You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Grant Ingersoll (JIRA)" <ji...@apache.org> on 2008/01/15 17:00:47 UTC
[jira] Created: (LUCENE-1133) WikipediaTokenizer needs a way of not
tokenizing
WikipediaTokenizer needs a way of not tokenizing
-------------------------------------------------
Key: LUCENE-1133
URL: https://issues.apache.org/jira/browse/LUCENE-1133
Project: Lucene - Java
Issue Type: Improvement
Components: contrib/wikipedia
Affects Versions: 2.3
Reporter: Grant Ingersoll
Priority: Minor
It would be nice if the WikipediaTokenizer had a way of, via a flag, leaving categories, links, etc. as single tokens (or at least some parts of them)
Thus, if we came across [[Category:My Big Idea]] there would be a way of outputting, as a single token "My Big Idea".
Optionally, it would be good to output both "My Big Idea" and the individual tokens as well.
I am not sure of how to do this in JFlex, so any insight would be appreciated.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
[jira] Updated: (LUCENE-1133) WikipediaTokenizer needs a way of not
tokenizing certain parts of the text
Posted by "Grant Ingersoll (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-1133?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Grant Ingersoll updated LUCENE-1133:
------------------------------------
Attachment: LUCENE_1133_1137.patch
Here's a patch that also includes LUCENE-1137 and sets a flag for the untokenized token when BOTH or UNTOKENIZED_ONLY is the output type.
This patch is an alternative to the existing one and doesn't necessarily have to replace that one, it just demonstrates using 1137 for something fun.
> WikipediaTokenizer needs a way of not tokenizing certain parts of the text
> --------------------------------------------------------------------------
>
> Key: LUCENE-1133
> URL: https://issues.apache.org/jira/browse/LUCENE-1133
> Project: Lucene - Java
> Issue Type: Improvement
> Components: contrib/wikipedia
> Affects Versions: 2.3
> Reporter: Grant Ingersoll
> Assignee: Grant Ingersoll
> Priority: Minor
> Attachments: LUCENE-1133.patch, LUCENE_1133_1137.patch
>
>
> It would be nice if the WikipediaTokenizer had a way of, via a flag, leaving categories, links, etc. as single tokens (or at least some parts of them)
> Thus, if we came across [[Category:My Big Idea]] there would be a way of outputting, as a single token "My Big Idea".
> Optionally, it would be good to output both "My Big Idea" and the individual tokens as well.
> I am not sure of how to do this in JFlex, so any insight would be appreciated.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
[jira] Resolved: (LUCENE-1133) WikipediaTokenizer needs a way of
not tokenizing certain parts of the text
Posted by "Grant Ingersoll (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-1133?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Grant Ingersoll resolved LUCENE-1133.
-------------------------------------
Resolution: Fixed
Lucene Fields: (was: [New])
Committed on 614895
> WikipediaTokenizer needs a way of not tokenizing certain parts of the text
> --------------------------------------------------------------------------
>
> Key: LUCENE-1133
> URL: https://issues.apache.org/jira/browse/LUCENE-1133
> Project: Lucene - Java
> Issue Type: Improvement
> Components: contrib/wikipedia
> Affects Versions: 2.3
> Reporter: Grant Ingersoll
> Assignee: Grant Ingersoll
> Priority: Minor
> Attachments: LUCENE-1133.patch, LUCENE_1133_1137.patch
>
>
> It would be nice if the WikipediaTokenizer had a way of, via a flag, leaving categories, links, etc. as single tokens (or at least some parts of them)
> Thus, if we came across [[Category:My Big Idea]] there would be a way of outputting, as a single token "My Big Idea".
> Optionally, it would be good to output both "My Big Idea" and the individual tokens as well.
> I am not sure of how to do this in JFlex, so any insight would be appreciated.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
[jira] Updated: (LUCENE-1133) WikipediaTokenizer needs a way of not
tokenizing certain parts of the text
Posted by "Grant Ingersoll (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-1133?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Grant Ingersoll updated LUCENE-1133:
------------------------------------
Attachment: LUCENE-1133.patch
Here's a completely back-compatible (if there is such a thing for something that hasn't been released) patch that allows for 3 different output options:
1. Tokens only (the current functionality)
2. Untokenized only
3. Both
In case 2, a Set of Wiki types is passed in, and any token that has that type will not be tokenized. Case 3 is the combination of 1 and 2, with the untokenized token output first.
