lark

lark copied to clipboard

lark permutes all the optional rules/terminals - Loading a grammar file takes >5 mins for a rule with many optionals

4

I've create a grammar file off from the Hive CREATE syntax: https://cwiki.apache.org/confluence/display/Hive/LanguageManual+DDL#LanguageManualDDL-CreateTable It's ambiguous and uses some number of terminals, but loading takes too much time and the serialized parser...

yjung-lyft

Token templates

2

Rule templates are already here. Token templates are next? @MegaIng Do you think this will be ready in the near future? I want to make a release in a few...

erezsh

not finding ambiguous parse

1

Greetings, and thank you for your excellent package. I'm not sure if I'm doing something wrong here, but the following CFG gives me an unambiguous parse on the input "a...

christolmicrosoft

Some grammars can throw the CYK parser into an infinite loop

7

Not likely to happen naturally, but possible. For example: parser = lark.Lark("start: a\na: b\n b:c\n c:a\n | /a/") parser.parse('a') Throws all parsers into an infinite loop (the cyk parser already...

erezsh

Detect collision in regular expressions when creating a lexer

33

When using a standard or contextual lexer, colliding regular expressions may produce an incorrect parser, that only produces an error when encountering the obscured terminal in the input text. For...

erezsh

Earley recogniser - Leo transitives

3

Hi, I noticed that there's an implementation of Joop Leo's optimisation for right recursion, but it is commented [out](https://github.com/lark-parser/lark/blob/f814d91f9dc0d9989a9ef413cca0f2622427eb74/lark/parsers/earley.py#L81). I'm fairly new to the concept of this algorithm so I'm...

deniskyashif

Construct a Lark grammar from ABNF format (RFC 5234)

5

[RFC 5234](https://tools.ietf.org/html/rfc5234) describes the standard grammar format for internet standards, such as the notoriously-hard-to-validate email addresses. Because this standard is a dialect of EBNF and does not allow for embedded...

Zac-HD

I managed to create a non repeatable parser

8

It seems I have some real bad problems with my tentative implementation. Launching the program several times I end up with (at least) two different outputs, even if I'm parsing...

mcondarelli

```python >>> Lark('start: A\n A: "a"+', ambiguity='explicit', lexer='dynamic_complete').parse('aa') Tree(start, [Token(A, 'aa')]) ``` Expected: ambiguous node containing ['aa'] and ['a', 'a']

erezsh

Optimization: Merge anonymous terminals

3

It's possible to improve Lark's speed, for all algorithms, by performing grammar simplification. Specifically, it's possible to merge adjacent anonymous terminals together, to reduce the parser work load, without affecting...

erezsh