Maven build problem: blueprint jar missing
Eric
When trying to install Gremlin with Maven (zip file downloaded with
the 'Download Source' button) I received the following message: "Your
project has dependencies that are not resolved locally". Apparently
it's the "com.tinkerpop:blueprints:jar:0.1.2-SNAPSHOT" that is
missing.
Can you instruct me on how to proceed?
Thanks,
E.
Marko Rodriguez
So I deleted my local m2 tinkerpop.com repository. I then downloaded Gremlin 0.2.2 zip file from the Gremlin website. I proceeded with:
marko:~$ cd Desktop/gremlin-0.2.2
marko:~/Desktop/gremlin-0.2.2$ ls
data gremlin.sh src
gremlin.bat pom.xml target
marko:~/Desktop/gremlin-0.2.2$ mvn clean install
[INFO] Scanning for projects...
[INFO] ------------------------------------------------------------------------
[INFO] Building Gremlin: A Graph-Based Programming Language
[INFO] task-segment: [clean, install]
[INFO] ------------------------------------------------------------------------
[INFO] [clean:clean {execution: default-clean}]
[INFO] Deleting directory /Users/marko/Desktop/gremlin-0.2.2/target
[INFO] [resources:resources {execution: default-resources}]
...................................
...................................
[INFO] Installing /Users/marko/Desktop/gremlin-0.2.2/target/gremlin-0.2.2.zip to /Users/marko/.m2/repository/com/tinkerpop/gremlin/0.2.2/gremlin-0.2.2.zip
[INFO] ------------------------------------------------------------------------
[INFO] BUILD SUCCESSFUL
[INFO] ------------------------------------------------------------------------
[INFO] Total time: 36 seconds
[INFO] Finished at: Tue Jun 08 09:04:39 MDT 2010
[INFO] Final Memory: 42M/88M
[INFO] ------------------------------------------------------------------------
marko:~/Desktop/gremlin-0.2.2$
In other words---it worked for me. Can you provided me your mvn output so I can see where exactly its going wrong in your build sequence.
Finally, Blueprints 0.1.2-SNAPSHOT is currently in the TinkerPop maven2 repository:
http://tinkerpop.com/maven2/com/tinkerpop/blueprints/0.1.2-SNAPSHOT/maven-metadata.xml
?? Sorry, I don't have an immediate answer for your. Again, if you could provide your maven's output, I can get more information.
Thanks,
Marko.
http://markorodriguez.com
http://pipes.tinkerpop.com
Eric
I found some leftovers from an unsuccessful installation in the
tinkerpop/blueprints/0.1.2.snapshot folder. After removing this folder
Maven succeeded in reinstalling the missing jar.
Problem solved.
Thank you,
E.
> ?? Sorry, I don't have an immediate answer for your. Again, if you could provide your maven's output, I can get more information.
>
> Thanks,
> Marko.
>
> On Jun 8, 2010, at 4:19 AM, Eric wrote:
>
>
>
> > Hello,
> > When trying to install Gremlin with Maven (zip file downloaded with
> > the 'Download Source' button) I received the following message: "Your
> > project has dependencies that are not resolved locally". Apparently
> > it's the "com.tinkerpop:blueprints:jar:0.1.2-SNAPSHOT" that is
> > missing.
> > Can you instruct me on how to proceed?
> > Thanks,
>
> - Tekst uit oorspronkelijk bericht weergeven -
Marko Rodriguez
Good.
Question: were you trying to build Blueprints from scratch? If so, you have to do:
mvn clean install -Dmaven.test.skip=true
or else have MongoDB installed because the Object Document Model works over MongoDB. Building Blueprints is a bit complicated...
