Camelia, the Perl 6 bug

IRC log for #bioclipse, 2009-11-25

| Channels | #bioclipse index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
06:28 ChanServ joined #bioclipse
06:28 ilogger2 joined #bioclipse
06:28 CIA-57 joined #bioclipse
06:28 miguelrojasch joined #bioclipse
06:28 mgerlich joined #bioclipse
06:28 zarah joined #bioclipse
06:29 sneumann joined #bioclipse
07:32 egonw joined #bioclipse
07:59 sneumann joined #bioclipse
08:01 Gpox joined #bioclipse
08:15 olass joined #bioclipse
08:20 Gpox joined #bioclipse
08:22 samuell joined #bioclipse
08:50 jonalv joined #bioclipse
08:58 edrin joined #bioclipse
09:05 Gpox joined #bioclipse
09:06 miguelrojasch joined #bioclipse
09:08 edrin hi
09:08 zarah hello edrin, you fantastic person you
10:38 samuell egonw: Provided the in-memory RDF Store is created with pellet.createStore(), what would be the best name for the disk-based (but memory cached) pure Java store? pellet.createTDBStore(), pellet.createDiskBasedStore() ... ?
10:38 samuell egonw. Just thinking it should be easy enough for the end user / scientist to understand which to use
10:38 egonw second
10:39 egonw or
10:39 egonw use createStore()
10:39 egonw and rename the current one into: createInMemoryStore()
10:39 egonw that would actually be better
10:39 samuell hmm, infact they will have a little different signatures
10:39 samuell Since the disk-based one takes a path to a directory too
10:39 egonw ah, indeed, well, then keep the same name
10:40 samuell But, it has some variants too... I guess one might want to store the db in a default place if no directory is not provided?
10:41 samuell also.
10:41 samuell Regarding signatures... I was just thinking whether it is a problem to drastically change the behavior of an existing method, in case someone is using it?
10:42 egonw OK
10:42 egonw make createStore() use a temp file, but disk based
10:42 egonw make createInMemoryStore() be the current impl
10:43 egonw make createStore("/Path") access any file for disk based store
10:43 egonw only question...
10:43 egonw should the path be workspace based or not...
10:43 egonw jonalv: ping
10:43 jonalv egonw: mm?
10:43 egonw jonalv: are your databases inside the workspace?
10:43 * jonalv was just going for lunch...
10:43 egonw should a triple store be in the workspace?
10:44 jonalv egonw: uhm,
10:44 samuell egonw: Yup, important question
10:44 jonalv egonw: I think yes to both questions
10:44 egonw OK
10:44 samuell So, is there any existing utility to convert a workspace path into a full path?
10:45 egonw IFile.getPath().getAbsolutePath() or so
10:45 samuell (Since the Jena TDB store constructor takes a full file path)
10:45 samuell Ok
10:45 jonalv egonw: the databases are sored in .hsqldbDatabases in the workspace path
10:45 egonw which is a hidden fodler in the navigator?
10:45 * jonalv is off for lunch now
10:46 jonalv egonw: I guess so yea
10:46 egonw ok
10:46 samuell Hmm... good tip to use a dot ...
10:46 * jonalv @lunch
10:46 egonw samuell: use the workspace path "/.tdbTripleStores/foo.store.ext"
10:46 egonw what ever .ext is used by TDB
10:47 samuell egonw: Yep. the store will be a folder actually
10:47 samuell so no need for ext.
10:48 samuell (A folder with a variety of files, for data, indexes etc.)
12:08 egonw ack
12:40 sneumann joined #bioclipse
12:52 egonw I wrote up a first blog about the SWAT4LS meeting last week Friday: http://chem-bla-ics.blogspot.com/2​009/11/swat4ls-wrapping-up-1.html (more will follow)
12:52 zarah egonw's link is also http://tinyurl.com/yko3p5e
12:56 masak joined #bioclipse
13:00 * jonalv restarts his sad computer
13:00 samuell egonw: Interesting with the Amino acid ontology :)
13:00 egonw yes
13:03 egonw by all, cu all at 10am tomorrow! (I will live coverage Mark's talk on twitter)
13:05 masak egonw: see you!
