mahout for news recommendation?

classic Classic list List threaded Threaded
29 messages Options
12
Reply | Threaded
Open this post in threaded view
|

mahout for news recommendation?

Josh-5
I'm working on an experimental web-based feed reader[1], and in our next
release we would like to feature collaborative filtering-based article
recommendation. For starters, articles will be recommended to you based on
how similar they are to other articles that either you or people you're
following have starred. I am just getting started reading up on mahout and
the problem space in general[2], and thought I would inquire here about
whether it would be a good choice for us.
Thanks!
Josh

P.S. Do you guys hang out in an IRC channel by any chance?


[1] http://melkjug.org, http://melkjug.openplans.org/about
[2] http://oreilly.com/catalog/9780596529321/
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Sean Owen
Sounds like a classic application of CF indeed, so, yes it is a fit.

The first step is to decide what your 'users' and 'items' are. Here it is
clearly users, and articles. Then decide whether you have any notion of
'likes', 'loves', 'dislikes' between users and articles, or whether you
simply have an association or no association. Sounds like the latter -
unless users are rating articles?

I only have a sec to describe roughly how to begin with Mahout since I am
away from a proper workstation but we can follow on later.

First make a file with 'userID,itemID' on each line. They can be whatever
you like. Make a FileDataModel with this file. (Unless I am crazy and never
submitted the change, should work with this input format - normally there is
a third element per line, the preference value.

Then make a TanimotoCoefficientSimilarity with this model. Then make a
NearestNUserNeighborhood with these objects. Then make a
GenericUserBasedRecommender with these. Try calling recommend() and see what
happens!

This basic setup can be further tweaked, customized and optimized for your
domain but that is the basic approach.

On Mar 31, 2009 4:07 AM, "Joshua Bronson" <[hidden email]> wrote:

I'm working on an experimental web-based feed reader[1], and in our next
release we would like to feature collaborative filtering-based article
recommendation. For starters, articles will be recommended to you based on
how similar they are to other articles that either you or people you're
following have starred. I am just getting started reading up on mahout and
the problem space in general[2], and thought I would inquire here about
whether it would be a good choice for us.
Thanks!
Josh

P.S. Do you guys hang out in an IRC channel by any chance?


[1] http://melkjug.org, http://melkjug.openplans.org/about
[2] http://oreilly.com/catalog/9780596529321/
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Grant Ingersoll-2
Great description, Sean.  I would also throw in: have a look at http://lucene.apache.org/mahout/taste.html 
  and try out the Group Lens demo.

-Grant

On Mar 31, 2009, at 2:16 AM, Sean Owen wrote:

> Sounds like a classic application of CF indeed, so, yes it is a fit.
>
> The first step is to decide what your 'users' and 'items' are. Here  
> it is
> clearly users, and articles. Then decide whether you have any notion  
> of
> 'likes', 'loves', 'dislikes' between users and articles, or whether  
> you
> simply have an association or no association. Sounds like the latter -
> unless users are rating articles?
>
> I only have a sec to describe roughly how to begin with Mahout since  
> I am
> away from a proper workstation but we can follow on later.
>
> First make a file with 'userID,itemID' on each line. They can be  
> whatever
> you like. Make a FileDataModel with this file. (Unless I am crazy  
> and never
> submitted the change, should work with this input format - normally  
> there is
> a third element per line, the preference value.
>
> Then make a TanimotoCoefficientSimilarity with this model. Then make a
> NearestNUserNeighborhood with these objects. Then make a
> GenericUserBasedRecommender with these. Try calling recommend() and  
> see what
> happens!
>
> This basic setup can be further tweaked, customized and optimized  
> for your
> domain but that is the basic approach.
>
> On Mar 31, 2009 4:07 AM, "Joshua Bronson" <[hidden email]> wrote:
>
> I'm working on an experimental web-based feed reader[1], and in our  
> next
> release we would like to feature collaborative filtering-based article
> recommendation. For starters, articles will be recommended to you  
> based on
> how similar they are to other articles that either you or people  
> you're
> following have starred. I am just getting started reading up on  
> mahout and
> the problem space in general[2], and thought I would inquire here  
> about
> whether it would be a good choice for us.
> Thanks!
> Josh
>
> P.S. Do you guys hang out in an IRC channel by any chance?
>
>
> [1] http://melkjug.org, http://melkjug.openplans.org/about
> [2] http://oreilly.com/catalog/9780596529321/

--------------------------
Grant Ingersoll
http://www.lucidimagination.com/

Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids)  
using Solr/Lucene:
http://www.lucidimagination.com/search

Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Jason Rennie-2
In reply to this post by Josh-5
Hi Josh,

One thing to consider is that CF approaches will typically ignore
"similarity" between items/articles except for implied similarity based on
stars/ratings.  I.e. if you want your model to account for textual
similarity as well as star/rating relations, a basic CF model probably isn't
what you want.  Instead, you might consider jointly solving many
classification problems (one for each user) where the item/article feature
set is the text.  Here's an example I worked on which was a bit more general
(5-star ratings rather than the on/off input it sounds like you have):

http://people.csail.mit.edu/jrennie/papers/ijcai05-preference.pdf

With text, you may need to be a bit careful about the size of the feature
set (words) so that your parameter set doesn't become intractable.

Note that if you want the system to exhibit real-time feedback, Mahout may
not be what you want since it is intended for batch-processing, IIUC.

Jason

On Mon, Mar 30, 2009 at 5:07 PM, Joshua Bronson <[hidden email]> wrote:

> I'm working on an experimental web-based feed reader[1], and in our next
> release we would like to feature collaborative filtering-based article
> recommendation. For starters, articles will be recommended to you based on
> how similar they are to other articles that either you or people you're
> following have starred. I am just getting started reading up on mahout and
> the problem space in general[2], and thought I would inquire here about
> whether it would be a good choice for us.
> Thanks!
> Josh
>
> P.S. Do you guys hang out in an IRC channel by any chance?
>
>
> [1] http://melkjug.org, http://melkjug.openplans.org/about
> [2] http://oreilly.com/catalog/9780596529321/
>



--
Jason Rennie
Research Scientist, ITA Software
http://www.itasoftware.com/
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Grant Ingersoll-2

On Mar 31, 2009, at 9:47 AM, Jason Rennie wrote:

>
> Note that if you want the system to exhibit real-time feedback,  
> Mahout may
> not be what you want since it is intended for batch-processing, IIUC.
>

Just to clarify a little bit, the CF part of Mahout is intended for  
real time, while the other parts (clustering, classification) are batch.
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Jason Rennie-2
Sorry for my misunderstanding.  Thanks for the clarification!

