Getting an AST for a Cypher query?

[Rich Morin]

Ronja looks like a great addition to Neo4j.  I'd like to leverage it for my own (exploratory) work.

By way of background, the first stage of Neo4j 2.2's Cypher processing is to:

Convert the input query string into an abstract syntax tree (AST)

The input query string is first tokenized and then parsed into an AST. 

 

Using this AST, we perform semantic checking of the variable types and scoping of variables within the tree. Any errors regarding basic typing information – such as attempting to divide two strings – are caught and returned to the user.

-- Introducing the new Cypher Query Optimizer

    http://neo4j.com/blog/introducing-new-cypher-query-optimizer

I would like to be able to send a Cypher query to Neo4j and (without executing the query) get the resulting AST. For extra credit, I'd like to see some of the data structures (eg, optimized and normalized AST, query graph, logical plan, execution plan) that Ronja generates.

Can anyone point me to (a) information on how I might do this and/or (b) documentation on these data structures?

-r

[Michael Hunger]

I think you would have to write code to access the internal structures in the cypher compiler

Most probably to return the data that's cached in the two execution plan caches

The code is in the cypher-compiler modules for each version

Von meinem iPhone gesendet

--
You received this message because you are subscribed to the Google Groups "Neo4j" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neo4j+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Rich Morin]

I'd like to minimize the amount of Java hacking I need to do,

particularly when it involves experimental, special-purpose code. Fortunately, I think I see a general solution that would serve a number of disparate purposes.

Basically, I'd like to have a way to register callbacks (dynamically) to the Ronja control flow. So, I've written this up as a Neo4j issue (ducks :-).

-r

[Clark Richey]

I do have to say that being able to access the internal Cypher interpreter / parser would be of great help. I looked at trying to access it with Neo 2.1.6 but simply couldn’t figure out how to properly do so (lots of dependencies of things maintained by the neo kernel). I am interested in using it for parsing out cypher that is sent via an API so that I can do things such as add in additional security clauses (match patterns) and look for things that may not be allowed such as deletions, merges, access to certain properties, etc.

We are currently writing our own parser to handle this but it would be SO much more efficient if there was an easy was to ask the neo kernel to parse text into cypher and get access to its constituent parts.

Clark Richey

clark....@gmail.com

[Rich Morin]

Chris Leishman indicated (in an issue response) that this thread is the best place to make a case for the callback registration idea, ask for help, etc. So, here goes...

My immediate motivation has to do with some ideas I have for using Cypher in an Elixir-based implementation of Property Graphs. I'd like to use parsed (and possibly, optimized) queries as a way to specify actor-based graph algorithms. See Cyphex for details.

Another use case has to do with extending and improving Cypher's syntax. For example, Cypher has very limited modularity. If I could access and modify the incoming Cypher query, I might be able to add "pattern variables", generalized macro processing, etc.

-r