Documentation

The Cityflows road cutter is a service that is part of the Cityflows ecosystem. It is designed to help in the preprocessing of data before it can beprocessed by the cityflows model. It takes input from a street network and a list of polygons, then labels the steets with the overlapping polygons. The GIS procedures are performed in a POSTGIS database. More information on how the road cutter operates can be found in the procedure steps.

Launch a PostGIS instance locally

docker run --name some-postgis -e POSTGRES_DB=postgis -e POSTGRES_USER=postgis -e POSTGRES_PASSWORD=rootpass -e POSTGIS_GEOS_VERSION=3.8.0 -p "5432:5432" -d mdillon/postgis

alternatively, you can use another base image (which hase a newer version of GEOS installed)

docker run --name some-postgis -e POSTGRES_PASSWD=rootpass -p "5432:5432" -d geographica/postgis

You can then connect to the DB using these credentials:

    host: "localhost",
    port: 5432,
    user: "postgres",
    password: "rootpass",
    database: "postgres",

(if you have an issue where you can't create a table because type geometry doesn't exist, try running CREATE EXTENSION postgis;)

If you change some of the env variables, make sure to reflect the change in src/index.js where the knex instance is created.

Input data

The inputdata is defined in the config.js file, there a inputfile is determined for the countGeometries and for the segments. The files you specify are required to be available in the following locations:

data/COUNTGEOMETRIES_INPUTFILE
data/managed_data_files/shapefiles/SEGMENTS_INPUTFILE

Where to get this data

1. CountGeometries:
   This is the result of the data preparation in the data model repository. Where first the data scraping/collecting needs to be done then the projecting to the correct format. Lastly all the different datasources need to be merged using src/tooling/merge/main.py. If the resulting file is large (most often because of a large area of interest) then it's recommended to do an additional step: run the tools/all_shapes.py script. This extra step will reduce the file to the unique geometries which is sufficient for the roadcutter procedure (NOT for the data model!).
2. segments:
   This is an extract from the flemish roadregister. To get this you first need to download it: here (look for wegenregister) and then run the notebook notebooks/extract_wegenregister.ipynb. This results in an useable formatted inputfile.

Insert the segments and counts geometries

Execute

node src/index.js

Note that this will drop and recreate the road_cutter schema and all the temporary tables at every execution.

Visualize the results

QGIS is a nice visualization software for spatial data. It can be used easily to configure a connection to the PostGIS database and visualize any geometry from the work tables or event the outputs of the whole cutting procedure.

Troubleshouting

Previous issues:

• wrong formatting of input data:
  • polygons (all_ref data):
    • some geometries are multipolygons. This will throw
    • check naming of column 'index', 'gridIndex' or 'grid_index'. This may vary.
    • ID's are sometimes not unique. Currently they are overwritten by the loop index.

Make sure in postgis that the polygons table has id's and indexes. If this is null, the query will fail down the road. Make sure you are checking the correct database ;)

Connected repositories in the Cityflows ecosystem

• the data model: repository
• weighted distribution: repository