13:09 jonalv joined #bioclipse
13:14 CIA-57 bioclipse.rdf: Samuel Lampa master * r5260b0f / (8 files in 4 dirs): Added file based RDF Storage, via Jena TDB library - http://bit.ly/4CsDyS
13:14 CIA-57 bioclipse.rdf: Samuel Lampa master * r06604e8 / (5 files):
13:14 CIA-57 bioclipse.rdf: Renamed methods, to comply with discussion with egonw.
13:14 CIA-57 bioclipse.rdf: createStore() use a temp file, but disk based
13:14 CIA-57 bioclipse.rdf: createInMemoryStore() be the current impl
13:14 CIA-57 bioclipse.rdf: createStore("/Path") access any file for disk based store - http://bit.ly/5RulMU
13:26 samuell jonalv: IFile handles directories too, no?
13:27 jonalv masak: uhm how this again?
13:27 masak jonalv: good question.
13:27 masak don't know offhand. try it and see what happens.
13:28 samuell Ok thanks anyway
13:29 samuell Well... now with autocompletion I found out that there is also a IFolder ^^
13:29 Gpox I think it is IFolder
13:29 samuell Yep :)
13:29 samuell Ok, issue resolved.
13:39 samuell jonalv: Will the same signature magic as for IFile, work for IFolder, so that you can use method(String folderPath) when you want to use method(IFolder folderPath)
13:39 samuell ?
13:39 jonalv samuell: probably not...
13:39 jonalv :(
13:40 samuell Ah, Ok
13:40 jonalv maybe we should fix that...
13:40 jonalv :)
13:40 samuell jonalv: Would be nice :)
13:40 samuell Let me know if I can help with anything
13:41 jonalv I don't think I have time for this right now maybe you could write tests for the behavior?
13:42 samuell I have not been too much into testing... but is interested to learn it... so maybe with some guidance from egonw or so, I can do that
13:43 samuell And, it's not a panic thing, but would be needed sooner or later I assume.
13:45 samuell (For now I'll just not let the user decide the path to the triplestore)
14:24 masak hard core hardcode!
14:24 jonalv google trouble with tribles
14:24 zarah jonalv: http://www.benbellabooks.com/bookstore/​cart.php%3Fm%3Dproduct_detail%26p%3D512
14:25 jonalv google trouble with tribbles
14:25 zarah jonalv: http://en.wikipedia.org/wik​i/The_Trouble_With_Tribbles
14:34 masak zarah: google zarah bioclipse
14:34 zarah masak: http://irclog.perlgeek.de/bioclipse/2008-08-11
14:45 egonw samuell: your patches seem to remove the jena jar
14:45 egonw and does not add any tdb jar
14:45 samuell Ok, yeah... but I thought I had fixed that... will check
14:47 CIA-57 bioclipse.rdf: Samuel Lampa master * r9c7df0e / (6 files): Added missing jars, required for Jena TDB - http://bit.ly/5HKZKu
14:47 samuell There we go.
14:47 samuell egonw: You saw my discussion with jonalv regarding file path automagics with IFolder, which does not work at the moment?
14:49 Gpox @pz 747
14:49 zarah http://pele.farmbio.uu.se/cgi-b​in/bugzilla/show_bug.cgi?id=747
14:49 samuell It might be worth waiting with merging my patches until we've fixed that?
14:49 egonw samuell: hald
14:49 egonw samuell: what's the plan?
14:49 egonw jonalv: you're working on IFolder support in the manager API?
14:49 jonalv egonw: no
14:50 samuell egonw: "hald" means?
14:50 egonw half
14:50 samuell Gpox: Thanks :)
14:50 samuell egonw: Ok :)
14:51 samuell egonw: Gpox's link had some relevant info
14:51 samuell (in the comment by masak)
14:52 Gpox @pz 1683
14:52 zarah http://pele.farmbio.uu.se/cgi-bi​n/bugzilla/show_bug.cgi?id=1683
14:54 egonw samuell: so, what patches are you depending on to happen before the bioclipse.rdf stuff?
