Catchment structure for the Lachlan valley¶

Network definition for a shapefile with information on the Lachlan valley, as of ~8th August 2025. This is use case for customisable column names as well as being more liberal with the types of columns in the geopandas data frame resulting from reading the shapefile.

In [1]:

from pathlib import Path

import geopandas as gpd

from geosdhydro import ShapefileToSwiftConverter

from pathlib import Path import geopandas as gpd from geosdhydro import ShapefileToSwiftConverter

In [2]:

fpath = Path.home() / "data"/"wnsw"/"Lachlan"/"Lachlan_links4swift.shp"

fpath = Path.home() / "data"/"wnsw"/"Lachlan"/"Lachlan_links4swift.shp"

In [3]:

fpath.exists()

fpath.exists()

Out[3]:

True

In [4]:

link_specs = gpd.read_file(fpath)

link_specs.head()

link_specs = gpd.read_file(fpath) link_specs.head()

Out[4]:

	OBJECTID	Shape_Leng	LinkID	FromNodeID	ToNodeID	HeadLink	SPathCnt	LPathCnt	SPathLen	LPathLen	...	Area_sqkm	SHORTNAME	Elevation	Shape_Le_1	Shape_Le_2	Shape_Area	DArea	NodeID	SubCatID	geometry
0	1	0.146763	1	43408002	43407878	0	23	23	36156.428635	36156.428635	...	76.225305	CADIANGULLONG @ PAN.	6.381	62604.167491	62604.167491	7.622530e+07	7.622146e+07	43408002.0	2.0	LINESTRING (148.97972 -33.5148, 148.8481 -33.5...
1	2	0.125647	2	43332557	43407878	0	22	22	26988.248489	26988.248489	...	190.273377	BELUBULA THE NEEDLES	366.293	303059.700345	303059.700345	1.011929e+09	1.902571e+08	43332557.0	7.0	LINESTRING (148.97361 -33.57389, 148.8481 -33....
2	5	0.038594	5	43323436	43407903	0	3	3	18098.684816	18098.684816	...	360.595819	LACHLAN R @ COWRA	274.300	299445.505950	299445.505950	1.294608e+09	3.606294e+08	43323436.0	70.0	LINESTRING (148.72056 -33.82222, 148.68359 -33...
3	6	0.220645	6	43328702	43407903	0	25	25	40402.031075	40402.031075	...	234.353118	LACHLAN R @ COWRA	274.300	299445.505950	299445.505950	1.294608e+09	2.343552e+08	43328702.0	76.0	LINESTRING (148.87306 -33.94639, 148.68359 -33...
4	7	0.254465	7	43636627	43407903	0	32	32	49606.117846	49606.117846	...	340.169996	HOVELLS CK @ NO.2	323.000	174352.740146	174352.740146	3.401700e+08	3.401840e+08	43636627.0	20.0	LINESTRING (148.87246 -34.00384, 148.68359 -33...

5 rows × 29 columns

In [5]:

{x: link_specs[x].dtype for x in link_specs.columns}

{x: link_specs[x].dtype for x in link_specs.columns}

Out[5]:

{'OBJECTID': dtype('int64'),
 'Shape_Leng': dtype('float64'),
 'LinkID': dtype('int64'),
 'FromNodeID': dtype('int64'),
 'ToNodeID': dtype('int64'),
 'HeadLink': dtype('int32'),
 'SPathCnt': dtype('int64'),
 'LPathCnt': dtype('int64'),
 'SPathLen': dtype('float64'),
 'LPathLen': dtype('float64'),
 'FID_': dtype('int32'),
 'OBJECTID_1': dtype('int64'),
 'OBJECTID_2': dtype('int64'),
 'TAG': dtype('O'),
 'Station': dtype('float64'),
 'HydroID': dtype('int64'),
 'STNAME': dtype('O'),
 'Longitude': dtype('float64'),
 'Latitude': dtype('float64'),
 'Area_sqkm': dtype('float64'),
 'SHORTNAME': dtype('O'),
 'Elevation': dtype('float64'),
 'Shape_Le_1': dtype('float64'),
 'Shape_Le_2': dtype('float64'),
 'Shape_Area': dtype('float64'),
 'DArea': dtype('float64'),
 'NodeID': dtype('float64'),
 'SubCatID': dtype('float64'),
 'geometry': <geopandas.array.GeometryDtype at 0x7f1472d25e80>}

Note that some of the input columns, LinkID, ToNodeID, FromNodeID, are integers, because of habits. It is preferable to have them as strings, but the converter will transparently convert them to string. Another thing is that there is a duplicated ID in the links:

In [6]:

gdf = link_specs
# Check for duplicates in the 'LinkID' column
duplicates = gdf["LinkID"][gdf["LinkID"].duplicated(keep=False)]

# Display the duplicated IDs
print(duplicates)

gdf = link_specs # Check for duplicates in the 'LinkID' column duplicates = gdf["LinkID"][gdf["LinkID"].duplicated(keep=False)] # Display the duplicated IDs print(duplicates)

Series([], Name: LinkID, dtype: int64)

Let's see what happens:

In [7]:

link_specs = link_specs.drop(index=[3,17])

link_specs = link_specs.drop(index=[3,17])

and now we expect the converter to do the job:

In [8]:

converter = ShapefileToSwiftConverter(
    gdf = link_specs,
    include_coordinates = True,
    linkid_field = "LinkID",
    fromnodeid_field = "FromNodeID",
    tonodeid_field = "ToNodeID",
    spathlen_field = "SPathLen",
    darea2_field = "DArea",
    geometry_field = "geometry",
)

converter = ShapefileToSwiftConverter( gdf = link_specs, include_coordinates = True, linkid_field = "LinkID", fromnodeid_field = "FromNodeID", tonodeid_field = "ToNodeID", spathlen_field = "SPathLen", darea2_field = "DArea", geometry_field = "geometry", )

In [9]:

result = converter.convert()

result = converter.convert()

result is a python dictionary

As expected given that some areas were negative in the input file (i.e. links without a contributing subarea), we have less subareas than links

In [10]:

f"there are {len(result['Links'])} links, {len(result['Nodes'])} nodes, {len(result['SubAreas'])} subareas"

f"there are {len(result['Links'])} links, {len(result['Nodes'])} nodes, {len(result['SubAreas'])} subareas"

Out[10]:

'there are 119 links, 122 nodes, 76 subareas'

The object converter has a save_to_file method, or you can use the json module to save the above result:

In [11]:

import json

import json

In [12]:

fp = Path.home() / "tmp" / "lachlan_swift.json"
with open(fp, "w") as f:
    json.dump(result, f, indent=2)

fp = Path.home() / "tmp" / "lachlan_swift.json" with open(fp, "w") as f: json.dump(result, f, indent=2)

Checking the json output loads as a catchment structure¶

This can be done if you have swift2 in your python env using the following.

See load_lachlan.ipynb.

In [13]:

# from swift2.model_definitions import model_from_json_file
# sim = model_from_json_file(fp)

# from swift2.model_definitions import model_from_json_file # sim = model_from_json_file(fp)

Feedback