Good luck with Gremlin,
Marko.
http://markorodriguez.com
http://tinkerpop.com
Eric Van Horenbeeck
After a final hassle with MS Vista I managed the gremlin.jars to pass all
the tests. For your information I include a zip file with the faulty
blueprint snapshot files that are/were generated on my machine at every
install from scratch. I installed MongoDB as you suggested. I didn't know
this product. From the 'Future Functionality' wiki tab I learn that you
removed the graph model for MongoDB, but continue to use its key/value
stores. Can you please explain what this means for the persistence layer?
Any plans on HyperGraphDB?
I'm a computational linguist and I will try to implement work I did for my
PhD on topical facets (http://users.telenet.be/ericvanhorenbeeck/) starting
at the bottom of the learning curve.
Keep you posted,
Regards,
Eric
Hi,
Good.
mvn clean install -Dmaven.test.skip=true
http://markorodriguez.com
http://tinkerpop.com
__________ Information from ESET NOD32 Antivirus, version of virus signature
database 5182 (20100608) __________
The message was checked by ESET NOD32 Antivirus.
__________ Information from ESET NOD32 Antivirus, version of virus signature
database 5182 (20100608) __________
The message was checked by ESET NOD32 Antivirus.
__________ Information from ESET NOD32 Antivirus, version of virus signature
database 5183 (20100608) __________
The message was checked by ESET NOD32 Antivirus.
__________ Information from ESET NOD32 Antivirus, version of virus signature
database 5184 (20100609) __________
The message was checked by ESET NOD32 Antivirus.
__________ Information from ESET NOD32 Antivirus, version of virus signature
database 5184 (20100609) __________
The message was checked by ESET NOD32 Antivirus.
__________ Information from ESET NOD32 Antivirus, version of virus signature
database 5185 (20100609) __________
The message was checked by ESET NOD32 Antivirus.
__________ Information from ESET NOD32 Antivirus, version of virus signature
database 5185 (20100609) __________
The message was checked by ESET NOD32 Antivirus.
Marko Rodriguez
> After a final hassle with MS Vista I managed the gremlin.jars to pass all
> the tests. For your information I include a zip file with the faulty
> blueprint snapshot files that are/were generated on my machine at every
> install from scratch.
I don't quite understand how you were getting "faulty" snapshots? Can you explain your step-by-step process. If you are building Gremlin from scratch, Maven will just download the Blueprint SNAPSHOTs (and other dependencies) from the various M2 repositories stated in the Gremlin pom.xml.
> I installed MongoDB as you suggested. I didn't know
> this product. From the 'Future Functionality' wiki tab I learn that you
> removed the graph model for MongoDB, but continue to use its key/value
> stores. Can you please explain what this means for the persistence layer?
First---you don't have to build Blueprints from scratch to use Gremlin. Again, when building Gremlin, Maven will grab the pre-built Blueprint jars from the TinkerPop Maven repository. However, if you do decide to build Blueprints from scratch, then yes, you need to have MongoDB installed if you are *not* going to skip the test cases (i.e. mvn install -Dmaven.test.skip=true)
Now, to answer your question---we removed MongoDB as a persistent store for the Property Graph Model (the graph stuff) as its too slow given that MongoDB does not make use of direct references between its documents. However, the reason why MongoDB is still referred to is because Blueprints is not JUST about graphs. It also provides an Object Document Model (a set of interfaces to easily plug and play databases like CouchDB, MongoDB, and TinkerDoc). I haven't done much work in this area, but it exists... See the README.textile for the realization that Blueprints provides both Graph and Document related interfaces -- http://github.com/tinkerpop/blueprints/blob/master/README.textile ...
> Any plans on HyperGraphDB?
Unfortunately, HyperGraphDB is too general and can't be easily subsumed into the Property Graph Model. However, the Property Graph Model can be easily subsumed by HyperGraphDB. Thus, its possible to connect Blueprints to HyperGraphDB, but it would not be using HyperGraphDB to its fullest :(. In other words, Blueprints would effect a binary graph in an n-ary graph system. Note that Ian Holsman (on this mailing list) has written a SAIL connector for HyperGraphDB and thus, if you are into RDF, there is a Blueprints-to-HyperSAIL exist (that Ian also wrote). He can say more on this...