See the unit test for how it should work.
I will commit after the release, pending improvements, etc.
> WikipediaTokenizer needs a way of not tokenizing certain parts of the text
> --------------------------------------------------------------------------
>
> Key: LUCENE-1133
> URL: https://issues.apache.org/jira/browse/LUCENE-1133
> Project: Lucene - Java
> Issue Type: Improvement
> Components: contrib/wikipedia
> Affects Versions: 2.3
> Reporter: Grant Ingersoll
> Assignee: Grant Ingersoll
> Priority: Minor
> Attachments: LUCENE-1133.patch
>
>
> It would be nice if the WikipediaTokenizer had a way of, via a flag, leaving categories, links, etc. as single tokens (or at least some parts of them)
> Thus, if we came across [[Category:My Big Idea]] there would be a way of outputting, as a single token "My Big Idea".
> Optionally, it would be good to output both "My Big Idea" and the individual tokens as well.
> I am not sure of how to do this in JFlex, so any insight would be appreciated.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
[jira] Commented: (LUCENE-1133) WikipediaTokenizer needs a way of
not tokenizing certain parts of the text
Posted by "Grant Ingersoll (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-1133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12559158#action_12559158 ]
Grant Ingersoll commented on LUCENE-1133:
-----------------------------------------
I use a TokenFilter to decide what to with the links
> WikipediaTokenizer needs a way of not tokenizing certain parts of the text
> --------------------------------------------------------------------------
>
> Key: LUCENE-1133
> URL: https://issues.apache.org/jira/browse/LUCENE-1133
> Project: Lucene - Java
> Issue Type: Improvement
> Components: contrib/wikipedia
> Affects Versions: 2.3
> Reporter: Grant Ingersoll
> Priority: Minor
>
> It would be nice if the WikipediaTokenizer had a way of, via a flag, leaving categories, links, etc. as single tokens (or at least some parts of them)
> Thus, if we came across [[Category:My Big Idea]] there would be a way of outputting, as a single token "My Big Idea".
> Optionally, it would be good to output both "My Big Idea" and the individual tokens as well.
> I am not sure of how to do this in JFlex, so any insight would be appreciated.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
[jira] Updated: (LUCENE-1133) WikipediaTokenizer needs a way of not
tokenizing certain parts of the text
Posted by "Grant Ingersoll (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-1133?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Grant Ingersoll updated LUCENE-1133:
------------------------------------
Summary: WikipediaTokenizer needs a way of not tokenizing certain parts of the text (was: WikipediaTokenizer needs a way of not tokenizing )
> WikipediaTokenizer needs a way of not tokenizing certain parts of the text
> --------------------------------------------------------------------------
>
> Key: LUCENE-1133
> URL: https://issues.apache.org/jira/browse/LUCENE-1133
> Project: Lucene - Java
> Issue Type: Improvement
> Components: contrib/wikipedia
> Affects Versions: 2.3
> Reporter: Grant Ingersoll
> Priority: Minor
>
> It would be nice if the WikipediaTokenizer had a way of, via a flag, leaving categories, links, etc. as single tokens (or at least some parts of them)
> Thus, if we came across [[Category:My Big Idea]] there would be a way of outputting, as a single token "My Big Idea".
> Optionally, it would be good to output both "My Big Idea" and the individual tokens as well.
> I am not sure of how to do this in JFlex, so any insight would be appreciated.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
[jira] Commented: (LUCENE-1133) WikipediaTokenizer needs a way of
not tokenizing certain parts of the text
Posted by "Grant Ingersoll (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-1133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12559175#action_12559175 ]
Grant Ingersoll commented on LUCENE-1133:
-----------------------------------------
Custom.
