Add support for bounded graph queries

[tuxdna]

At present BayesianGraph.execute() does not support queries with bounded parameters:

fabric8-analytics-data-model/src/graph_manager.py

Lines 16 to 30 in 13bd745

	def execute(cls, str_gremlin_dsl):
	"""Execute the query prepared for the graph database."""
	logger.debug("BayesianGraph::execute() Gremlin DSL: %s", str_gremlin_dsl)
	payload = {'gremlin': str_gremlin_dsl}
	response = requests.post(config.GREMLIN_SERVER_URL_REST,
	data=json.dumps(payload))
	json_response = response.json()
	logger.debug("BayesianGraph::execute(): %s", response)
	if response.status_code != 200:
	logger.error("ERROR %d(%s): %s" % (response.status_code, response.reason,
	json_response.get("message")))
	return False, json_response
	else:
	return True, json_response

Support for bounded parameters will help us write better queries and avoid many bugs.

[tuxdna]

Bounded parameters work similar to prepared statements in SQL. Here is an example:

[miteshvp]

Its a matter of implementing it in data importer. We do for front-ending services as in https://github.com/fabric8-analytics/f8a-server-backbone/blob/aa74d11653d05f6e7bb4d1daad50c188d0a6168f/src/recommender.py#L129
But since data importer endpoint was not public till now, so was ok, but, we may need to implement it .