Jason

On Tue, Mar 31, 2009 at 10:22 AM, Grant Ingersoll <[hidden email]>wrote:

>
> On Mar 31, 2009, at 9:47 AM, Jason Rennie wrote:
>
>
>> Note that if you want the system to exhibit real-time feedback, Mahout may
>> not be what you want since it is intended for batch-processing, IIUC.
>>
>>
> Just to clarify a little bit, the CF part of Mahout is intended for real
> time, while the other parts (clustering, classification) are batch.
>



--
Jason Rennie
Research Scientist, ITA Software
http://www.itasoftware.com/
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Tim Bass
Most prior-work in news related classification has been done with
Bayesian classifiers / networks.

I kindly suggest that if you are interested in processing RSS, you use
Bayesian classifiers as your core.



On Tue, Mar 31, 2009 at 10:39 PM, Jason Rennie <[hidden email]> wrote:

> Sorry for my misunderstanding.  Thanks for the clarification!
>
> Jason
>
> On Tue, Mar 31, 2009 at 10:22 AM, Grant Ingersoll <[hidden email]>wrote:
>
>>
>> On Mar 31, 2009, at 9:47 AM, Jason Rennie wrote:
>>
>>
>>> Note that if you want the system to exhibit real-time feedback, Mahout may
>>> not be what you want since it is intended for batch-processing, IIUC.
>>>
>>>
>> Just to clarify a little bit, the CF part of Mahout is intended for real
>> time, while the other parts (clustering, classification) are batch.
>>
>
>
>
> --
> Jason Rennie
> Research Scientist, ITA Software
> http://www.itasoftware.com/
>
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Vinicius Carvalho-3
"Just to clarify a little bit, the CF part of Mahout is intended for real
time, while the other parts (clustering, classification) are batch."

Sorry to just bump in the discussion. I've started with Taste a few months
ago to use in my MD project. I've found mahout and I'm still studying hadoop
first.

My question is: So the item recommender inside mahout runs on a single node?
Or does it uses the map-reduce features from hadoop?

Sorry for the dumb question.

Regards

On Tue, Mar 31, 2009 at 1:32 PM, Tim Bass <[hidden email]> wrote:

> Most prior-work in news related classification has been done with
> Bayesian classifiers / networks.
>
> I kindly suggest that if you are interested in processing RSS, you use
> Bayesian classifiers as your core.
>
>
>
> On Tue, Mar 31, 2009 at 10:39 PM, Jason Rennie <[hidden email]> wrote:
> > Sorry for my misunderstanding.  Thanks for the clarification!
> >
> > Jason
> >
> > On Tue, Mar 31, 2009 at 10:22 AM, Grant Ingersoll <[hidden email]
> >wrote:
> >
> >>
> >> On Mar 31, 2009, at 9:47 AM, Jason Rennie wrote:
> >>
> >>
> >>> Note that if you want the system to exhibit real-time feedback, Mahout
> may
> >>> not be what you want since it is intended for batch-processing, IIUC.
> >>>
> >>>
> >> Just to clarify a little bit, the CF part of Mahout is intended for real
> >> time, while the other parts (clustering, classification) are batch.
> >>
> >
> >
> >
> > --
> > Jason Rennie
> > Research Scientist, ITA Software
> > http://www.itasoftware.com/
> >
>



--
The intuitive mind is a sacred gift and the
rational mind is a faithful servant. We have
created a society that honors the servant and
has forgotten the gift.
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Otis Gospodnetic-2

it's the former.  Taste is still not parallelized, but other parts of Mahout are, and they make use of Hadoop.

 Otis
--
Sematext -- http://sematext.com/ -- Lucene - Solr - Nutch



----- Original Message ----

> From: Vinicius Carvalho <[hidden email]>
> To: [hidden email]
> Sent: Tuesday, March 31, 2009 12:57:35 PM
> Subject: Re: mahout for news recommendation?
>
> "Just to clarify a little bit, the CF part of Mahout is intended for real
> time, while the other parts (clustering, classification) are batch."
>
> Sorry to just bump in the discussion. I've started with Taste a few months
> ago to use in my MD project. I've found mahout and I'm still studying hadoop
> first.
>
> My question is: So the item recommender inside mahout runs on a single node?
> Or does it uses the map-reduce features from hadoop?
>
> Sorry for the dumb question.
>
> Regards
>
> On Tue, Mar 31, 2009 at 1:32 PM, Tim Bass wrote:
>
> > Most prior-work in news related classification has been done with
> > Bayesian classifiers / networks.
> >
> > I kindly suggest that if you are interested in processing RSS, you use
> > Bayesian classifiers as your core.
> >
> >
> >
> > On Tue, Mar 31, 2009 at 10:39 PM, Jason Rennie wrote:
> > > Sorry for my misunderstanding.  Thanks for the clarification!
> > >
> > > Jason
> > >
> > > On Tue, Mar 31, 2009 at 10:22 AM, Grant Ingersoll
> > >wrote:
> > >
> > >>
> > >> On Mar 31, 2009, at 9:47 AM, Jason Rennie wrote:
> > >>
> > >>
> > >>> Note that if you want the system to exhibit real-time feedback, Mahout
> > may
> > >>> not be what you want since it is intended for batch-processing, IIUC.
> > >>>
> > >>>
> > >> Just to clarify a little bit, the CF part of Mahout is intended for real
> > >> time, while the other parts (clustering, classification) are batch.
> > >>
> > >
> > >
> > >
> > > --
> > > Jason Rennie
> > > Research Scientist, ITA Software
> > > http://www.itasoftware.com/
> > >
> >
>
>
>
> --
> The intuitive mind is a sacred gift and the
> rational mind is a faithful servant. We have
> created a society that honors the servant and
> has forgotten the gift.

Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Sean Owen
Couple clarifications -

The CF components are oriented to on-line, real-time use, though of course
one can trivially build a batch job out of that. That is what I did with the
EC2 image that cranks out recommendations for all users.

The CF component is also already parallelized as much as is practical. There
are already Hadoop jobs for parallel, batch operation.

Finally if you have some external notion of item similarity, like text
similarity between articles, you can and should include this info by
creating an ItemSimilarity with this knowledge. In that case you want to use
an item-based recommender, since it is only in such a case that item-based
recommenders have a distinct advantage.

On Apr 1, 2009 10:32 AM, "Otis Gospodnetic" <[hidden email]>
wrote:


it's the former.  Taste is still not parallelized, but other parts of Mahout
are, and they make use of Hadoop.

 Otis
--
Sematext -- http://sematext.com/ -- Lucene - Solr - Nutch

----- Original Message ---- > From: Vinicius Carvalho <
[hidden email]> > To: mahout-...

> On Tue, Mar 31, 2009 at 1:32 PM, Tim Bass wrote: > > > Most prior-work in
news related classifica...

> > On Tue, Mar 31, 2009 at 10:39 PM, Jason Rennie wrote: > > > Sorry for my
misunderstanding. Than...

> > >wrote: > > > > > >> > > >> On Mar 31, 2009, at 9:47 AM, Jason Rennie
wrote: > > >> > > >> > > >...
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Josh-5
Thanks all for the good info. Taste definitely sounds like a promising
direction for us to go in for our recommendation service.
I'm working through the installation of the GroupLens demo, but the mvn
package step is failing with the output at
http://paste.pocoo.org/show/110618/. Haven't looked into this yet, just
thought I'd post to the list first with my progress. If anyone else uses
IRC, I've created (and am currently the only one in) the #mahout channel on
freenode. Hope to see some of you in there!

Josh

On Wed, Apr 1, 2009 at 5:48 AM, Sean Owen <[hidden email]> wrote:

> Couple clarifications -
>
> The CF components are oriented to on-line, real-time use, though of course
> one can trivially build a batch job out of that. That is what I did with
> the
> EC2 image that cranks out recommendations for all users.
>
> The CF component is also already parallelized as much as is practical.
> There
> are already Hadoop jobs for parallel, batch operation.
>
> Finally if you have some external notion of item similarity, like text
> similarity between articles, you can and should include this info by
> creating an ItemSimilarity with this knowledge. In that case you want to
> use
> an item-based recommender, since it is only in such a case that item-based
> recommenders have a distinct advantage.
>
> On Apr 1, 2009 10:32 AM, "Otis Gospodnetic" <[hidden email]>
> wrote:
>
>
> it's the former.  Taste is still not parallelized, but other parts of
> Mahout
> are, and they make use of Hadoop.
>
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Grant Ingersoll-2
Do a "mvn install" from the top level directory first: http://lucene.apache.org/mahout/taste.html#demo

HTH,
Grant

On Apr 1, 2009, at 11:35 AM, Joshua Bronson wrote:

> Thanks all for the good info. Taste definitely sounds like a promising
> direction for us to go in for our recommendation service.
> I'm working through the installation of the GroupLens demo, but the  
> mvn
> package step is failing with the output at
> http://paste.pocoo.org/show/110618/. Haven't looked into this yet,  
> just
> thought I'd post to the list first with my progress. If anyone else  
> uses
> IRC, I've created (and am currently the only one in) the #mahout  
> channel on
> freenode. Hope to see some of you in there!
>
> Josh
>
> On Wed, Apr 1, 2009 at 5:48 AM, Sean Owen <[hidden email]> wrote:
>
>> Couple clarifications -
>>
>> The CF components are oriented to on-line, real-time use, though of  
>> course
>> one can trivially build a batch job out of that. That is what I did  
>> with
>> the
>> EC2 image that cranks out recommendations for all users.
>>
>> The CF component is also already parallelized as much as is  
>> practical.
>> There
>> are already Hadoop jobs for parallel, batch operation.
>>
>> Finally if you have some external notion of item similarity, like  
>> text
>> similarity between articles, you can and should include this info by
>> creating an ItemSimilarity with this knowledge. In that case you  
>> want to
>> use
>> an item-based recommender, since it is only in such a case that  
>> item-based
>> recommenders have a distinct advantage.
>>
>> On Apr 1, 2009 10:32 AM, "Otis Gospodnetic" <[hidden email]
>> >
>> wrote:
>>
>>
>> it's the former.  Taste is still not parallelized, but other parts of
>> Mahout
>> are, and they make use of Hadoop.
>>

--------------------------
Grant Ingersoll
http://www.lucidimagination.com/

Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids)  
using Solr/Lucene:
http://www.lucidimagination.com/search

Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Josh-5
You mean you're supposed to do step 4 *before* step 8?!? ;p
I did run mvn install, and thought I got a bunch of warnings like the
following:

[WARNING] Entry:
> mahout-0.2-SNAPSHOT/usr/local/melk/mahout/core/src/main/java/org/apache/mahout/cf/taste/impl/common/
> longer than 100 characters.
>

after a couple hours it said it completed successfully:

[INFO]
> ------------------------------------------------------------------------
>
[INFO] Reactor Summary:
>
[INFO]
> ------------------------------------------------------------------------
>
[INFO] Mahout core ........................................... SUCCESS
> [8:46.665s]
>
[INFO] Mahout Taste Webapp ................................... SUCCESS
> [55.496s]
>
[INFO] Mahout examples ....................................... SUCCESS
> [55.317s]
>
[INFO] Apache Lucene Mahout .................................. SUCCESS
> [2:02:03.392s]
>
[INFO]
> ------------------------------------------------------------------------
>
[INFO]
> ------------------------------------------------------------------------
>
[INFO] BUILD SUCCESSFUL
>
[INFO]
> ------------------------------------------------------------------------
>
[INFO] Total time: 132 minutes 41 seconds
>
[INFO] Finished at: Wed Apr 01 00:59:27 EDT 2009
>
[INFO] Final Memory: 61M/80M
>
[INFO]
> ------------------------------------------------------------------------
>


So I proceeded through steps 5, 6, and 7, and then step 8's "mvn package"
command failed with the output I linked to.

