You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@opennlp.apache.org by Jörn Kottmann <ko...@gmail.com> on 2011/03/21 11:12:39 UTC
OpenNLP 1.5.1 Release Candidate 5
Hello,
our next RC is ready for testing, compared to RC 4 it fixes
a bug in the creation of the sample UIMA PEAR package (OPENNLP-143).
It can be found here:
http://people.apache.org/~joern/releases/opennlp-1.5.1-incubating/rc5/
Testing is now mostly done, and we are ready to release soon.
See our test plan for details:
https://cwiki.apache.org/confluence/display/OPENNLP/TestPlan1.5.1
Jörn
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Jörn Kottmann <ko...@gmail.com>.
On 4/10/11 9:23 AM, Tommaso Teofili wrote:
> Hello,
> everything worked as expected, I tested the generated models both manually
> and via the OpenNLP API.
Thanks, I will update the test plan.
Jörn
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Tommaso Teofili <to...@gmail.com>.
Hello,
everything worked as expected, I tested the generated models both manually
and via the OpenNLP API.
Tommaso
2011/4/7 Jörn Kottmann <ko...@gmail.com>
> Ok, I am just optimistic that it will work and I will prepare the next RC
> today.
>
> Jörn
>
>
> On 4/7/11 1:42 PM, Tommaso Teofili wrote:
>
>> Hello,
>>
>> 2011/4/7 Jörn Kottmann<ko...@gmail.com>
>>
>> Hello,
>>>
>>> would be nice if you could test again.
>>>
>>> I also put the test result you reported into our test plan:
>>>
>>> https://cwiki.apache.org/confluence/display/OPENNLP/TestPlan1.5.1
>>>
>>> Please have a short look there, but it should be correct.
>>>
>>>
>>>
>>> Sure, I can rerun those tests with the just made fixes in a couple of
>> days.
>> The report is correct.
>> Tommaso
>>
>>
>
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Jörn Kottmann <ko...@gmail.com>.
Ok, I am just optimistic that it will work and I will prepare the next
RC today.
Jörn
On 4/7/11 1:42 PM, Tommaso Teofili wrote:
> Hello,
>
> 2011/4/7 Jörn Kottmann<ko...@gmail.com>
>
>> Hello,
>>
>> would be nice if you could test again.
>>
>> I also put the test result you reported into our test plan:
>>
>> https://cwiki.apache.org/confluence/display/OPENNLP/TestPlan1.5.1
>>
>> Please have a short look there, but it should be correct.
>>
>>
>>
> Sure, I can rerun those tests with the just made fixes in a couple of days.
> The report is correct.
> Tommaso
>
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Tommaso Teofili <to...@gmail.com>.
Hello,
2011/4/7 Jörn Kottmann <ko...@gmail.com>
> Hello,
>
> would be nice if you could test again.
>
> I also put the test result you reported into our test plan:
>
> https://cwiki.apache.org/confluence/display/OPENNLP/TestPlan1.5.1
>
> Please have a short look there, but it should be correct.
>
>
>
Sure, I can rerun those tests with the just made fixes in a couple of days.
The report is correct.
Tommaso
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Jörn Kottmann <ko...@gmail.com>.
Hello,
would be nice if you could test again.
I also put the test result you reported into our test plan:
https://cwiki.apache.org/confluence/display/OPENNLP/TestPlan1.5.1
Please have a short look there, but it should be correct.
Thanks,
Jörn
On 4/5/11 7:52 PM, Jörn Kottmann wrote:
> Thanks for reporting this, I directly fixed it:
> https://issues.apache.org/jira/browse/OPENNLP-151
>
> Jörn
>
> On 4/5/11 7:39 PM, Tommaso Teofili wrote:
>> Just today I was testing the TokenizerTrainer and I found a bug there
>> with
>> the isSkipAlphaNumerics parameter: in the initialize() method, I see
>> that
>> it's defined as a local variable too so the instance variable gets never
>> assigned and this causes a NPE on the collectionProcessComplete().
