The CSV validator is a web application to validate CSV data against Table Schema. The application provides a fully reusable core that requires only configuration to determine the supported specifications, configured validation types and other validator customisations. The web application allows validation via:
- A SOAP web service API for contract-based machine-machine integrations.
- A REST web service API for machine-machine integrations.
- A web form for validation via user interface.
The SOAP web service API conforms to the GITB validation service API which makes it usable as a building block in GITB Test Description Language (TDL) conformance test cases for the verification of content (as a verify step handler). Note additionally that the validator can also be used to build a command-line tool as an executable JAR with pre-packaged or provided configuration.
This validator is maintained by the European Commission's DIGIT and specifically the Interoperability Test Bed, a conformance testing service for projects involved in the delivery of cross-border public services. Find out more here.
Usage and configuration of this validator is documented as a step-by-step tutorial in the Test Bed's CSV validation guide.
The validator's key principle is that the software is built as a generic core that can be configured to validate any supported specifications. Configuration is organised in domains which represent logically separate setups supported by the same application instance. Each such domain defines the offered validation types and their options along with the validation artefacts needed to carry out validation (local, remote or user-provided). A domain's configuration is grouped in a folder that contains a configuration property file along with any other necessary resources.
When built from source, the simplest way to get started using the validator is to use the all-in-one executable JAR
built from the csvvalidator-war module. You can use and configure this as described in the
validator installation guide.
If you do not plan on modifying the validator's source you can reuse the Test Bed's provided packages. Specifically:
- Via Docker, using the isaitb/csv-validator image from the Docker Hub.
- Via JAR file, using the generic web application package.
- Via command line tool, using the generic command line tool package.
It is interesting to note that the second option (executable web application JAR) matches what you would build from this
repository. The command line package is produced from the csvvalidator-jar although this requires an additional step
of JAR post-processing to configure the validator's domain(s).
Once the validator's web application is up you can use it as follows:
- SOAP API: http://localhost:8080/csv/soap/DOMAIN/validation?wsdl
- REST API: http://localhost:8080/csv/DOMAIN/api/validate (Swagger docs at http://localhost:8080/csv/swagger-ui/index.html)
- Web form: http://localhost:8080/csv/DOMAIN/upload
Note that the DOMAIN placeholder in the above URLs is the name of a domain configuration folder beneath your configured validator.resourceRoot.
This can be adapted by providing validator.domainName.DOMAIN mapping(s) for your domain(s). These are application-level configuration properties
that can be set in the default application.properties
or via environment variables and system properties.
The validator is a multi-module Maven project from which the artefact to use is the web application package, produced
from module csvvalidator-war. This is an all-in-one Spring Boot web application. To build issue mvn clean install.
To build this project's libraries you require:
- A JDK installation (17+).
- Maven (3+)
- Locally installed itb-commons dependencies (see below).
Building this validator from source depends on libraries that are available on public repositories. The exception is currently the set of itb-commons dependencies, common libraries that are shared by all Test Bed validators. To be able to build you need to first clone itb-commons and install its artefacts in your local Maven repository.
When building for development you should provide the validator's basic configuration to allow it to bootstrap. The simplest approach to do this is to use environment variables that set the validator's configuration properties.
The minimum properties you should define this way are:
validator.resourceRoot: The root folder from which all domain configurations will be loaded.logging.file.path: The validator’s log output folder.validator.tmpFolder: The validator’s temporary work folder.
In addition, you should include within the validator.resourceRoot folder additional folder(s) for your configuration domains,
each with its configuration property file and any other needed resources. A simple example of such configuration that you
can also download and reuse, is provided in the CSV validation guide's configuration step.
If you decide not to load any configurations, the generic CSV validator configuration will be loaded instead. This generic configuration does not contain any validation artefacts and will instead allow you to upload your own validation schemas from the validator's UI and APIs. The domain of this generic validator is called any. you can access it like any other domain, from http://localhost:8080/csv/any/upload.
If you choose Docker to run your validator you can use the sample Dockerfile as a starting point. To use this:
- Create a folder and copy within it the Dockerfile and JAR produced from the
csvvalidator-warmodule. - (Optional) Create a sub-folder (e.g.
resources) as your resource root within which you place your domain configuration folder(s). - (Optional) Adapt the Dockerfile to also copy the
resourcesfolder and set its path within the image as thevalidator.resourceRoot:
...
COPY resources /validator/resources/
ENV validator.resourceRoot /validator/resources/
...
Note: If you skip step 2 and 3, only the generic validator will be available under the any domain.
The CSV validator supports custom plugins to extend the validation report. Plugins are implementations of the GITB validation service API for which the following applies. Note that plugin JAR files need to be built as "all-in-one" JARs.
The CSV validator calls plugins in sequence passing in the following input:
| Input name | Type | Description |
|---|---|---|
contentToValidate |
String |
The absolute and full path to the input provided to the validator. |
domain |
String |
The validation domain relevant to the specific validation call. |
validationType |
String |
The validation type of the domain that is selected for the specific validation call. |
tempFolder |
String |
The absolute and full path to a temporary folder for plugins. This will be automatically deleted after all plugins complete validation. |
hasHeaders |
String |
true or false depending on whether the input should be considered as having a header row. |
delimiter |
String |
The character to use as the field delimiter. |
quote |
String |
The character to use for the quote character. |
locale |
String |
The locale (language code) to use for reporting of results (e.g. "fr", "fr_FR"). |
The output of plugins is essentially a GITB ValidationResponse that wraps a TAR instance. The report items within this TAR instance are merged
with any reports produced by Table Schema validation.
Plugin configuration for a validator instance is part of its domain configuration.
Once you have your plugins implemented you can configure them using the validator.defaultPlugins
and validator.plugins properties where you list each plugin by providing:
- The path to its JAR file.
- The fully qualified class name of the plugin entry point.
This software is shared using the European Union Public Licence (EUPL) version 1.2.
The authors of this library waive any and all liability linked to its usage or the interpretation of results produced by its downstream validators.
For feedback or questions regarding this library you are invited to post issues in the current repository. In addition, feel free to contact the Test Bed team via email at [email protected].
The Test Bed provides similar validators for other content types. Check these out for more information:
- The RDF validator: see overview, source, detailed guide and Docker Hub image.
- The XML validator: see overview, source, detailed guide and Docker Hub image.
- The JSON validator: see overview, source, detailed guide and Docker Hub image.