csvw: omit "tables" in JSON output

[ktk]

We do not support the tables part of the CSVW spec in our own CSVW parser, as our parser is a pure streaming parser and we cannot do some of the things in the specification where tablesmakes sense when we are streaming data.

I don't mind if we leave it in for other CSVW processors but I need at least an optional way to get rid of that nesting in the output. I see two ways of doing that:

• check if there is more than one file mentioned in the XRM file and only add tables nesting when there is more than one file. Like that I can solve the problem by simply creating a unique mapping file per CSV file
• add additional functionality to the dialect part (that's where I think it makes sense at least) that would create a new CSVW mapping file per output without table so we could still have more than one file defined per XRM source. Not sure how we should generate the output filename in that case though

[mchlrch]

Let's do the first one: Only add the "tables": [ ] nesting in the output, if there is more than one map in one .xrm file

Generated output, if there's only one map:

{
    "@context": "http://www.w3.org/ns/csvw",
    "url": "http://www.example.com/departments.csv",
    "tableSchema": {
        "aboutUrl": "http://www.example.com/department/{DEPTNO}",
        "columns": [...]
    }
}

[nnamtug]

simple change in one commit -> pushing this directly to master

[mchlrch]

Works as expected. Closing this.