>> The fix is in just removing the "Boolean" type definition at line 111 of
>> TokenizerTrainer [1] which allows assignment of configuration parameter
>> value to the instance variable.
>> Tommaso
>>
>> [1] :
>> http://svn.apache.org/viewvc/incubator/opennlp/trunk/opennlp-uima/src/main/java/opennlp/uima/tokenize/TokenizerTrainer.java?view=markup
>>
>>
>>
>> 2011/4/1 Tommaso Teofili<to...@gmail.com>
>>
>>> 2011/4/1 Jörn Kottmann<ko...@gmail.com>
>>>
>>>> On 4/1/11 12:58 PM, Tommaso Teofili wrote:
>>>>
>>>>> One issue I found is that the opennlp.uima.Language parameter is not
>>>>> defined
>>>>> in the trainers' descriptors causing them to fail during
>>>>> initialization
>>>>> since the *Trainer classes need the language as a mandatory parameter
>>>>> (that
>>>>> is good I think since the statistical model built is language
>>>>> dependent).
>>>>> Am I right or am I missing something?
>>>>>
>>>> No, that really sounds like a mistake, seems like I simply forgot
>>>> to put
>>>> the parameter
>>>> deceleration into the descriptor. I will change it on Monday, or of
>>>> course
>>>> a patch is welcome :)
>>>
>>> I didn't run in any other issues, will provide a patch for the
>>> descriptors
>>> tomorrow or sunday :)
>>> Tommaso
>>>
>>>
>>>> p.s.:
>>>>> Also within that fail case it seems
>>>>> the org.apache.uima.UIMAException_Messages is missing, but I'd not
>>>>> consider
>>>>> this a bug at the moment since I am doing tests in a separate project
>>>>> which
>>>>> could need some tweaks but I though it was still useful to report
>>>>>
>>>> I will have a look, thanks for pointing out, even its not a bug we
>>>> might
>>>> want
>>>> to improve it.
>>>>
>>>> Jörn
>>>>
>>>
>
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Jörn Kottmann <ko...@gmail.com>.
Thanks for reporting this, I directly fixed it:
https://issues.apache.org/jira/browse/OPENNLP-151
Jörn
On 4/5/11 7:39 PM, Tommaso Teofili wrote:
> Just today I was testing the TokenizerTrainer and I found a bug there with
> the isSkipAlphaNumerics parameter: in the initialize() method, I see that
> it's defined as a local variable too so the instance variable gets never
> assigned and this causes a NPE on the collectionProcessComplete().
> The fix is in just removing the "Boolean" type definition at line 111 of
> TokenizerTrainer [1] which allows assignment of configuration parameter
> value to the instance variable.
> Tommaso
>
> [1] :
> http://svn.apache.org/viewvc/incubator/opennlp/trunk/opennlp-uima/src/main/java/opennlp/uima/tokenize/TokenizerTrainer.java?view=markup
>
>
> 2011/4/1 Tommaso Teofili<to...@gmail.com>
>
>> 2011/4/1 Jörn Kottmann<ko...@gmail.com>
>>
>>> On 4/1/11 12:58 PM, Tommaso Teofili wrote:
>>>
>>>> One issue I found is that the opennlp.uima.Language parameter is not
>>>> defined
>>>> in the trainers' descriptors causing them to fail during initialization
>>>> since the *Trainer classes need the language as a mandatory parameter
>>>> (that
>>>> is good I think since the statistical model built is language dependent).
>>>> Am I right or am I missing something?
