Add getter and setter for tags.

[BuczynskiRafal]

Hi,
I needed to get the tags from the inp file so I added a getter and setter to this section.
For the class "inp" they work fine. For the class "Model" there is a problem reading the file.

@aerispaha can you suggest where could be the problem?

This is the output from test_model_tags (model.tags.dataframe)

Empty DataFrame
Columns: []
Index: [Subcatch, Subcatch, Subcatch, Subcatch]

It looks like it reads the names correctly.

[aerispaha]

@BuczynskiRafal I think the reason this fails is because we don't actually have a tags property in the swmmio.Model object.

The Model object is a higher level abstraction whose properties bring together multiple related tables in the inp file. For example, the Model.nodes property combines the [JUNCTIONS], [OUTFALLS], and [STORAGE] tables. However, I don't think this higher-level abstraction is necessary in the case of tags. Is there a use-case for this that I'm not thinking of?

That said, instead of adding a tags property to the swmmio.Model object, we could certainly include tags within the Model.nodes section. This way, we when we want fetch all the relevant information about nodes in our models, we can include any tags that may be assigned in the dataframe that is returned by Model.nodes.dataframe.

Does that make sense to you?

[BuczynskiRafal]

Yes, I agree with you. I left the setter and getter for the inp class.

I'm don't know how to solve tag assignment using the approach you propose i.e. by referencing the key from the "section_headers.yml" file.
The tag section is indexed by the element type and not by its name/id.

[TAGS]
Subcatch  CA-1             CA              
Subcatch  CA-7             CA              
Subcatch  CA-8             CA              
Subcatch  CA-11            CA
Link      P001             CA

I modified the yaml

composite:
  nodes:
    inp_sections: [junctions, outfalls, storage, tags]
    rpt_sections: [Node Depth Summary, Node Flooding Summary, Node Inflow Summary]
    columns: [InvertElev, MaxDepth, InitDepth, SurchargeDepth, PondedArea,
               OutfallType, StageOrTimeseries, TideGate, MaxD, StorageCurve,
               Coefficient, Exponent, Constant, EvapFrac, SuctionHead,
               Conductivity, InitialDeficit, coords, Tag]

output

          InvertElev  MaxDepth  ...                coords  Tag
Link             NaN       NaN  ...                  None   CA
Outlet          0.00       NaN  ...      [(500.0, 500.0)]  NaN
P001            1.10      20.0  ...      [(100.0, 100.0)]  NaN
P005            1.58       NaN  ...      [(200.0, 200.0)]  NaN
P009            3.60      20.0  ...  [(498.395, 215.716)]  NaN
P011            2.89      20.0  ...      [(400.0, 400.0)]  NaN
Subcatch         NaN       NaN  ...                  None   CA
Subcatch         NaN       NaN  ...                  None   CA
Subcatch         NaN       NaN  ...                  None   CA
Subcatch         NaN       NaN  ...                  None   CA

This worked with conduits method

        tags = dataframe_from_inp(inp.path, "TAGS")
        if "Link" in set(tags.index):
            df = df.merge(dataframe_from_inp(inp.path, "Tags"), left_on="Name", right_on="Name", how="left")

output

   Name InletNode  ... SlopeFtPerFt  Tag
0  P001      P001  ...        0.005   CA
1  P005      P005  ...        0.002  NaN
2  P009      P009  ...        0.004  NaN
3  P011      P011  ...        0.002  NaN

[aerispaha]

@BuczynskiRafal thanks so much for this contribution. I have a couple minor comments, but I'm going to approve this PR. I'll work on getting the new tags feature released in a couple hours.

[aerispaha]

Because you've added tags to the inp object, you can simplify this line to

tags = inp.tags

[aerispaha]

Instead of re-instantiating a tags dataframe, you could reuse the tags dataframe created on line 216, right? Then this line could be simplified to

df = df.merge(tags, left_on="Name", right_on="Name", how="left")

[aerispaha]

perfect!

[aerispaha]

This makes progress on #57 by adding coverage for [TAGS]. Thanks again @BuczynskiRafal!