Fix for Issue #724

[mehak16163]

For this fix of issue #724 I have

1. Added ignore_error method for each of the parsers included in rdflib/plugins/parsers.
2. Added a couple of test cases to verify whether the new keyword works as expected.
3. In graph.py where the parse method is called, I have made the changes to include the keyword.
4. Added the logging functionality when the ignore_errors = True, so that we can identify the lines which were not correctly parsed.

[coveralls]

Coverage increased (+0.04%) to 75.801% when pulling 4300559 on mehak16163:Issue724 into 037ea51 on RDFLib:master.

[mehak16163]

Hi @nicholascar,
Can you please go through this pull request and suggest any changes to be done?

[nicholascar]

Hi @mehak16163, thanks for this PR but there are some problems with it that mean we, the maintainers, have to reject it.

It doesn't make sense to ignore parsing errors in any RDF format other than n-triples as the triples are inter-connected so if you ignore a parsing error, you will not get a valid graph. in n-triples, you could ignore a single line parsing fail as each line is a subject, predicate, object, triple, but this is the only format you can do this for.

Also, you've tried to present a passing and a failing test for your changes but your passing test will always pass and doesn't notice what the state of the parsed RDF is, since you've ignored errors, and your failing test will always fail, since you are parsing invalid RDF. So your given tests don't actually tell us the code is working (the code can't work, except for n-triples, as I said).

Sorry but I have to close this PR.