QWC ConfigGenerator¶

Generates JSON files for service configs and permissions from WMS GetCapabilities, QGS projects and QWC ConfigDB.

This service is integrated into qwc-docker, consult qwc-services.github.io for the general qwc-services documentation.

Setup¶

Create a ConfigGenerator config file tenantConfig.json below INPUT_CONFIG_PATH (see below) for each tenant.

Configuration¶

• JSON schema
• Example tenantConfig.json

NOTE: the Search service configuration takes its resources directly from the ConfigGenerator configuration. Its Permissions are collected from the ConfigDB (solr_facet resources), unless they are defined in the ConfigGenerator configuration.

Environment variables¶

Config options in the config file can be overridden by equivalent uppercase environment variables.

In addition, the following environment variables are supported:

Name	Default	Description
INPUT_CONFIG_PATH	config-in/	Base directory where the input configuration files are located.
OUTPUT_CONFIG_PATH	/tmp/	Base directory where generated service and permission configuration files are located.
JSON_SCHEMAS_PATH	/tmp/	Directory where service schemas are loaded.

Schema validation¶

By default, the ConfigGenerator will validate the service configurations in tenantConfig.json against the schema definition of the service. The JSON Schemas are loaded from local files in JSON_SCHEMAS_PATH, or else downloaded from https://github.com/qwc-services/ if no schema files are present. You can disable the schema validation by setting "validate_schema": false in the ConfigGenerator's config block in tenantConfig.json.

You can change the versions of the schemas that the ConfigGenerator uses for verification inside schema-versions.json.

A helper script to download the JSON schemas of all QWC services registered in schema-versions.json is available at src/download_json_schemas.py.

Custom resource types¶

If you want to define custom resource types for a custom service, you can add a record for the resource type to the configdb

INSERT INTO qwc_config.resource_types(name, description, list_order) values ('<resource_name>', '<resource_description>', <list_order>);

and then add it to the custom_resource_types setting.

Additional services¶

For any additional service (without specific resources), ConfigGenerator generates the configuration in OUTPUT_CONFIG_PATH directory.

Add the following configuration and adapt it to your service in tenantConfig.json:

{
    "name": "<service_name>",
    "schema_url": "<service_schema_url>",
    "config": {...}
}

Note: service_name is expected to be camel case (i.e. adminGui), and the service name in the generated config will lowercase and hyphenated (i.e. admin-gui).

Usage¶

CLI interface¶

Show command options:

uv run src/config_generator_cli.py --help

Generate both service configs and permissions:

uv run src/config_generator_cli.py ./tenantConfig.json all

Generate service config files:

uv run src/config_generator_cli.py ./tenantConfig.json service_configs

Generate permissions file:

uv run src/config_generator_cli.py ./tenantConfig.json permissions

Additionally, the following command line args may be specified:

• --use_cached_project_metadata=1: Whether to use cached project metadata
• --force_readonly_datasets=1: Whether to force read-only dataset permissions

Service interface¶

Install dependencies and run:

# Setup venv
uv venv .venv

export INPUT_CONFIG_PATH=<INPUT_CONFIG_PATH>
export OUTPUT_CONFIG_PATH=<INPUT_CONFIG_PATH>
uv run src/server.py

Set FLASK_DEBUG=1 for additional debug output.

Set FLASK_RUN_PORT=<port> to change the default port (default: 5000).

Generate both service configs and permissions for default tenant:

# Non-blocking call, returns task_id
curl "http://localhost:5000/generate_configs?tenant=default"

# Blocking call, streams log output
curl "http://localhost:5000/generate_configs?tenant=default&stream_response=1"

Additionally, the following query parameters may be specified:

• use_cached_project_metadata=1: Whether to use cached project metadata
• force_readonly_datasets=1: Whether to force read-only dataset permissions

Watch config generator task status:

curl "http://localhost:5000/generate_configs_status?task_id=<task_id>"

Cancel config generator task:

curl "http://localhost:5000/generate_configs_cancel?task_id=<task_id>"

Docker usage¶

The Docker image is published on Dockerhub.

See sample docker-compose.yml of qwc-docker.