>>>>
>>> No, that really sounds like a mistake, seems like I simply forgot to put
>>> the parameter
>>> deceleration into the descriptor. I will change it on Monday, or of course
>>> a patch is welcome :)
>>
>> I didn't run in any other issues, will provide a patch for the descriptors
>> tomorrow or sunday :)
>> Tommaso
>>
>>
>>> p.s.:
>>>> Also within that fail case it seems
>>>> the org.apache.uima.UIMAException_Messages is missing, but I'd not
>>>> consider
>>>> this a bug at the moment since I am doing tests in a separate project
>>>> which
>>>> could need some tweaks but I though it was still useful to report
>>>>
>>> I will have a look, thanks for pointing out, even its not a bug we might
>>> want
>>> to improve it.
>>>
>>> Jörn
>>>
>>
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Tommaso Teofili <to...@gmail.com>.
Just today I was testing the TokenizerTrainer and I found a bug there with
the isSkipAlphaNumerics parameter: in the initialize() method, I see that
it's defined as a local variable too so the instance variable gets never
assigned and this causes a NPE on the collectionProcessComplete().
The fix is in just removing the "Boolean" type definition at line 111 of
TokenizerTrainer [1] which allows assignment of configuration parameter
value to the instance variable.
Tommaso
[1] :
http://svn.apache.org/viewvc/incubator/opennlp/trunk/opennlp-uima/src/main/java/opennlp/uima/tokenize/TokenizerTrainer.java?view=markup
2011/4/1 Tommaso Teofili <to...@gmail.com>
> 2011/4/1 Jörn Kottmann <ko...@gmail.com>
>
>> On 4/1/11 12:58 PM, Tommaso Teofili wrote:
>>
>>> One issue I found is that the opennlp.uima.Language parameter is not
>>> defined
>>> in the trainers' descriptors causing them to fail during initialization
>>> since the *Trainer classes need the language as a mandatory parameter
>>> (that
>>> is good I think since the statistical model built is language dependent).
>>> Am I right or am I missing something?
>>>
>>
>> No, that really sounds like a mistake, seems like I simply forgot to put
>> the parameter
>> deceleration into the descriptor. I will change it on Monday, or of course
>> a patch is welcome :)
>
>
> I didn't run in any other issues, will provide a patch for the descriptors
> tomorrow or sunday :)
> Tommaso
>
>
>>
>> p.s.:
>>> Also within that fail case it seems
>>> the org.apache.uima.UIMAException_Messages is missing, but I'd not
>>> consider
>>> this a bug at the moment since I am doing tests in a separate project
>>> which
>>> could need some tweaks but I though it was still useful to report
>>>
>>
>> I will have a look, thanks for pointing out, even its not a bug we might
>> want
>> to improve it.
>>
>> Jörn
>>
>
>
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Tommaso Teofili <to...@gmail.com>.
2011/4/1 Jörn Kottmann <ko...@gmail.com>
> On 4/1/11 12:58 PM, Tommaso Teofili wrote:
>
>> One issue I found is that the opennlp.uima.Language parameter is not
>> defined
>> in the trainers' descriptors causing them to fail during initialization
>> since the *Trainer classes need the language as a mandatory parameter
>> (that
>> is good I think since the statistical model built is language dependent).
>> Am I right or am I missing something?
>>
>
> No, that really sounds like a mistake, seems like I simply forgot to put
> the parameter
> deceleration into the descriptor. I will change it on Monday, or of course
> a patch is welcome :)
I didn't run in any other issues, will provide a patch for the descriptors
tomorrow or sunday :)
Tommaso
>
> p.s.:
>> Also within that fail case it seems
>> the org.apache.uima.UIMAException_Messages is missing, but I'd not
>> consider
>> this a bug at the moment since I am doing tests in a separate project
>> which
>> could need some tweaks but I though it was still useful to report
>>
>
> I will have a look, thanks for pointing out, even its not a bug we might
> want
> to improve it.
>
> Jörn
>
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Jörn Kottmann <ko...@gmail.com>.
On 4/1/11 12:58 PM, Tommaso Teofili wrote:
> One issue I found is that the opennlp.uima.Language parameter is not defined
> in the trainers' descriptors causing them to fail during initialization
> since the *Trainer classes need the language as a mandatory parameter (that
> is good I think since the statistical model built is language dependent).