> WikipediaTokenizer needs a way of not tokenizing certain parts of the text
> --------------------------------------------------------------------------
>
> Key: LUCENE-1133
> URL: https://issues.apache.org/jira/browse/LUCENE-1133
> Project: Lucene - Java
> Issue Type: Improvement
> Components: contrib/wikipedia
> Affects Versions: 2.3
> Reporter: Grant Ingersoll
> Priority: Minor
>
> It would be nice if the WikipediaTokenizer had a way of, via a flag, leaving categories, links, etc. as single tokens (or at least some parts of them)
> Thus, if we came across [[Category:My Big Idea]] there would be a way of outputting, as a single token "My Big Idea".
> Optionally, it would be good to output both "My Big Idea" and the individual tokens as well.
> I am not sure of how to do this in JFlex, so any insight would be appreciated.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
[jira] Commented: (LUCENE-1133) WikipediaTokenizer needs a way of
not tokenizing certain parts of the text
Posted by "Otis Gospodnetic (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-1133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12559126#action_12559126 ]
Otis Gospodnetic commented on LUCENE-1133:
------------------------------------------
I'd actually love the option to skip that stuff all together. The last time I checked one Wikipedia dump (Croatian, not English), text/pointers in angle bracket was really junk, at lest for my purposes of getting a representative list of tokens and types for a language.
> WikipediaTokenizer needs a way of not tokenizing certain parts of the text
> --------------------------------------------------------------------------
>
> Key: LUCENE-1133
> URL: https://issues.apache.org/jira/browse/LUCENE-1133
> Project: Lucene - Java
> Issue Type: Improvement
> Components: contrib/wikipedia
> Affects Versions: 2.3
> Reporter: Grant Ingersoll
> Priority: Minor
>
> It would be nice if the WikipediaTokenizer had a way of, via a flag, leaving categories, links, etc. as single tokens (or at least some parts of them)
> Thus, if we came across [[Category:My Big Idea]] there would be a way of outputting, as a single token "My Big Idea".
> Optionally, it would be good to output both "My Big Idea" and the individual tokens as well.
> I am not sure of how to do this in JFlex, so any insight would be appreciated.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
[jira] Assigned: (LUCENE-1133) WikipediaTokenizer needs a way of
not tokenizing certain parts of the text
Posted by "Grant Ingersoll (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-1133?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Grant Ingersoll reassigned LUCENE-1133:
---------------------------------------
Assignee: Grant Ingersoll
> WikipediaTokenizer needs a way of not tokenizing certain parts of the text
> --------------------------------------------------------------------------
>
> Key: LUCENE-1133
> URL: https://issues.apache.org/jira/browse/LUCENE-1133
> Project: Lucene - Java
> Issue Type: Improvement
> Components: contrib/wikipedia
> Affects Versions: 2.3
> Reporter: Grant Ingersoll
> Assignee: Grant Ingersoll
> Priority: Minor
>
> It would be nice if the WikipediaTokenizer had a way of, via a flag, leaving categories, links, etc. as single tokens (or at least some parts of them)
> Thus, if we came across [[Category:My Big Idea]] there would be a way of outputting, as a single token "My Big Idea".
> Optionally, it would be good to output both "My Big Idea" and the individual tokens as well.
> I am not sure of how to do this in JFlex, so any insight would be appreciated.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org
[jira] Commented: (LUCENE-1133) WikipediaTokenizer needs a way of
not tokenizing certain parts of the text
Posted by "Otis Gospodnetic (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/LUCENE-1133?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12559163#action_12559163 ]
Otis Gospodnetic commented on LUCENE-1133:
------------------------------------------
Aha. Which existing one? Custom?
> WikipediaTokenizer needs a way of not tokenizing certain parts of the text
> --------------------------------------------------------------------------
>
> Key: LUCENE-1133
> URL: https://issues.apache.org/jira/browse/LUCENE-1133
> Project: Lucene - Java
> Issue Type: Improvement
> Components: contrib/wikipedia
> Affects Versions: 2.3
> Reporter: Grant Ingersoll
> Priority: Minor
>
> It would be nice if the WikipediaTokenizer had a way of, via a flag, leaving categories, links, etc. as single tokens (or at least some parts of them)
> Thus, if we came across [[Category:My Big Idea]] there would be a way of outputting, as a single token "My Big Idea".
> Optionally, it would be good to output both "My Big Idea" and the individual tokens as well.
> I am not sure of how to do this in JFlex, so any insight would be appreciated.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-dev-help@lucene.apache.org