rdf-parser-csvw

A CSV on the Web parser with RDFJS Stream interface.

⛓️‍💥 Fork of rdf-ext/rdf-parser-csvw converted to ESM, including some tweaks and fixes

Usage

The package exports the parser as a class, so an instance must be created before it can be used. The .import method, as defined in the RDFJS specification, must be called to do the actual parsing. It expects a stream of strings. The method will return a stream which emits the parsed quads.

The constructor accepts an options object with the following optional keys:

• metadata: Use the metadata to convert the CSV to RDF. The metadata must be given as a Dataset using the CSV on the Web ontology. This options is required.
• baseIRI: Use the IRI to create Named Nodes. The value must be a String. This options is required.
• factory: Use an alternative RDFJS data factory. By default the reference implementation us used.
• timezone: Use an alternative timezone to parse date and time values. The value must be given as a String as defined in the Luxon documentation. By default local will be used.
• relaxColumnCount: Don't throw an error if a row has a column count which doesn't match the headers column count.
• skipLinesWithError: Skip lines with error instead of throwing an error and stop parsing. This is mainly useful for debugging and should not be used in production environments.
• skipEmptyLines: Ignore empty lines in the CSV file.
• trimHeaders: Remove whitespace from the beginning and end of the column names.

It's also possible to pass options as second argument to the .import method. The options from the constructor and the .import method will be merged together.

Compliance with the CSVW specification

Dates and times

• Date/Time formats are generally supported with possible issues around the timezone specifier X and the exact formatting of the timezone offset.
• RFC2822 dates are supported via explict 'RFC2822' datatype format (case insensitive).
• Time formats are implemented but the output will always be normalised to UTC.