Juanjo Alvarez Martinez

Or even better, XQuery. But AFAIK, we didn't find any good open source libraries implementing these (but let's keep this open in case we find something in the future).

You could also use the node positions in the tree relative to the root as "hash" (key, really). We'll almost surely be adding hashes or keys as part of the...

Sorry, I wasn't clear, I didn't mean the token positions in the source code but the node positions in the tree, take a look at the ascii "art" at the...

These times are not from a query - there are several levels of anidated foreachs inside every iteration each doing queries plus a complete parse on each iteration. Timing the...

I don't know it 0.05 per query in this case over a non trivial UAST can be considered slow or not so I'll leave this open until we decide, @smola...

I don't think it's a problem of query efficiency so much as the libuast calling to some function pointers (like callback) that are implemented on the cpython side which, even...

@smola, definitely, libxml does a lot of allocations when starting up and so does libuast when creating the pseudo-xml. Ad-hoc code on Python works on already allocated and initialized nodes...

These issues should go to the Python and Javascript drivers projects. That script is cool!

For the Python driver, we currently fix tokens that at least have a line number but not column (with are a lot) using a tokenizer, but there are some nodes...

The current token synchronizer removes a token from the list of tokens in lines (list of list) when it gets the position of it so at the time of getting...