Repository navigation
uast_extract does not allow to extract nested properties #702
Description
Activity
It's doable, but we have to discuss in what format we want to return extracted data.
Here is a one big object in json:[ { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 8, "line": 4, "offset": 31 }, "start": { "@type": "uast:Position", "col": 5, "line": 4, "offset": 28 } }, "@role": [ "Function", "Declaration", "Name", "Identifier" ], "@token": "sum", "@type": "FunctionDef", "args": { "@pos": { "@type": "uast:Positions" }, "@role": [ "Function", "Declaration", "Argument", "Incomplete" ], "@type": "arguments", "args": [ { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 10, "line": 4, "offset": 33 }, "start": { "@type": "uast:Position", "col": 9, "line": 4, "offset": 32 } }, "@role": [ "Function", "Declaration", "Argument", "Name" ], "@token": "a", "@type": "arg", "annotation": null, "noops_previous": { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 1, "line": 3, "offset": 23 }, "start": { "@type": "uast:Position", "col": 1, "line": 1, "offset": 0 } }, "@role": [ "Noop" ], "@type": "PreviousNoops", "lines": [ { "@pos": { "@type": "uast:Positions", "start": { "@type": "uast:Position", "col": 1, "line": 2, "offset": 1 } }, "@role": [ "Comment", "Noop" ], "@token": "#!/usr/bin/env python\\n", "@type": "NoopLine" } ] } }, { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 13, "line": 4, "offset": 36 }, "start": { "@type": "uast:Position", "col": 12, "line": 4, "offset": 35 } }, "@role": [ "Function", "Declaration", "Argument", "Name" ], "@token": "b", "@type": "arg", "annotation": null } ] }, "body": { "@role": [ "Function", "Declaration", "Body" ], "@type": "FunctionDef.body", "body_stmts": [ { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 8, "line": 5, "offset": 46 }, "start": { "@type": "uast:Position", "col": 2, "line": 5, "offset": 40 } }, "@role": [ "Return", "Statement" ], "@token": "return", "@type": "Return", "value": { "@pos": { "@type": "uast:Positions", "start": { "@type": "uast:Position", "col": 9, "line": 5, "offset": 47 } }, "@role": [ "Expression", "Binary" ], "@type": "BinOp", "left": { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 10, "line": 5, "offset": 48 }, "start": { "@type": "uast:Position", "col": 9, "line": 5, "offset": 47 } }, "@role": [ "Identifier", "Expression", "Binary", "Left" ], "@token": "a", "@type": "Name", "ctx": "Load" }, "op": { "@pos": { "@type": "uast:Positions" }, "@role": [ "Operator", "Arithmetic", "Add", "Binary" ], "@token": "+", "@type": "Add" }, "right": { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 14, "line": 5, "offset": 52 }, "start": { "@type": "uast:Position", "col": 13, "line": 5, "offset": 51 } }, "@role": [ "Identifier", "Expression", "Binary", "Right" ], "@token": "b", "@type": "Name", "ctx": "Load" } } } ] }, "decorator_list": { "@role": [ "Function", "Declaration", "Incomplete" ], "@type": "FunctionDef.decorators", "decorators": [] }, "returns": null }, { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 10, "line": 5, "offset": 48 }, "start": { "@type": "uast:Position", "col": 9, "line": 5, "offset": 47 } }, "@role": [ "Identifier", "Expression", "Binary", "Left" ], "@token": "a", "@type": "Name", "ctx": "Load" }, { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 14, "line": 5, "offset": 52 }, "start": { "@type": "uast:Position", "col": 13, "line": 5, "offset": 51 } }, "@role": [ "Identifier", "Expression", "Binary", "Right" ], "@token": "b", "@type": "Name", "ctx": "Load" }, { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 10, "line": 7, "offset": 63 }, "start": { "@type": "uast:Position", "col": 7, "line": 7, "offset": 60 } }, "@role": [ "Identifier", "Expression", "Call", "Callee" ], "@token": "sum", "@type": "Name", "ctx": "Load" }, { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 6, "line": 7, "offset": 59 }, "start": { "@type": "uast:Position", "col": 1, "line": 7, "offset": 54 } }, "@role": [ "Identifier", "Expression", "Call", "Callee" ], "@token": "print", "@type": "Name", "ctx": "Load", "noops_previous": { "@pos": { "@type": "uast:Positions", "end": { "@type": "uast:Position", "col": 1, "line": 6, "offset": 53 }, "start": { "@type": "uast:Position", "col": 1, "line": 6, "offset": 53 } }, "@role": [ "Noop" ], "@type": "PreviousNoops", "lines": [] } } ]Same as we're retuning now, right? As for the syntax, maybe jsonpath? we already use it in JSON_EXTRACT UDF.
@erizocosmico - this is what we get from bblfsh (whole node) converted to json. But the question is about
extract. So far, extract returns a slice of all extracted sub-node values, e.g.:["FunctionDef","Name","Name","Name","Name"] ["sum","a","b","sum","print"] [["Function","Declaration","Name","Identifier"],["Identifier","Expression","Binary","Left"],["Identifier","Expression","Binary","Right"],["Identifier","Expression","Call","Callee"],["Identifier","Expression","Call","Callee"]]@ajnavarro what shall we do with this?
I would leave it here just to have it in mind.
There is currently no way to allow extraction of nested properties of a node, which would be useful for getting, for example, the line and column of the position.