Comments on: Getting OpenNLP parser to work http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/ > From inner thoughts to the outer limits of Alexandre Rafalovitch Mon, 03 Oct 2011 22:01:43 +0000 hourly 1 http://wordpress.org/?v=3.1.4 By: Dave http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/comment-page-1/#comment-71389 Dave Sun, 08 May 2011 02:19:52 +0000 http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/#comment-71389 Those OpenNLP tutorials are still a work-in-progress. Some are complete, but only the basics. I spent a lot of time recently digging into v1.5.0 (current as I write this) and am hoping to write up my own lessons learned so I don't forget in the future. Might be useful to someone else, too. Here's the start: http://blog.dpdearing.com/2011/05/opennlp-1-5-0-basics-sentence-detection-and-tokenizing/ Those OpenNLP tutorials are still a work-in-progress. Some are complete, but only the basics. I spent a lot of time recently digging into v1.5.0 (current as I write this) and am hoping to write up my own lessons learned so I don’t forget in the future.

Might be useful to someone else, too. Here’s the start:
http://blog.dpdearing.com/2011/05/opennlp-1-5-0-basics-sentence-detection-and-tokenizing/

]]> By: Jörn http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/comment-page-1/#comment-37029 Jörn Wed, 22 Sep 2010 09:48:03 +0000 http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/#comment-37029 There are now OpenNLP tutorials in the project wiki: https://sourceforge.net/apps/mediawiki/opennlp/index.php?title=Main_Page There are now OpenNLP tutorials in the project wiki:
https://sourceforge.net/apps/mediawiki/opennlp/index.php?title=Main_Page

]]> By: Laughing Man http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/comment-page-1/#comment-13228 Laughing Man Mon, 25 May 2009 08:41:42 +0000 http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/#comment-13228 I wrote a small program which takes a simple XML file, cleans it and uses the detector/tokenizer/tagger/chunker (in that order). The download link and the install instructions (it's an Ant build.xml) can be found <a href="http://coadsy.blogspot.com/2009/05/language-guesser-and-opennlp-pipe.html" rel="nofollow">here</a>. I wrote a small program which takes a simple XML file, cleans it and uses the detector/tokenizer/tagger/chunker (in that order). The download link and the install instructions (it’s an Ant build.xml) can be found here.

]]> By: Alexandre Rafalovitch http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/comment-page-1/#comment-238 Alexandre Rafalovitch Wed, 28 May 2008 12:14:57 +0000 http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/#comment-238 Quek, I don't buy 'read the source' argument in the context of this article. Code, and especially NLP code, is not easy to read and therefore has a higher learning curve than tutorials and good documentations. Therefore, people with not enough time (or patience) will drop out early and go look at something else. That's why one of the measure of success for open source projects is whether they have a good introductory book published. Obviously, not everything can be documented, but it should be easy to at least start quickly and judge whether it is worth spending extra time on looking through the code. That's what happened with Stanford parser for me. There was just enough documentation to start and then I read the source to figure out <a href="http://blog.outerthoughts.com/2007/07/duplicating-tagseparator-effect-when-using-stanford-parser-programmatically/" rel="nofollow">the additional options</a>. Quek,

I don’t buy ‘read the source’ argument in the context of this article. Code, and especially NLP code, is not easy to read and therefore has a higher learning curve than tutorials and good documentations. Therefore, people with not enough time (or patience) will drop out early and go look at something else. That’s why one of the measure of success for open source projects is whether they have a good introductory book published.

Obviously, not everything can be documented, but it should be easy to at least start quickly and judge whether it is worth spending extra time on looking through the code.

That’s what happened with Stanford parser for me. There was just enough documentation to start and then I read the source to figure out the additional options.

]]> By: quek http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/comment-page-1/#comment-237 quek Wed, 28 May 2008 03:52:50 +0000 http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/#comment-237 dude... daniel maclaren's tutorial is based on code within the OpenNLP package itself. In fact almost all the code's taken from the english.* java files. It is open source after all, all you have to do is read through the source. chill man chill. dude… daniel maclaren’s tutorial is based on code within the OpenNLP package itself. In fact almost all the code’s taken from the english.* java files. It is open source after all, all you have to do is read through the source. chill man chill.

]]> By: Karalabe http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/comment-page-1/#comment-239 Karalabe Tue, 08 Apr 2008 19:41:44 +0000 http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/#comment-239 You can use 2 lines of code to get you're sentences parsed :) [figured it out partly thanks to the tutorial you linked in] <code> String sentence = "This is some random sentence to get you started"; // Load the serialized parser ParserME parser = TreebankParser.getParser("wherever-your-parser-dir-is/"); // Parse the sentence, and fetch an array containing the best 1 parses, access the first (0) element and display it TreebankParser.parseLine(sentence, parser, 1)[0].show(); </code> You can use 2 lines of code to get you’re sentences parsed :) [figured it out partly thanks to the tutorial you linked in]


String sentence = "This is some random sentence to get you started";

// Load the serialized parser
ParserME parser = TreebankParser.getParser("wherever-your-parser-dir-is/");

// Parse the sentence, and fetch an array containing the best 1 parses, access the first (0) element and display it
TreebankParser.parseLine(sentence, parser, 1)[0].show();

]]> By: Alexandre Rafalovitch http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/comment-page-1/#comment-236 Alexandre Rafalovitch Sat, 16 Feb 2008 17:33:31 +0000 http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/#comment-236 Thank you Alyona, This looks interesting as well. Thank you Alyona,

This looks interesting as well.

]]> By: Alyona http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/comment-page-1/#comment-235 Alyona Wed, 13 Feb 2008 21:58:01 +0000 http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/#comment-235 I found this tutorial useful: http://www.codeproject.com/KB/recipes/englishparsing.aspx I found this tutorial useful:
http://www.codeproject.com/KB/recipes/englishparsing.aspx

]]> By: Jeryl http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/comment-page-1/#comment-234 Jeryl Tue, 11 Dec 2007 00:44:42 +0000 http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/#comment-234 worked for me, whats your email i can send u sample code(as eclipse project with junit tests.) worked for me, whats your email i can send u sample code(as eclipse project with junit tests.)

]]> By: Alexandre Rafalovitch http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/comment-page-1/#comment-233 Alexandre Rafalovitch Sun, 12 Aug 2007 12:56:05 +0000 http://blog.outerthoughts.com/2007/08/getting-opennlp-parser-to-work/#comment-233 Matthew, Same application as with the previous two parsers: trying to extract complex named entities out of long, non-trivial sentences. Matthew,

Same application as with the previous two parsers: trying to extract complex named entities out of long, non-trivial sentences.

]]>