convert complete service

background: https://github.com/timrdf/csv2rdf4lod-automation/wiki/A-quick-and-easy-conversion was done years ago. It sets up shop from an RDFa encoding of the conversion parameters.

client URL: http://gemini.tw.rpi.edu/dev/elixir/services/start

This accepts UI changes and reflects the results in the DOM's RDFa. This also uses jqueryrdf to parse itself for RDFa and POST to:

service URL: http://gemini.tw.rpi.edu/dev/elixir/services/convert-complete

accept POST:

Include directly (we're using this approach now):

<http://logd.tw.rpi.edu/source/data-gov/dataset/4383/version/2011-Jan-24>
    a dcat:Dataset;
    dcterms:source <http://explore.data.gov/download/wfna-38ey/XLS> ;
    a conversion:VersionedDataset .
    conversion:base_uri           "http://logd.tw.rpi.edu" ;
    conversion:source_identifier  "data-gov" ;
    conversion:dataset_identifier "4383" ;
    conversion:version_identifier "2011-Jan-24" ;
    # we need delimiter

If Linked data and provides appropriate access information (in dcat or dcterms:source):

    a dcat:Dataset;

• set up directory structure
• retrieve URL into source
• create conversion trigger
• convert raw
• publish raw to endpoint

response:

a RetrievedDataset
conversion:source_identifier
conversion:dataset_identifier
conversion:version_identifier
dcterms:source 
some provenance and void that associates the dataset to a named graph in an endpoint.

RePOSTing will reperform the same operations (as it it hadn't been done before) UNLESS there is some indication that has been successful (e1 exists ==> successful)

Possible errors that we'll need to handle:

• csv file does not exist (404)
• the e1 of versioned dataset already exists

client follows that provenance and queries the raw named graph to populate first 10 rows.

UI to let user select different/additional exemplars

dom tricks to build RDFa tree

parse RDFa from dom and POST to convert sample service