Just for the heck of it I tried "mvn install" again (from the top-level
directory) and after getting a bunch of the "longer-than-100-characters"
warnings again, this time after 7 minutes it failed with:

[ERROR] BUILD ERROR
>
[INFO]
> ------------------------------------------------------------------------
>
[INFO] Failed to create assembly: Error creating assembly archive project: A
> tar file cannot include itself.
>


I posted the full transcript of my console session at
http://melkjug.org/_static/grouplens-install-log.txt. Seems like something
funky's going on with tar, but I'm not sure what.


On Wed, Apr 1, 2009 at 12:11 PM, Grant Ingersoll <[hidden email]>wrote:

> Do a "mvn install" from the top level directory first:
> http://lucene.apache.org/mahout/taste.html#demo
>
> HTH,
> Grant
>
>
> On Apr 1, 2009, at 11:35 AM, Joshua Bronson wrote:
>
>  Thanks all for the good info. Taste definitely sounds like a promising
>> direction for us to go in for our recommendation service.
>> I'm working through the installation of the GroupLens demo, but the mvn
>> package step is failing with the output at
>> http://paste.pocoo.org/show/110618/. Haven't looked into this yet, just
>> thought I'd post to the list first with my progress. If anyone else uses
>> IRC, I've created (and am currently the only one in) the #mahout channel
>> on
>> freenode. Hope to see some of you in there!
>>
>> Josh
>>
>> On Wed, Apr 1, 2009 at 5:48 AM, Sean Owen <[hidden email]> wrote:
>>
>>  Couple clarifications -
>>>
>>> The CF components are oriented to on-line, real-time use, though of
>>> course
>>> one can trivially build a batch job out of that. That is what I did with
>>> the
>>> EC2 image that cranks out recommendations for all users.
>>>
>>> The CF component is also already parallelized as much as is practical.
>>> There
>>> are already Hadoop jobs for parallel, batch operation.
>>>
>>> Finally if you have some external notion of item similarity, like text
>>> similarity between articles, you can and should include this info by
>>> creating an ItemSimilarity with this knowledge. In that case you want to
>>> use
>>> an item-based recommender, since it is only in such a case that
>>> item-based
>>> recommenders have a distinct advantage.
>>>
>>> On Apr 1, 2009 10:32 AM, "Otis Gospodnetic" <[hidden email]>
>>> wrote:
>>>
>>>
>>> it's the former.  Taste is still not parallelized, but other parts of
>>> Mahout
>>> are, and they make use of Hadoop.
>>>
>>>
> --------------------------
> Grant Ingersoll
> http://www.lucidimagination.com/
>
> Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids) using
> Solr/Lucene:
> http://www.lucidimagination.com/search
>
>
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Josh-5
Oh, it just occurred to me I should have said that /usr/local/melk/mahout is
a checkout of http://svn.apache.org/repos/asf/lucene/mahout/trunk/. When I
got to step 2, which says...

> 2. Unpack the archive and copy movies.dat and ratings.dat to
> trunk/taste-web/src/main/resources/org/apache/mahout/cf/taste/example/grouplens under
> the Mahout distribution directory.


...I assumed the instructions had left out the step of running "svn
checkout http://svn.apache.org/repos/asf/lucene/mahout/trunk/". Was this
assumption incorrect?

I did have to "mkdir
-p trunk/taste-web/src/main/resources/org/apache/mahout/cf/taste/example/grouplens"
before I could copy the .dat files there as
the trunk/taste-web/src/main/resources directory of the checkout doesn't
contain anything in it. Did I go off on the wrong track?

On Wed, Apr 1, 2009 at 4:27 PM, Joshua Bronson <[hidden email]> wrote:

> You mean you're supposed to do step 4 *before* step 8?!? ;p
> I did run mvn install, and though I got a bunch of warnings like the
> following:
>
> [WARNING] Entry:
>> mahout-0.2-SNAPSHOT/usr/local/melk/mahout/core/src/main/java/org/apache/mahout/cf/taste/impl/common/
>> longer than 100 characters.
>>
>
> after a couple hours it said it completed successfully:
>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO] Reactor Summary:
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO] Mahout core ........................................... SUCCESS
>> [8:46.665s]
>>
> [INFO] Mahout Taste Webapp ................................... SUCCESS
>> [55.496s]
>>
> [INFO] Mahout examples ....................................... SUCCESS
>> [55.317s]
>>
> [INFO] Apache Lucene Mahout .................................. SUCCESS
>> [2:02:03.392s]
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO] BUILD SUCCESSFUL
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO] Total time: 132 minutes 41 seconds
>>
> [INFO] Finished at: Wed Apr 01 00:59:27 EDT 2009
>>
> [INFO] Final Memory: 61M/80M
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
>
>
> So I proceeded through steps 5, 6, and 7, and then step 8's "mvn package"
> command failed with the output I linked to.
>
> Just for the heck of it I tried "mvn install" again (from the top-level
> directory) and after getting a bunch of the "longer-than-100-characters"
> warnings again, this time after 7 minutes it failed with:
>
> [ERROR] BUILD ERROR
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO] Failed to create assembly: Error creating assembly archive project:
>> A tar file cannot include itself.
>>
>
>
> I posted the full transcript of my console session at
> http://melkjug.org/_static/grouplens-install-log.txt. Seems like something
> funky's going on with tar, but I'm not sure what.
>
>
> On Wed, Apr 1, 2009 at 12:11 PM, Grant Ingersoll <[hidden email]>wrote:
>
>> Do a "mvn install" from the top level directory first:
>> http://lucene.apache.org/mahout/taste.html#demo
>>
>> HTH,
>> Grant
>>
>>
>> On Apr 1, 2009, at 11:35 AM, Joshua Bronson wrote:
>>
>>  Thanks all for the good info. Taste definitely sounds like a promising
>>> direction for us to go in for our recommendation service.
>>> I'm working through the installation of the GroupLens demo, but the mvn
>>> package step is failing with the output at
>>> http://paste.pocoo.org/show/110618/. Haven't looked into this yet, just
>>> thought I'd post to the list first with my progress. If anyone else uses
>>> IRC, I've created (and am currently the only one in) the #mahout channel
>>> on
>>> freenode. Hope to see some of you in there!
>>>
>>> Josh
>>>
>>> On Wed, Apr 1, 2009 at 5:48 AM, Sean Owen <[hidden email]> wrote:
>>>
>>>  Couple clarifications -
>>>>
>>>> The CF components are oriented to on-line, real-time use, though of
>>>> course
>>>> one can trivially build a batch job out of that. That is what I did with
>>>> the
>>>> EC2 image that cranks out recommendations for all users.
>>>>
>>>> The CF component is also already parallelized as much as is practical.
>>>> There
>>>> are already Hadoop jobs for parallel, batch operation.
>>>>
>>>> Finally if you have some external notion of item similarity, like text
>>>> similarity between articles, you can and should include this info by
>>>> creating an ItemSimilarity with this knowledge. In that case you want to
>>>> use
>>>> an item-based recommender, since it is only in such a case that
>>>> item-based
>>>> recommenders have a distinct advantage.
>>>>
>>>> On Apr 1, 2009 10:32 AM, "Otis Gospodnetic" <[hidden email]
>>>> >
>>>> wrote:
>>>>
>>>>
>>>> it's the former.  Taste is still not parallelized, but other parts of
>>>> Mahout
>>>> are, and they make use of Hadoop.
>>>>
>>>>
>> --------------------------
>> Grant Ingersoll
>> http://www.lucidimagination.com/
>>
>> Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids) using
>> Solr/Lucene:
>> http://www.lucidimagination.com/search
>>
>>
>
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Sean Owen
On Thu, Apr 2, 2009 at 3:58 AM, Joshua Bronson <[hidden email]> wrote:
> ...I assumed the instructions had left out the step of running "svn
> checkout http://svn.apache.org/repos/asf/lucene/mahout/trunk/". Was this
> assumption incorrect?

Well you do need some copy of the Mahout distro, whether from SVN or a
tarball. I imagin the latter is actually more common. The instructions
are indeed silent on this and assume you start with copy of the distro
from some source.

>
> I did have to "mkdir
> -p trunk/taste-web/src/main/resources/org/apache/mahout/cf/taste/example/grouplens"
> before I could copy the .dat files there as
> the trunk/taste-web/src/main/resources directory of the checkout doesn't
> contain anything in it. Did I go off on the wrong track?

Possible, the locations have been moving about and not sure those
changes are in sync with my brain or the documentation. This sounds
right, and I guess it works? then we should make the directory in SVN.
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Grant Ingersoll-2
In reply to this post by Josh-5
What version of Maven do you have?

On Apr 1, 2009, at 4:27 PM, Joshua Bronson wrote:

> You mean you're supposed to do step 4 *before* step 8?!? ;p
> I did run mvn install, and thought I got a bunch of warnings like the
> following:
>
> [WARNING] Entry:
>> mahout-0.2-SNAPSHOT/usr/local/melk/mahout/core/src/main/java/org/
>> apache/mahout/cf/taste/impl/common/
>> longer than 100 characters.
>>
>
> after a couple hours it said it completed successfully:
>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO] Reactor Summary:
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO] Mahout core ........................................... SUCCESS
>> [8:46.665s]
>>
> [INFO] Mahout Taste Webapp ................................... SUCCESS
>> [55.496s]
>>
> [INFO] Mahout examples ....................................... SUCCESS
>> [55.317s]
>>
> [INFO] Apache Lucene Mahout .................................. SUCCESS
>> [2:02:03.392s]
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO] BUILD SUCCESSFUL
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO] Total time: 132 minutes 41 seconds
>>
> [INFO] Finished at: Wed Apr 01 00:59:27 EDT 2009
>>
> [INFO] Final Memory: 61M/80M
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
>
>
> So I proceeded through steps 5, 6, and 7, and then step 8's "mvn  
> package"
> command failed with the output I linked to.
>
> Just for the heck of it I tried "mvn install" again (from the top-
> level
> directory) and after getting a bunch of the "longer-than-100-
> characters"
> warnings again, this time after 7 minutes it failed with:
>
> [ERROR] BUILD ERROR
>>
> [INFO]
>> ------------------------------------------------------------------------
>>
> [INFO] Failed to create assembly: Error creating assembly archive  
> project: A
>> tar file cannot include itself.
>>
>
>
> I posted the full transcript of my console session at
> http://melkjug.org/_static/grouplens-install-log.txt. Seems like  
> something
> funky's going on with tar, but I'm not sure what.
>
>
> On Wed, Apr 1, 2009 at 12:11 PM, Grant Ingersoll  
> <[hidden email]>wrote:
>
>> Do a "mvn install" from the top level directory first:
>> http://lucene.apache.org/mahout/taste.html#demo
>>
>> HTH,
>> Grant
>>
>>
>> On Apr 1, 2009, at 11:35 AM, Joshua Bronson wrote:
>>
>> Thanks all for the good info. Taste definitely sounds like a  
>> promising
>>> direction for us to go in for our recommendation service.
>>> I'm working through the installation of the GroupLens demo, but  
>>> the mvn
>>> package step is failing with the output at
>>> http://paste.pocoo.org/show/110618/. Haven't looked into this yet,  
>>> just
>>> thought I'd post to the list first with my progress. If anyone  
>>> else uses
>>> IRC, I've created (and am currently the only one in) the #mahout  
>>> channel
>>> on
>>> freenode. Hope to see some of you in there!
>>>
>>> Josh
>>>
>>> On Wed, Apr 1, 2009 at 5:48 AM, Sean Owen <[hidden email]> wrote:
>>>
>>> Couple clarifications -
>>>>
>>>> The CF components are oriented to on-line, real-time use, though of
>>>> course
>>>> one can trivially build a batch job out of that. That is what I  
>>>> did with
>>>> the
>>>> EC2 image that cranks out recommendations for all users.
>>>>
>>>> The CF component is also already parallelized as much as is  
>>>> practical.
>>>> There
>>>> are already Hadoop jobs for parallel, batch operation.
>>>>
>>>> Finally if you have some external notion of item similarity, like  
>>>> text
>>>> similarity between articles, you can and should include this info  
>>>> by
>>>> creating an ItemSimilarity with this knowledge. In that case you  
>>>> want to
>>>> use
>>>> an item-based recommender, since it is only in such a case that
>>>> item-based
>>>> recommenders have a distinct advantage.
>>>>
>>>> On Apr 1, 2009 10:32 AM, "Otis Gospodnetic" <[hidden email]
>>>> >
>>>> wrote:
>>>>
>>>>
>>>> it's the former.  Taste is still not parallelized, but other  
>>>> parts of
>>>> Mahout
>>>> are, and they make use of Hadoop.
>>>>
>>>>
>> --------------------------
>> Grant Ingersoll
>> http://www.lucidimagination.com/
>>
>> Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids)  
>> using
>> Solr/Lucene:
>> http://www.lucidimagination.com/search
>>
>>

