Schema function additions

[schloerke]

To keep in the spirit of libgraphqlparser (speed), I added a second function schema2json(txt) that allows for both queries and schema definitions while still keeping the original 'query only' function graphql2json(txt). (link) My opinion is to only use the schema and query parsing function, but I see the point of a small speed increase for query only parsing.

Feel free to change the name to something else. :-), just exposing the required functionality.

I also upgraded the tests for kitchen sink queries and schema definitions.

Side note...
This update is so nice for me! It saves me about a >1000 lines of code for a small to midsize schema definition. I will now be able to do something similar to https://github.com/matthewmueller/graph.ql

[jeroen]

Thanks! Perhaps we can avoid duplicating a lot of C code binding code by creating a single c++ function with a parameter that specifies which parser to use?

// [[Rcpp::export]]
Rcpp::String dump_json_ast(Rcpp::String graph, bool schema) {
  graph.set_encoding(CE_UTF8);
  const char *error;
  auto AST = schema ? 
    facebook::graphql::parseStringWithExperimentalSchemaSupport(graph.get_cstring(), &error) :
    facebook::graphql::parseString(graph.get_cstring(), &error);
  if (!AST) {
    char buf[1000];
    strncpy(buf, error, 1000);
    free((void*) error);
    throw std::runtime_error(buf);
  }
  const char *json = graphql_ast_to_json((const struct GraphQLAstNode *)AST.get());
  Rcpp::String out(json);
  out.set_encoding(CE_UTF8);
  free((void*) json);
  return out;
}

Would that make sense?

[schloerke]

Sounds great to me! Plus it's much less confusing

From the "speed" of R standpoint, I would enable schema parsing by default.

[schloerke]

Oops, I made a bad example. Forgot to use the new function... but that'll get fixed with the merging of the functionality

[schloerke]

Merged into one function now. Tests updated. Added schema example to graphql2json() fn.

[jeroen]

OK looks good. Can you add yourself as a contributor to DESCRIPTION? Also would you mind rebasing to squash the commits into a single commit?

[schloerke]

squashed in #7

[jeroen]

Thanks. Next time you don't have to do a new PR, just git push -f :)

[schloerke]

Ah! I've usually just done with within the Github PR (but can't cuz I'm not owner here), didn't know how to do it command line side. Thank you!

[jeroen]

I always have to look it up but it's really easy: https://ariejan.net/2011/07/05/git-squash-your-latests-commits-into-one/

[jeroen]

This is on CRAN now.