Summary
-------

Controller that gives users rql read/ write capabilities.


Sample usage
------------

Users of this service must perform a HTTP POST request to its endpoint,
that is the base url of the CubicWeb application instance appended with
the "rqlio/1.0" url path.

The posted data must use the application/json MIME type, and contain a list of
pairs of the form `(rql_string, rql_args)`, where:

* `rql_string` is any valid RQL query that may contain mapping keys with
  their usual form

* `rql_args` is a dictionary, whose keys are the mapping keys from
  `rql_string`, and the values can be:

  - actual values

  - string references to a previous RQL query's result, with the
    assumption that the referenced RQL query returns a single line and
    single column rset; under such conditions, a string reference
    must be "__rXXX" where `XXX` is the (0-based) index of the RQL query in
    the json-encoded list of queries.

The HTTP request's response (in case where there is no error), is a
json-encoded list.  Its length is the number of RQL queries in the request,
and each element contains the json-encoded result set rows from the
corresponding query.

In case of an error, a json object with a `reason` key will explain the
problem.

Python client example using python-requests::

    import requests
    import json

    args = [('INSERT CWUser U: U login %(l)s, U upassword %(p)s',
            {'l': 'Babar', 'p': 'cubicweb rulez & 42'}),

           ('INSERT CWGroup G: G name "pachyderms"', {}),

           ('SET U in_group G WHERE U eid %(u)s, G eid %(g)s',
            {'u': '__r0', 'g': '__r1'})
           ]

    resp = requests.post('https://myinstance.example.com/rqlio/1.0'),
                         data=json.dumps(args),
                         headers={'Content-Type': 'application/json'})
    assert resp.status_code == 200