--------------------------
Grant Ingersoll
http://www.lucidimagination.com/

Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids)  
using Solr/Lucene:
http://www.lucidimagination.com/search

Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Grant Ingersoll-2
Hmm, when I check out clean and do install it works fine.

What platform are you on?


On Apr 2, 2009, at 8:12 AM, Grant Ingersoll wrote:

> What version of Maven do you have?
>
> On Apr 1, 2009, at 4:27 PM, Joshua Bronson wrote:
>
>> You mean you're supposed to do step 4 *before* step 8?!? ;p
>> I did run mvn install, and thought I got a bunch of warnings like the
>> following:
>>
>> [WARNING] Entry:
>>> mahout-0.2-SNAPSHOT/usr/local/melk/mahout/core/src/main/java/org/
>>> apache/mahout/cf/taste/impl/common/
>>> longer than 100 characters.
>>>
>>
>> after a couple hours it said it completed successfully:
>>
>> [INFO]
>>> ------------------------------------------------------------------------
>>>
>> [INFO] Reactor Summary:
>>>
>> [INFO]
>>> ------------------------------------------------------------------------
>>>
>> [INFO] Mahout core ...........................................  
>> SUCCESS
>>> [8:46.665s]
>>>
>> [INFO] Mahout Taste Webapp ...................................  
>> SUCCESS
>>> [55.496s]
>>>
>> [INFO] Mahout examples .......................................  
>> SUCCESS
>>> [55.317s]
>>>
>> [INFO] Apache Lucene Mahout ..................................  
>> SUCCESS
>>> [2:02:03.392s]
>>>
>> [INFO]
>>> ------------------------------------------------------------------------
>>>
>> [INFO]
>>> ------------------------------------------------------------------------
>>>
>> [INFO] BUILD SUCCESSFUL
>>>
>> [INFO]
>>> ------------------------------------------------------------------------
>>>
>> [INFO] Total time: 132 minutes 41 seconds
>>>
>> [INFO] Finished at: Wed Apr 01 00:59:27 EDT 2009
>>>
>> [INFO] Final Memory: 61M/80M
>>>
>> [INFO]
>>> ------------------------------------------------------------------------
>>>
>>
>>
>> So I proceeded through steps 5, 6, and 7, and then step 8's "mvn  
>> package"
>> command failed with the output I linked to.
>>
>> Just for the heck of it I tried "mvn install" again (from the top-
>> level
>> directory) and after getting a bunch of the "longer-than-100-
>> characters"
>> warnings again, this time after 7 minutes it failed with:
>>
>> [ERROR] BUILD ERROR
>>>
>> [INFO]
>>> ------------------------------------------------------------------------
>>>
>> [INFO] Failed to create assembly: Error creating assembly archive  
>> project: A
>>> tar file cannot include itself.
>>>
>>
>>
>> I posted the full transcript of my console session at
>> http://melkjug.org/_static/grouplens-install-log.txt. Seems like  
>> something
>> funky's going on with tar, but I'm not sure what.
>>
>>
>> On Wed, Apr 1, 2009 at 12:11 PM, Grant Ingersoll  
>> <[hidden email]>wrote:
>>
>>> Do a "mvn install" from the top level directory first:
>>> http://lucene.apache.org/mahout/taste.html#demo
>>>
>>> HTH,
>>> Grant
>>>
>>>
>>> On Apr 1, 2009, at 11:35 AM, Joshua Bronson wrote:
>>>
>>> Thanks all for the good info. Taste definitely sounds like a  
>>> promising
>>>> direction for us to go in for our recommendation service.
>>>> I'm working through the installation of the GroupLens demo, but  
>>>> the mvn
>>>> package step is failing with the output at
>>>> http://paste.pocoo.org/show/110618/. Haven't looked into this  
>>>> yet, just
>>>> thought I'd post to the list first with my progress. If anyone  
>>>> else uses
>>>> IRC, I've created (and am currently the only one in) the #mahout  
>>>> channel
>>>> on
>>>> freenode. Hope to see some of you in there!
>>>>
>>>> Josh
>>>>
>>>> On Wed, Apr 1, 2009 at 5:48 AM, Sean Owen <[hidden email]> wrote:
>>>>
>>>> Couple clarifications -
>>>>>
>>>>> The CF components are oriented to on-line, real-time use, though  
>>>>> of
>>>>> course
>>>>> one can trivially build a batch job out of that. That is what I  
>>>>> did with
>>>>> the
>>>>> EC2 image that cranks out recommendations for all users.
>>>>>
>>>>> The CF component is also already parallelized as much as is  
>>>>> practical.
>>>>> There
>>>>> are already Hadoop jobs for parallel, batch operation.
>>>>>
>>>>> Finally if you have some external notion of item similarity,  
>>>>> like text
>>>>> similarity between articles, you can and should include this  
>>>>> info by
>>>>> creating an ItemSimilarity with this knowledge. In that case you  
>>>>> want to
>>>>> use
>>>>> an item-based recommender, since it is only in such a case that
>>>>> item-based
>>>>> recommenders have a distinct advantage.
>>>>>
>>>>> On Apr 1, 2009 10:32 AM, "Otis Gospodnetic" <[hidden email]
>>>>> >
>>>>> wrote:
>>>>>
>>>>>
>>>>> it's the former.  Taste is still not parallelized, but other  
>>>>> parts of
>>>>> Mahout
>>>>> are, and they make use of Hadoop.
>>>>>
>>>>>
>>> --------------------------
>>> Grant Ingersoll
>>> http://www.lucidimagination.com/
>>>
>>> Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids)  
>>> using
>>> Solr/Lucene:
>>> http://www.lucidimagination.com/search
>>>
>>>
>
> --------------------------
> Grant Ingersoll
> http://www.lucidimagination.com/
>
> Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids)  
> using Solr/Lucene:
> http://www.lucidimagination.com/search
>

