sfvalidator

In-browser dataset validation.

Goals

Validating data sets against the rules listed here within a terminal or browser.

Legal

• Check license: Look for DCAT licenses

Technical

• Fetches the page using fetch-ponyfill Check for Cache-Control, ETag HTTP headers.

• CORS detection: Is only in header if CORS is supported. Detect Access-Control-Allow-Origin HTTP header.

Syntactic

• Check if RDF1.1: Parse with ldfetch

Semantic

• Maximum use of standards

Querying

• Fragmented data: Check for hydra links and timestamps

• European DCAT-AP standard

Usage

Node

First install the package:

npm install sfvalidator

Then require it and start validating:

constsfv=require("sfvalidator")sfv.validate_url("https://www.example.com").then(result=>{console.log(result);}

Terminal

Install the package globaly:

npm install -g sfvalidator

Now you are able to run:

sfvalidator <URL>
sfvalidator --file <file-path>

In-browser

1. Firstly install or download the module.

2. Secondely install browserify:

    npm install -g browserify
   

3. Then use browserify to bundle the module and its dependencies:

    browserify <path-to-module>/lib/browserify-validator.js -o bundle.js
   

4. Now you are able to include this bundle as a script in your webpage and use it:

    <script src="bundle.js"></script>
   <script>
   validator = new window.sfvalidator();
   validator.validate_url("https:://www.example.com").then(result => { console.log(result); }
   </script>
   

See the example folder for a more extensive example.

Output

{Accessible: {first_attempt: Score,seconde_attempt: Score},License: Score,Headers: {Cache: Score,ETag: Score,Cors: Score},Rdf: Score,Fragmented: Score,Timestamped: Score}

A Score is of this structure:

{
score: int->-1iffailed,0ifnotcheckedor1ifpassed
message: Object->themessagefromthevalidator}

The message attribute can have different values when the element passed the validation:

• In case of accessable, rdf and timestamped it is just a string.
• In case of the headers it is the value of the header in the returned package.
• In case of license and fragmented it is a list of found licenses and hydra links respectfully.

Examples

Terminal

Run ./bin/validate.js https://linked.open.gent/parking.

Or run ./bin/validate.js --file example/url_examples.txt

In-browser

Run npm run-script build-example and then run an http server in the example/in-browser folder.

Function Documentation

• validate_url: Use this function to validate a url, returns a report. Takes a url as argument in the form of a string.
• validate_headers: This function validates the headers of the package the server sends back. Takes a url as input and returns a promise with a report. This function gets called by validate_url.
• validate_rdf: This function tries to parse the triples present in the data fetched from the url. Takes a url as argument and returns a promise that returns a report. Gets called by validate_url.
• validate_license: Checks if there is a license on the data within the triples. Takes triples and the url as input and returns a score. This function gets called by validate_rdf.
• validate_fragmentation: Searches for hydra links within the triples of the data that have the url as subject. Takes triples and the url as input and returns a score. Gets called by validate_rdf.
• validate_metadata: Checks for the precense of triples with DCT or DCAT metadata predicates about the url. Takes triples and the url as input and returns a score. Gets called by validate_rdf.
• validate_timestamped: Checks if the file is timestamped. Takes triples and the url as input and returns a score. Gets called by validate_rdf.