> Am I right or am I missing something?
No, that really sounds like a mistake, seems like I simply forgot to put
the parameter
deceleration into the descriptor. I will change it on Monday, or of
course a patch is welcome :)
> p.s.:
> Also within that fail case it seems
> the org.apache.uima.UIMAException_Messages is missing, but I'd not consider
> this a bug at the moment since I am doing tests in a separate project which
> could need some tweaks but I though it was still useful to report
I will have a look, thanks for pointing out, even its not a bug we might
want
to improve it.
Jörn
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Tommaso Teofili <to...@gmail.com>.
One issue I found is that the opennlp.uima.Language parameter is not defined
in the trainers' descriptors causing them to fail during initialization
since the *Trainer classes need the language as a mandatory parameter (that
is good I think since the statistical model built is language dependent).
Am I right or am I missing something?
Regards,
Tommaso
p.s.:
Also within that fail case it seems
the org.apache.uima.UIMAException_Messages is missing, but I'd not consider
this a bug at the moment since I am doing tests in a separate project which
could need some tweaks but I though it was still useful to report
2011/4/1 Tommaso Teofili <to...@gmail.com>
> Hi Jörn,
> I've done some simple tests yesterday evening but I am running much of
> those tests today, will let you know by the end of the day how it's going.
> Regards,
> Tommaso
>
>
> 2011/3/31 Jörn Kottmann <ko...@gmail.com>
>
>> On 3/30/11 11:38 AM, Tommaso Teofili wrote:
>>
>>> Hi all,
>>> I can help with testing the UIMA training support with annotated test
>>> data
>>> crawled from various websites since I'm going to do that for a college
>>> course I'm speaking to, so I'd be happy to help (any hints on which
>>> pieces
>>> need stronger tests is welcome).
>>>
>>
>> How are your tests going? I would like to wait with the preparation of the
>> next RC for your results. Can you give us a short update.
>>
>> Thanks,
>> Jörn
>>
>
>
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Tommaso Teofili <to...@gmail.com>.
Hi Jörn,
I've done some simple tests yesterday evening but I am running much of those
tests today, will let you know by the end of the day how it's going.
Regards,
Tommaso
2011/3/31 Jörn Kottmann <ko...@gmail.com>
> On 3/30/11 11:38 AM, Tommaso Teofili wrote:
>
>> Hi all,
>> I can help with testing the UIMA training support with annotated test data
>> crawled from various websites since I'm going to do that for a college
>> course I'm speaking to, so I'd be happy to help (any hints on which pieces
>> need stronger tests is welcome).
>>
>
> How are your tests going? I would like to wait with the preparation of the
> next RC for your results. Can you give us a short update.
>
> Thanks,
> Jörn
>
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Jörn Kottmann <ko...@gmail.com>.
On 3/30/11 11:38 AM, Tommaso Teofili wrote:
> Hi all,
> I can help with testing the UIMA training support with annotated test data
> crawled from various websites since I'm going to do that for a college
> course I'm speaking to, so I'd be happy to help (any hints on which pieces
> need stronger tests is welcome).
How are your tests going? I would like to wait with the preparation of the
next RC for your results. Can you give us a short update.
Thanks,
Jörn
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Jörn Kottmann <ko...@gmail.com>.
On 3/30/11 11:38 AM, Tommaso Teofili wrote:
> Hi all,
> I can help with testing the UIMA training support with annotated test data
> crawled from various websites since I'm going to do that for a college
> course I'm speaking to, so I'd be happy to help (any hints on which pieces
> need stronger tests is welcome).
Thanks, your help is really welcome. No, just train all the components
where we have training support for, and roughly check that all
the training data was passed to the trainers. This you can do e.g.
by having a look at the event count which is printed to the console.
After training is done, you should quickly test the model with our
command line tools to see that it can be loaded and produces some
meaningful output.