--------------------------
Grant Ingersoll
http://www.lucidimagination.com/

Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids)  
using Solr/Lucene:
http://www.lucidimagination.com/search

Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Josh-5
 The machine I'm having problems with the demo on is a Macbook with
Apple-distributed java tools:

$ uname -a

Darwin voodoo.openplans.org 9.6.0 Darwin Kernel Version 9.6.0: Mon Nov 24
17:37:00 PST 2008; root:xnu-1228.9.59~1/RELEASE_I386 i386 i386 MacBook2,1
Darwin



$ for i in mvn java javac; do which $i; $i -version; echo; done
/usr/bin/mvn
Maven version: 2.0.6

/usr/bin/java
java version "1.6.0_07"
Java(TM) SE Runtime Environment (build 1.6.0_07-b06-153)
Java HotSpot(TM) 64-Bit Server VM (build 1.6.0_07-b06-57, mixed mode)

/usr/bin/javac
javac 1.6.0_07




I retried the demo on a Gentoo Linux virtual machine, however, and all went
well. Here is the info on that machine:

$ uname -a
Linux dev.melkjug.org 2.6.21-xen #1 SMP Tue May 20 03:08:24 EDT 2008 x86_64
Intel(R) Xeon(R) CPU E5430 @ 2.66GHz GenuineIntel GNU/Linux


$ for i in mvn java javac; do which $i; $i -version; echo; done
/usr/bin/mvn
Maven version: 2.0.9
Java version: 1.6.0_11
OS name: "linux" version: "2.6.21-xen" arch: "amd64" Family: "unix"

/usr/bin/java
java version "1.6.0_11"
Java(TM) SE Runtime Environment (build 1.6.0_11-b03)
Java HotSpot(TM) 64-Bit Server VM (build 11.0-b16, mixed mode)

/usr/bin/javac
javac 1.6.0_11



Note that the difference in the maven versions. Is the demo not compatible
with maven <= 2.0.6?


On Thu, Apr 2, 2009 at 8:51 AM, Grant Ingersoll <[hidden email]> wrote:

> Hmm, when I check out clean and do install it works fine.
>
> What platform are you on?
>
>
>
> On Apr 2, 2009, at 8:12 AM, Grant Ingersoll wrote:
>
>  What version of Maven do you have?
>>
>> On Apr 1, 2009, at 4:27 PM, Joshua Bronson wrote:
>>
>>  You mean you're supposed to do step 4 *before* step 8?!? ;p
>>> I did run mvn install, and though I got a bunch of warnings like the
>>> following:
>>>
>>> [WARNING] Entry:
>>>
>>>>
>>>> mahout-0.2-SNAPSHOT/usr/local/melk/mahout/core/src/main/java/org/apache/mahout/cf/taste/impl/common/
>>>> longer than 100 characters.
>>>>
>>>>
>>> after a couple hours it said it completed successfully:
>>>
>>> [INFO]
>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>>  [INFO] Reactor Summary:
>>>
>>>>
>>>>  [INFO]
>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>>  [INFO] Mahout core ........................................... SUCCESS
>>>
>>>> [8:46.665s]
>>>>
>>>>  [INFO] Mahout Taste Webapp ................................... SUCCESS
>>>
>>>> [55.496s]
>>>>
>>>>  [INFO] Mahout examples ....................................... SUCCESS
>>>
>>>> [55.317s]
>>>>
>>>>  [INFO] Apache Lucene Mahout .................................. SUCCESS
>>>
>>>> [2:02:03.392s]
>>>>
>>>>  [INFO]
>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>>  [INFO]
>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>>  [INFO] BUILD SUCCESSFUL
>>>
>>>>
>>>>  [INFO]
>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>>  [INFO] Total time: 132 minutes 41 seconds
>>>
>>>>
>>>>  [INFO] Finished at: Wed Apr 01 00:59:27 EDT 2009
>>>
>>>>
>>>>  [INFO] Final Memory: 61M/80M
>>>
>>>>
>>>>  [INFO]
>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>>
>>>
>>> So I proceeded through steps 5, 6, and 7, and then step 8's "mvn package"
>>> command failed with the output I linked to.
>>>
>>> Just for the heck of it I tried "mvn install" again (from the top-level
>>> directory) and after getting a bunch of the "longer-than-100-characters"
>>> warnings again, this time after 7 minutes it failed with:
>>>
>>> [ERROR] BUILD ERROR
>>>
>>>>
>>>>  [INFO]
>>>
>>>> ------------------------------------------------------------------------
>>>>
>>>>  [INFO] Failed to create assembly: Error creating assembly archive
>>> project: A
>>>
>>>> tar file cannot include itself.
>>>>
>>>>
>>>
>>> I posted the full transcript of my console session at
>>> http://melkjug.org/_static/grouplens-install-log.txt. Seems like
>>> something
>>> funky's going on with tar, but I'm not sure what.
>>>
>>>
>>> On Wed, Apr 1, 2009 at 12:11 PM, Grant Ingersoll <[hidden email]
>>> >wrote:
>>>
>>>  Do a "mvn install" from the top level directory first:
>>>> http://lucene.apache.org/mahout/taste.html#demo
>>>>
>>>> HTH,
>>>> Grant
>>>>
>>>>
>>>> On Apr 1, 2009, at 11:35 AM, Joshua Bronson wrote:
>>>>
>>>> Thanks all for the good info. Taste definitely sounds like a promising
>>>>
>>>>> direction for us to go in for our recommendation service.
>>>>> I'm working through the installation of the GroupLens demo, but the mvn
>>>>> package step is failing with the output at
>>>>> http://paste.pocoo.org/show/110618/. Haven't looked into this yet,
>>>>> just
>>>>> thought I'd post to the list first with my progress. If anyone else
>>>>> uses
>>>>> IRC, I've created (and am currently the only one in) the #mahout
>>>>> channel
>>>>> on
>>>>> freenode. Hope to see some of you in there!
>>>>>
>>>>> Josh
>>>>>
>>>>> On Wed, Apr 1, 2009 at 5:48 AM, Sean Owen <[hidden email]> wrote:
>>>>>
>>>>> Couple clarifications -
>>>>>
>>>>>>
>>>>>> The CF components are oriented to on-line, real-time use, though of
>>>>>> course
>>>>>> one can trivially build a batch job out of that. That is what I did
>>>>>> with
>>>>>> the
>>>>>> EC2 image that cranks out recommendations for all users.
>>>>>>
>>>>>> The CF component is also already parallelized as much as is practical.
>>>>>> There
>>>>>> are already Hadoop jobs for parallel, batch operation.
>>>>>>
>>>>>> Finally if you have some external notion of item similarity, like text
>>>>>> similarity between articles, you can and should include this info by
>>>>>> creating an ItemSimilarity with this knowledge. In that case you want
>>>>>> to
>>>>>> use
>>>>>> an item-based recommender, since it is only in such a case that
>>>>>> item-based
>>>>>> recommenders have a distinct advantage.
>>>>>>
>>>>>> On Apr 1, 2009 10:32 AM, "Otis Gospodnetic" <
>>>>>> [hidden email]>
>>>>>> wrote:
>>>>>>
>>>>>>
>>>>>> it's the former.  Taste is still not parallelized, but other parts of
>>>>>> Mahout
>>>>>> are, and they make use of Hadoop.
>>>>>>
>>>>>>
>>>>>>  --------------------------
>>>> Grant Ingersoll
>>>> http://www.lucidimagination.com/
>>>>
>>>> Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids) using
>>>> Solr/Lucene:
>>>> http://www.lucidimagination.com/search
>>>>
>>>>
>>>>
>> --------------------------
>> Grant Ingersoll
>> http://www.lucidimagination.com/
>>
>> Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids) using
>> Solr/Lucene:
>> http://www.lucidimagination.com/search
>>
>>
> --------------------------
> Grant Ingersoll
> http://www.lucidimagination.com/
>
> Search the Lucene ecosystem (Lucene/Solr/Nutch/Mahout/Tika/Droids) using
> Solr/Lucene:
> http://www.lucidimagination.com/search
>
>
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Josh-5
In reply to this post by Sean Owen
On Thu, Apr 2, 2009 at 7:11 AM, Sean Owen <[hidden email]> wrote:

> On Thu, Apr 2, 2009 at 3:58 AM, Joshua Bronson <[hidden email]>
> wrote:
> > ...I assumed the instructions had left out the step of running "svn
> > checkout http://svn.apache.org/repos/asf/lucene/mahout/trunk/". Was this
> > assumption incorrect?
>
> Well you do need some copy of the Mahout distro, whether from SVN or
> a tarball. I imagin the latter is actually more common.


Where would you get a tarball, by the way? There is none linked to from
http://lucene.apache.org/mahout/releases.html. As for getting it from SVN,
the "Version Control" link under "Resources" in the sidebar of
http://lucene.apache.org/mahout/ points to a ViewVC instance (
http://svn.apache.org/viewvc/lucene/mahout/). I had to dig around a tiny bit
to find http://svn.apache.org/repos/asf/lucene/mahout/trunk/.


The instructions are indeed silent on this and assume you start with copy of
> the distro from some source.
>

I think it would be helpful to include this step in the demo explicitly.


> I did have to "mkdir
> > -p
> trunk/taste-web/src/main/resources/org/apache/mahout/cf/taste/example/grouplens"
> > before I could copy the .dat files there as
> > the trunk/taste-web/src/main/resources directory of the checkout doesn't
> > contain anything in it. Did I go off on the wrong track?
>
> Possible, the locations have been moving about and not sure those
> changes are in sync with my brain or the documentation. This sounds
> right, and I guess it works? then we should make the directory in SVN.
>

+1.
Reply | Threaded
Open this post in threaded view
|

Re: mahout for news recommendation?

Grant Ingersoll-2
In reply to this post by Josh-5

On Apr 2, 2009, at 2:08 PM, Joshua Bronson wrote:

> The machine I'm having problems with the demo on is a Macbook with
> Apple-distributed java tools:
>
> $ uname -a
>
> Darwin voodoo.openplans.org 9.6.0 Darwin Kernel Version 9.6.0: Mon  
> Nov 24
> 17:37:00 PST 2008; root:xnu-1228.9.59~1/RELEASE_I386 i386 i386  
> MacBook2,1
> Darwin
>
>
>
> $ for i in mvn java javac; do which $i; $i -version; echo; done
> /usr/bin/mvn
> Maven version: 2.0.6

Please upgrade to 2.0.9 or later (2.0.10 is the latest).   By all  
accounts 2.0.6 is a real dog: http://www.lucidimagination.com/search/document/3e7dfef6281482dd/packaging_step_taking_forever_is_this_right

-Grant
12