo also seems to be an excellent resource for sample SWMM models and data files. There are also some particularly big models in there that would be nice to use for documentation and examples.

For example, here is a plot of the 10070_H_Elements.inp model:

m = swmmio.Model('10070_H_Elements.inp')
m.links.geodataframe.plot('Geom1', linewidth=m.links.dataframe['Geom1']*0.5, capstyle='round', figsize=(10,10))