Erez Shinan
|
00811a7c16
|
Don't construct a list of tokens, to reduce memory footprint
|
7 years ago |
Erez Shinan
|
c90b933a33
|
Made load_grammar a tiny bit nicer
|
7 years ago |
Erez Shinan
|
c1968b7e30
|
Fixed tools.nearley (converter)
|
7 years ago |
Erez Shinan
|
3dcf3380d3
|
Bumped patch-version
|
7 years ago |
Erez Shinan
|
9ef3670ea2
|
Improved error messages
|
7 years ago |
Erez Shinan
|
89edd9f30f
|
Bugfix in scanless mode
|
7 years ago |
Erez Shinan
|
3630f5a5a4
|
Version bump. And not for nothing
|
7 years ago |
Erez Shinan
|
794a1c4965
|
More flags work
|
7 years ago |
Erez Shinan
|
9d6b496f3e
|
Fixed bug in unless: Now uses regexp, not string matching
|
7 years ago |
Erez Shinan
|
34acc4674e
|
Added flag options to terminals (just 'i' for now)
|
7 years ago |
Erez Shinan
|
2dbf5d7b26
|
Little stuff
|
7 years ago |
Erez Shinan
|
24f86569da
|
Fixed issue #3 (infinite recursion in grammar)
|
7 years ago |
Erez Shinan
|
d40ddff5c0
|
Improved README
|
7 years ago |
Erez Shinan
|
a73cc9ad90
|
Re-wrote the Earley parser to use a parse-forest
It now knows how to resolve ambiguity! And in a memory-efficient way!
|
7 years ago |
Erez Shinan
|
616dc9f300
|
Added stuff to README
|
7 years ago |
Erez Shinan
|
29718a9a3a
|
Bugfix in scanless
|
7 years ago |
Erez Shinan
|
680449fb6b
|
A little bit of cleanup
|
7 years ago |
Erez Shinan
|
d8cbd239d7
|
Fixed a bug in string evaluation
|
7 years ago |
Erez Shinan
|
e151d22ea2
|
examples.reconstruct_json now working with scanless
|
7 years ago |
Erez Shinan
|
5c115acbd9
|
Final fixes for scanless (hopefully)
|
7 years ago |
Erez Shinan
|
c1fbb3214c
|
Earley.lexer = None as default now
|
7 years ago |
Erez Shinan
|
96ebe94963
|
Solved issue #3: infinite loop due to zero-length tokens
|
7 years ago |
Erez Shinan
|
d230e7744c
|
Pylint
|
7 years ago |
Erez Shinan
|
302dad36fe
|
Dramatically improved memory consumption of Earley parser
I no longer keep the entire table, only the relevant columns.
I let Python handle garbage collection for me.
|
7 years ago |
Erez Shinan
|
4397d38daa
|
Some fixes in examples
|
7 years ago |
Erez Shinan
|
a60c339ff7
|
Scanless Earley now working for all tests!
|
7 years ago |
Erez Shinan
|
be6aa4f726
|
Fixed string escaping and unicode handling
|
7 years ago |
Erez Shinan
|
df4d471641
|
Updated docs to match v0.2
|
7 years ago |
Erez Shinan
|
c9b45823ed
|
Better support for scanless parsing
|
7 years ago |
Erez Shinan
|
a5a20a423a
|
Changed parser/lexer interface in lark. Bumped minor version
|
7 years ago |
Erez Shinan
|
fbba305a9e
|
Fixed grammars
|
7 years ago |
Erez Shinan
|
5236e4a32a
|
Refactored TokenDef to store Pattern(Str/RE)
|
7 years ago |
Erez Shinan
|
a0bb661c61
|
Token import now working
|
7 years ago |
Erez Shinan
|
c70355389f
|
Initial support for EBNF in tokens (automatic compilation to regexps)
|
7 years ago |
Erez Shinan
|
8e7395c6d6
|
Improved earley.py and added some comments to both parsers
|
7 years ago |
Erez Shinan
|
a0fbecee70
|
Tiny improvement in Earley
|
7 years ago |
Erez Shinan
|
9f00b3fab7
|
Re-wrote Earley's mainloop implementation. Better performance.
Now uses Column & NewsList instead of sets.
|
7 years ago |
Erez Shinan
|
4076256faa
|
Reconstruct working again
|
7 years ago |
Erez Shinan
|
c14435ac95
|
Some cleanup
|
7 years ago |
Erez Shinan
|
c17558dd91
|
Official switched to my Earley implementation
|
7 years ago |
Erez Shinan
|
538f944602
|
My Earley parser is working
|
7 years ago |
Erez Shinan
|
e4d3e74f6a
|
Finished refactoring
|
7 years ago |
Erez Shinan
|
d5123812b0
|
Another step of refactoring
|
7 years ago |
Erez Shinan
|
b95567c4a7
|
Another refactoring step
|
7 years ago |
Erez Shinan
|
972034fd2d
|
Began refactoring of lalr_analysis -> grammar_analysis
|
7 years ago |
Erez Shinan
|
07df4b80eb
|
My Earley parser is now working. Not yet plugged in
|
7 years ago |
Erez Shinan
|
8189172701
|
recognizer working
|
7 years ago |
Erez Shinan
|
43da0ddad9
|
Initial reconstruct, working with example!
Definitely needs more work though
|
7 years ago |
Erez Shinan
|
0b902b1d03
|
Added the '!' prefix to keep all tokens in rule. Also removed tools (it's now in lark/tools)
|
7 years ago |
Erez Shinan
|
b7dcfbca58
|
Tools can now be installed via pip
|
7 years ago |