Jörn
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Tommaso Teofili <to...@gmail.com>.
Hi all,
I can help with testing the UIMA training support with annotated test data
crawled from various websites since I'm going to do that for a college
course I'm speaking to, so I'd be happy to help (any hints on which pieces
need stronger tests is welcome).
Tommaso
2011/3/29 Jörn Kottmann <ko...@gmail.com>
> On 3/29/11 7:27 PM, Grant Ingersoll wrote:
>
>> On Mar 29, 2011, at 12:57 PM, Jörn Kottmann wrote:
>>
>> On 3/29/11 6:03 PM, Grant Ingersoll wrote:
>>>
>>>> Overall, artifacts look good.
>>>>
>>>> From a packaging standpoint:
>>>>
>>>> * NOTICE.txt needs the JWNL copyright statement in it per the terms of
>>>> BSD
>>>>
>>>> * UIMA reference should also be in the NOTICE (as should any other
>>>> dependency that requires it). Technically, since we aren't redistributing
>>>> the library, I don't think need to, but I think we should for completeness.
>>>>
>>>> How come the UIMA stuff isn't in the binary? Just curious.
>>>>
>>> There are different ways to run OpenNLP in UIMA, one is to create the
>>> sample pear
>>> and start it up in the uima document analyzer, but that one is
>>> distributed with UIMA.
>>>
>>> Anyway I will fix the NOTICE file as suggested by Grant for RC 6, I hope
>>> we are then
>>> ready to release,
>>>
>> With the next candidate, you should call a vote. We need at least 3 PPMC
>> votes that are +1 and no -1's. Votes from others should be taken seriously
>> as well. I typically treat a -1 from anyone as binding or at least
>> warranting further investigation. Once the PPMC vote passes, then we must
>> ask the Incubator PMC to vote. This is only needed while in the Incubator.
>>
>> I would also recommend reading http://www.apache.org/dev/release.html
>>
>
> Thanks, through my work on UIMA I am aware with this process. The reason I
> did not ask for the vote yet was that a little testing
> is still missing and that I wanted the packaging being checked. If every
> thing goes well I will call the for the vote next week on RC6.
>
> Jörn
>
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Jörn Kottmann <ko...@gmail.com>.
On 3/29/11 7:27 PM, Grant Ingersoll wrote:
> On Mar 29, 2011, at 12:57 PM, Jörn Kottmann wrote:
>
>> On 3/29/11 6:03 PM, Grant Ingersoll wrote:
>>> Overall, artifacts look good.
>>>
>>> From a packaging standpoint:
>>>
>>> * NOTICE.txt needs the JWNL copyright statement in it per the terms of BSD
>>>
>>> * UIMA reference should also be in the NOTICE (as should any other dependency that requires it). Technically, since we aren't redistributing the library, I don't think need to, but I think we should for completeness.
>>>
>>> How come the UIMA stuff isn't in the binary? Just curious.
>> There are different ways to run OpenNLP in UIMA, one is to create the sample pear
>> and start it up in the uima document analyzer, but that one is distributed with UIMA.
>>
>> Anyway I will fix the NOTICE file as suggested by Grant for RC 6, I hope we are then
>> ready to release,
> With the next candidate, you should call a vote. We need at least 3 PPMC votes that are +1 and no -1's. Votes from others should be taken seriously as well. I typically treat a -1 from anyone as binding or at least warranting further investigation. Once the PPMC vote passes, then we must ask the Incubator PMC to vote. This is only needed while in the Incubator.
>
> I would also recommend reading http://www.apache.org/dev/release.html
Thanks, through my work on UIMA I am aware with this process. The reason
I did not ask for the vote yet was that a little testing
is still missing and that I wanted the packaging being checked. If every
thing goes well I will call the for the vote next week on RC6.
Jörn
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Grant Ingersoll <gs...@apache.org>.
On Mar 29, 2011, at 12:57 PM, Jörn Kottmann wrote:
> On 3/29/11 6:03 PM, Grant Ingersoll wrote:
>> Overall, artifacts look good.
