Douglas Lovell
c0d89cec2d
add a few word transition tests
2015-10-30 16:25:04 -06:00
Douglas Lovell
f9c2d9d4aa
all the tests pass
2015-10-30 15:03:06 -06:00
Douglas Lovell
b7bb0cbf5b
put text positions on the transitions and track in the token stream
2015-10-30 10:58:15 -06:00
Douglas Lovell
dd5f9b25fa
fix the off by ones
2015-10-29 15:29:44 -06:00
Douglas Lovell
7514478a65
Make the transition token the hash key for a transition
2015-10-29 15:28:23 -06:00
Douglas Lovell
8560af8cce
test trie in word transition mode
2015-10-28 17:03:29 -06:00
Douglas Lovell
51940af6e7
test state with word transitions
2015-10-28 16:57:37 -06:00
Douglas Lovell
1f63ae71d4
StringBuffer, StringIterator are old school
2015-10-28 16:56:27 -06:00
Douglas Lovell
4be3e115b6
add builder method for setting word transitions
2015-10-28 14:10:53 -06:00
Douglas Lovell
a646f233a5
improve the option name
2015-10-28 11:17:46 -06:00
Douglas Lovell
f05026cf90
added word transitions. all tests pass
2015-10-28 10:59:45 -06:00
Douglas Lovell
6283bf039d
first pass refactoring for word transitions
2015-10-27 16:55:35 -06:00
robert-bor
3393e4f51f
Issue #26 tokens report if they are 100% whitespace
2015-09-27 20:56:04 +02:00
robert-bor
438e546245
Issue #25 match tokens report back whether they are whole words or not
2015-09-27 18:22:38 +02:00
robert-bor
877a56c956
Issue #24 removed if condition that checked for empty emit strings, whereas emit() always returns a collection
2015-09-27 17:58:55 +02:00
robert-bor
b42c664796
Issue #24 big cleanup, removed all post-processing methods for whole words and non-overlapping sequences and integrated the same functionality closer to the AC algorithm.
2015-09-27 17:56:42 +02:00
robert-bor
b274844b75
Issue #24 stopOnHit removed; the functionality has been replaced by the superior firstMatch
2015-09-27 14:40:04 +02:00
robert-bor
bfaa32b20e
Issue #24 tokenize() method implementation extracted to separate class
2015-09-27 14:37:36 +02:00
robert-bor
5203efbbcb
Extra explanation on containsMatch
2015-09-23 20:56:24 +02:00
robert-bor
a46177415f
Updated README.md documentation
2015-09-23 08:39:13 +02:00
robert-bor
e365689391
v0.3.0
2015-09-22 22:27:30 +02:00
robert-bor
dc27d6e3e9
pull #17 changes adopted to implement a whole word check on the entire keyword, including whitespaces.
2015-09-22 22:22:20 +02:00
robert-bor
76ae8222ea
issue #12 adopted the suggestion by yim1990 with a small change, so that the keyword emit is lowercased as well
2015-09-22 22:10:19 +02:00
robert-bor
e2c5334234
pull #14 implemented pull request by rripken for containsMatch and firstMatch
2015-09-22 22:02:30 +02:00
robert-bor
4633b1ba2a
Merge branch 'rripken-master' into feature/footprint-reduction
...
Conflicts:
src/main/java/org/ahocorasick/trie/Trie.java
src/test/java/org/ahocorasick/trie/TrieTest.java
2015-09-22 20:38:29 +02:00
robert-bor
30f003c5ae
Issue #18 fixed link to broken PDF, now points to http://cr.yp.to/bib/1975/aho.pdf
2015-09-22 20:22:24 +02:00
robert-bor
c18e030459
Merge branch 'SubOptimal-contrib' into feature/footprint-reduction
2015-09-22 20:18:52 +02:00
robert-bor
023c253c93
Issue #16 #20 #21 adopted pull request from remen which makes sure the failure states are constructed as part of the trie construction. This prevents the NPE which the referenced issues are complaining about.
2015-09-22 20:14:48 +02:00
robert-bor
fcefdfdaf9
Merge branch 'remen-master' into feature/footprint-reduction
...
Conflicts:
src/main/java/org/ahocorasick/trie/Trie.java
2015-09-22 20:06:04 +02:00
robert-bor
b85f8fc08f
Issue #22 added possibility to stop processing on generating at least one emit
2015-09-22 19:31:05 +02:00
robert-bor
4399e42b99
Issue #23 apply CharSequence to top-level parseText as well
2015-09-22 06:25:50 +02:00
robert-bor
055e13c298
Issue #23 removed the ParseConfiguration, rely on CharSequence instead
2015-09-21 22:03:59 +02:00
robert-bor
88799fb3da
Issue #23 added callback handler concept which omits the custom setting up of a list, but instead places direct calls to the handler. The handler are only supported on the lowest level of aho-corasick, ie no overlap, whole words and token support
...
Also added the possibility to pass a reader to the same level as above.
2015-09-21 21:09:26 +02:00
Petter Remen
9bce51e001
Issue #16 Use builder pattern to create Trie
...
Previously, there was a race condition in Trie#parseText since
it called constructFailureStates on first run without synchronization.
See https://github.com/robert-bor/aho-corasick/issues/16
This commit fixes this by using the builder pattern in order to
create a fully initialized Trie.
N.B. This changes the API
2015-07-03 12:29:31 +02:00
Frank Dietrich
285a74c37f
fix broken link to the white paper
2015-04-30 01:10:47 +02:00
ryan
d1478c7480
HashMap has better performance in my test cases.
2014-10-06 13:34:03 -07:00
ryan
a46e7dfe1d
Fixed formatting changes.
2014-10-06 11:02:01 -07:00
ryan
df503bae43
Added method and tests for a faster path to return the first match.
2014-10-06 10:52:35 -07:00
robert-bor
25eeef5168
v0.2.4 with bugfix #10
2014-08-27 08:44:06 +02:00
robert-bor
2b125d2689
Issue #10 make sure that State emits a specific match only once
2014-08-27 08:42:46 +02:00
robert-bor
c96c57399a
update README.md
2014-08-26 10:11:05 +02:00
robert-bor
c572d234e1
v0.2.3 bugfix
2014-08-26 10:05:18 +02:00
robert-bor
e8b5be0497
Issue #8 fixed Unicode issue by converting characters individually, not the entire search text
2014-08-26 09:50:15 +02:00
robert-bor
7431c74a7f
Issue #7 bugfix release v0.2.2
2014-02-15 11:50:43 +01:00
robert-bor
d7421ead0f
Issue #7 Ignore keywords in the trie that are null or empty
2014-02-15 11:44:54 +01:00
robert-bor
a4fcfe8f20
Updated README for v0.2.1
2014-02-08 13:24:36 +01:00
robert-bor
4bd568836f
Releasing v0.2.1 bug release for issue #5
2014-02-08 12:46:48 +01:00
robert-bor
31117d6a6e
Solved issue #5 by introducing a proper boundary check for words that are at the end of a String
2014-02-08 12:45:55 +01:00
robert-bor
1656e862df
v0.2.0
2014-02-01 22:02:29 +01:00
robert-bor
bcde097070
Issue #4 Trie.tokenize() available. It returns a list of tokens. A token can be either a fragment (unmatched text) or a match. If it is a match, the original emit can be queried.
2014-02-01 22:01:15 +01:00