> I'm a computational linguist and I will try to implement work I did for my
> PhD on topical facets (http://users.telenet.be/ericvanhorenbeeck/) starting
> at the bottom of the learning curve.
Cool. Can you say more about this in layman terms?
Hope my answers were clear. If not, please identify any issues with my reply and I will restate those parts.
Thanks for your interest,
Eric
There is not much to add to my story: I download Gremlin, unzip it in
my Program Files folder, and in the NetBeans IDE I start a new "Maven
Project with Existing POM". That's it. Maven downloads everything, as
it should - except for the Blueprint part. I then remove the faulty
Blueprint folder and use an IDE facility that lets me reinstall any
missing dependencies. So, in fact I didn't build Blueprint from
scratch.
Thank you for enlighten me on the database issue. I do not take the
RDF road to text semantics. I think I will start with the Neo4J
flavor. My priority is to get the graph construction right with the
necessary traversal and selection algorithms. I developed Java-code of
labo-quality for the main algos but now I need a more robust and fast
implementation.
The first step is to build an unrestricted text network from documents
where every node is a token type linked to other tokens with a
directed edge. Stopwords are not removed and stemming is not applied
(hence the 'unrestricted' qualification). The result is a 'small-
world' type of network. For an intro see: "Ferrer i Cancho, R., &
Solé, R. V. (2001). 'The Small-World of Human Language'. Proceedings
of the Royal Society London, 268, 2261-65". Nodes and edges have key/
value properties that permit to identify the documents that
contributed to the network. Collections of documents can be formed and
the informative value of words and phrases calculated. Shared concepts
are detected with core extraction and other graph operations. It all
amounts to the multi-valued tensor algebra you described and that’s
how I ended up here.
E.
Marko Rodriguez
... I then remove the faulty
Blueprint folder and use an IDE facility that lets me reinstall any
missing dependencies. So, in fact I didn't build Blueprint from
scratch.
Thank you for enlighten me on the database issue. I do not take the
RDF road to text semantics. I think I will start with the Neo4J
flavor. My priority is to get the graph construction right with the
necessary traversal and selection algorithms. I developed Java-code of
labo-quality for the main algos but now I need a more robust and fast
implementation.
For an intro see: "Ferrer i Cancho, R., &
Solé, R. V. (2001). 'The Small-World of Human Language'. Proceedings
of the Royal Society London, 268, 2261-65". Nodes and edges have key/
value properties that permit to identify the documents that
contributed to the network. Collections of documents can be formed and
the informative value of words and phrases calculated. Shared concepts
are detected with core extraction and other graph operations. It all
amounts to the multi-valued tensor algebra you described and that’s
how I ended up here.
Eric
As promised I removed and reinstalled Gremlin with the new build. It
went off without a hitch this time. No need to tweak manually the
Blueprint folder. All tests passed. A minor inconvenience however with
the Windows Vista security settings: although I installed Gremlin with
Administrator rights, I had to adjust the user permission settings of
the Gremlin folder to allow a 'Clean and Build' of the project.
E.
Marko Rodriguez
> As promised I removed and reinstalled Gremlin with the new build. It
> went off without a hitch this time. No need to tweak manually the
> Blueprint folder. All tests passed.
Great. I wonder if there was some file encoding issue with the last SNAPSHOT build? ... At least it works now.
> A minor inconvenience however with
> the Windows Vista security settings: although I installed Gremlin with
> Administrator rights, I had to adjust the user permission settings of
> the Gremlin folder to allow a 'Clean and Build' of the project.
Hmmm... I dunno. Anyone on a Windows machine know why this might happen?
Thanks Eric, I appreciate you helping out and being the Windows guinea pig.
- uh, just in case that doesn't translate for you---"Windows guinea pig" means "Windows tester" (but has humor, not so boring as 'tester').
Take care,