>>
>> From a packaging standpoint:
>>
>> * NOTICE.txt needs the JWNL copyright statement in it per the terms of BSD
>>
>> * UIMA reference should also be in the NOTICE (as should any other dependency that requires it). Technically, since we aren't redistributing the library, I don't think need to, but I think we should for completeness.
>>
>> How come the UIMA stuff isn't in the binary? Just curious.
> There are different ways to run OpenNLP in UIMA, one is to create the sample pear
> and start it up in the uima document analyzer, but that one is distributed with UIMA.
>
> Anyway I will fix the NOTICE file as suggested by Grant for RC 6, I hope we are then
> ready to release,
With the next candidate, you should call a vote. We need at least 3 PPMC votes that are +1 and no -1's. Votes from others should be taken seriously as well. I typically treat a -1 from anyone as binding or at least warranting further investigation. Once the PPMC vote passes, then we must ask the Incubator PMC to vote. This is only needed while in the Incubator.
I would also recommend reading http://www.apache.org/dev/release.html
-Grant
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Jörn Kottmann <ko...@gmail.com>.
On 3/29/11 6:03 PM, Grant Ingersoll wrote:
> Overall, artifacts look good.
>
> From a packaging standpoint:
>
> * NOTICE.txt needs the JWNL copyright statement in it per the terms of BSD
>
> * UIMA reference should also be in the NOTICE (as should any other dependency that requires it). Technically, since we aren't redistributing the library, I don't think need to, but I think we should for completeness.
>
> How come the UIMA stuff isn't in the binary? Just curious.
There are different ways to run OpenNLP in UIMA, one is to create the
sample pear
and start it up in the uima document analyzer, but that one is
distributed with UIMA.
Anyway I will fix the NOTICE file as suggested by Grant for RC 6, I hope
we are then
ready to release, from the testing front everything looks good, expect a
few minor tests
which should still be done. I am a little worried about the UIMA
training support, since
I do not have real test data for it, but I believe it is in a good state
(which doesn't mean it really is).
Anyway the UIMA training support was released like that before, and we
might postpone proper testing
to our next release (I know that doesn't sounds nice).
Jörn
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Jörn Kottmann <ko...@gmail.com>.
On 3/29/11 6:03 PM, Grant Ingersoll wrote:
> Overall, artifacts look good.
>
> From a packaging standpoint:
>
> * NOTICE.txt needs the JWNL copyright statement in it per the terms of BSD
>
> * UIMA reference should also be in the NOTICE (as should any other dependency that requires it). Technically, since we aren't redistributing the library, I don't think need to, but I think we should for completeness.
Please review the updated NOTICE file, that is the corresponding jira issue:
https://issues.apache.org/jira/browse/OPENNLP-148
Thanks,
Jörn
Re: OpenNLP 1.5.1 Release Candidate 5
Posted by Grant Ingersoll <gs...@apache.org>.
Overall, artifacts look good.
From a packaging standpoint:
* NOTICE.txt needs the JWNL copyright statement in it per the terms of BSD
* UIMA reference should also be in the NOTICE (as should any other dependency that requires it). Technically, since we aren't redistributing the library, I don't think need to, but I think we should for completeness.
How come the UIMA stuff isn't in the binary? Just curious.
-Grant
On Mar 21, 2011, at 6:12 AM, Jörn Kottmann wrote:
> Hello,
>
> our next RC is ready for testing, compared to RC 4 it fixes
> a bug in the creation of the sample UIMA PEAR package (OPENNLP-143).
>
> It can be found here:
> http://people.apache.org/~joern/releases/opennlp-1.5.1-incubating/rc5/
>
> Testing is now mostly done, and we are ready to release soon.
> See our test plan for details:
> https://cwiki.apache.org/confluence/display/OPENNLP/TestPlan1.5.1
>
> Jörn
>