cjio, or CityJSON/io

Python CLI to process and manipulate CityJSON files. The different operators can be chained to perform several processing operations in one step, the CityJSON model goes through them and different versions of the CityJSON model can be saved as files along the pipeline.

Documentation

cjio.readthedocs.io

Installation

It uses Python 3.7+ only.

To install the latest release:

pip install cjio

Note

The commands export, triangulate, reproject, and validate require extra packages that are not install by default. You can install these packages by specifying the commands for pip.

pip install 'cjio[export,reproject,validate]'

To install the development branch, and still develop with it:

git checkout develop
virtualenv venv
. venv/bin/activate
pip install --editable '.[develop]'

Note for Windows users

If your installation fails based on a pyproj or pyrsistent error there is a small hack to get around it. Based on the python version you have installed you can download a wheel (binary of a python package) of the problem package/s. A good website to use is here. You then run:

pip install [name of wheel file]

You can then continue with:

pip install cjio

Supported CityJSON versions

Currently it supports CityJSON v2.0, but v1.1 and v1.0 files can be upgraded automatically with the operator upgrade`.

The operators (cjio --version) expect that your file is using the latest version CityJSON schema. If your file uses an earlier version, you can upgrade it with the upgrade operator: cjio old.json upgrade save newfile.city.json

Usage of the CLI

After installation, you have a small program called cjio, to see its possibilities:

cjio --help

Commands:
  attribute_remove  Remove an attribute.
  attribute_rename  Rename an attribute.
  crs_assign        Assign a (new) CRS (an EPSG).
  crs_reproject     Reproject to a new EPSG.
  crs_translate     Translate the coordinates.
  export            Export to another format.
  info              Output information about the dataset.
  lod_filter        Filter only one LoD for a dataset.
  materials_remove  Remove all materials.
  merge             Merge the current CityJSON with other ones.
  metadata_create   Add the +metadata-extended properties.
  metadata_get      Shows the metadata and +metadata-extended of this...
  metadata_remove   Remove the +metadata-extended properties.
  metadata_update   Update the +metadata-extended.
  print             print the (pretty formatted) JSON to the console.
  save              Save to a CityJSON file.
  subset            Create a subset, City Objects can be selected by: (1)...
  textures_locate   Output the location of the texture files.
  textures_remove   Remove all textures.
  textures_update   Update the location of the texture files.
  triangulate       Triangulate every surface.
  upgrade           Upgrade the CityJSON to the latest version.
  validate          Validate the CityJSON: (1) against its schemas (2)...
  vertices_clean    Remove duplicate vertices + orphan vertices

Or see the command-specific help by calling --help after a command:

Usage: cjio INPUT subset [OPTIONS]

Create a subset, City Objects can be selected by: (1) IDs of City Objects;
(2) bbox; (3) City Object type(s); (4) randomly.

These can be combined, except random which overwrites others.

Option '--exclude' excludes the selected objects, or "reverse" the
selection.

Usage examples:

  cjio myfile.city.json subset --bbox 104607 490148 104703 490257 save out.city.json
  cjio myfile.city.json subset --radius 500.0 610.0 50.0 --exclude save out.city.json
  cjio myfile.city.json subset --id house12 save out.city.json
  cjio myfile.city.json subset --random 5 save out.city.json
  cjio myfile.city.json subset --cotype LandUse --cotype Building save out.city.json

Options:
  --id TEXT          The ID of the City Objects; can be used multiple times.
  --bbox FLOAT...    2D bbox: minx miny maxx maxy.
  --radius FLOAT...  x y radius
  --random INTEGER   Number of random City Objects to select.
  --cotype TEXT      The City Object types; can be used multiple times.
  --exclude          Excludes the selection, thus delete the selected
                     object(s).
  --help             Show this message and exit.

Pipelines of operators

The input 3D city model opened is passed through all the operators, and it gets modified by some operators. Operators like info and validate output information in the console and just pass the 3D city model to the next operator.

cjio example.city.json subset --id house12 remove_materials save out.city.json
cjio example.city.json remove_textures info
cjio example.city.json upgrade validate save new.city.json
cjio myfile.city.json merge '/home/elvis/temp/*.city.json' save all_merged.city.json

stdin and stdout

Starting from v0.8, cjio allows to read/write from stdin/stdout (standard input/output streams).

For reading, it accepts at this moment only CityJSONL (text sequences with CityJSONFeatures). Instead of putting the file name, stdin must be used.

For writing, both CityJSON files and CityJSONL files can be piped to stdout. Instead of putting the file name, stdout must be used. Also, the different operators of cjio output messages/information, and those will get in the stdout stream, to avoid this add the flag --suppress_msg when reading the file, as shown below.

cat mystream.city.jsonl | cjio --suppress_msg stdin remove_materials save stdout
cjio --suppress_msg myfile.city.json remove_materials export jsonl stdout | less
cat myfile.city.json | cjio --suppress_msg stdin crs_reproject 7415 export jsonl mystream.txt

Generating Binary glTF

Convert the CityJSON example.city.json to a glb file /home/elvis/gltfs/example.glb

cjio example.json export glb /home/elvis/gltfs

Convert the CityJSON example.city.json to a glb file /home/elvis/test.glb

cjio example.city.json export glb /home/elvis/test.glb

Usage of the API

cjio.readthedocs.io/en/stable/tutorials.html

Docker

If docker is the tool of your choice, please read the following hints.

To run cjio via docker simply call:

docker run --rm  -v <local path where your files are>:/data tudelft3d/cjio:latest cjio --help

To give a simple example for the following lets assume you want to create a geojson which represents the bounding boxes of the files in your directory. Lets call this script gridder.py. It would look like this:

from cjio import cityjson
import glob
import ntpath
import json
import os
from shapely.geometry import box, mapping

def path_leaf(path):
    head, tail = ntpath.split(path)
    return tail or ntpath.basename(head)

files = glob.glob('./*.json')

geo_json_dict = {
    "type": "FeatureCollection",
    "features": []
}

for f in files:
    cj_file = open(f, 'r')
    cm = cityjson.reader(file=cj_file)
    theinfo = json.loads(cm.get_info())
    las_polygon = box(theinfo['bbox'][0], theinfo['bbox'][1], theinfo['bbox'][3], theinfo['bbox'][4])
    feature = {
        'properties': {
            'name': path_leaf(f)
        },
        'geometry': mapping(las_polygon)
    }
    geo_json_dict["features"].append(feature)
    geo_json_dict["crs"] = {
        "type": "name",
        "properties": {
            "name": "EPSG:{}".format(theinfo['epsg'])
        }
    }
geo_json_file = open(os.path.join('./', 'grid.json'), 'w+')
geo_json_file.write(json.dumps(geo_json_dict, indent=2))
geo_json_file.close()

This script will produce for all files with postfix ".json" in the directory a bbox polygon using cjio and save the complete geojson result in grid.json in place.

If you have a python script like this, simply put it inside your local data and call docker like this:

docker run --rm  -v <local path where your files are>:/data tudelft3d/cjio:latest python gridder.py

This will execute your script in the context of the python environment inside the docker image.

Example CityJSON datasets

There are a few example files on the CityJSON webpage.

Alternatively, any CityGML file can be automatically converted to CityJSON with the open-source project citygml-tools (based on citygml4j).

Acknowledgements

The glTF exporter is adapted from Kavisha's CityJSON2glTF.