14:55 samuell Well, there would need to be a standardized way of handling folders in the workspace
14:56 samuell Before this is in place I suppose it is best ...
14:56 samuell ... to require the user to specify the full path from JS.
14:56 samuell But I don't know if that is good enough for committing?
14:57 egonw and using IFile for folder paths is not a feasible workaround until then?
14:57 egonw samuell: btw, you filed a feature request for IFolder support?
14:57 samuell egonw: No... maybe I should... or otherwise the issues that Gpox linked would fix it any way
14:57 Gpox @pz 982, 1683, 747
14:57 zarah [747] UI commands to work with files/... http://tinyurl.com/ykrsvyg
14:57 zarah [982] ResourcePathTransformer only wo... http://tinyurl.com/y9u4844
14:57 zarah [1683] ResourcePathTransformer should... http://tinyurl.com/yjxor7e
14:57 zarah Entire list at http://tinyurl.com/yeo9afj
14:58 samuell And maybe in a better way... using IResource instead of IFolder etc.
14:58 samuell So, possibly uch feature request would be confusing
14:59 samuell Again, thanks Gpox
15:00 samuell If it is possible to use IFile for handling folders, I could go with that for the moment... didn't try that
15:05 samuell (trying now)
15:10 olass samuell: http://www.phurnace.com/blog/the-​eclipse-resources-plug-in-2.html
15:10 zarah olass's link is also http://tinyurl.com/ykewbht
15:12 olass samuell: IContainer is superclass for IFolder and IProject
15:12 olass you probably want that
15:13 samuell olass: Thanks... Yep, but seems I will need to use IFile at the moment though, if I want to use the method signature automagics...
15:13 olass yes
15:13 olass it does not support folders
15:13 olass which is a severe limitation
15:13 olass but an IFile can never be a folder
15:13 samuell Yep... thinking it would be possible to use in a hacky kind of way for the moment...
15:13 olass good luck :)
15:13 samuell Just letting it take care of translating relative paths (inside the workspace) to full paths
15:14 samuell Will see if it is possible.
15:14 samuell Of course hoping to do this in a better way some day.
15:14 samuell thx :)
15:19 samuell jonalv: When doing the IFile automagics, one should use IFile as type both in the annotation, and in the actual method signature in the interface (which makes two places, in the interface file), right?
15:19 jonalv samuell: you can just implement both a method variant takinga String and one taking an IFolder
15:19 jonalv samuell: "in the annotation"?
15:20 jonalv samuell: I am not sure I understand
15:20 samuell the Spring annotation, "params"
15:20 jonalv samuell: that's not  a Spring annotation
15:20 samuell Ok, sry
15:21 samuell But, I'd go something like:
15:21 samuell @Recorded
15:21 samuell @PublishedMethod(
15:21 samuell params = "IFile tdbDirectory",
15:21 samuell methodSummary = "Blabla..."
15:21 samuell )
15:21 samuell public IRDFStore createStore(IFile tdbDirectory);
15:21 jonalv samuell: but yea the params annotattion  parmeter should almost alwyas be the same as the the java one...
15:21 samuell And then implementing createStore(String ... ) ?
15:21 samuell Ok
15:22 jonalv samuell: no you implement the method taking an IFile not the one taking a tring
15:22 samuell Ok... So IFile on all places... and it will translate the string to an IFile anyway?
15:23 jonalv samuell: well not at all places... but yea the whole point is to translate from String to IFIle
15:23 samuell According to http://tinyurl.com/yjv6uar , I understood that I will use "String" somewere
15:24 samuell So, there are three places, for which I wonder which type to use where:
15:24 samuell Interface annotation
15:24 samuell Interface signature
15:24 samuell Implementation signature
15:25 samuell (The two first are in IPelletManager.java in my case, and the third in PelletManager.java)
15:27 jonalv samuell: you must define both methods in the interface
15:27 jonalv both the one takinga Strign and the one taking an IFIle
15:27 samuell Ah, ok
15:27 samuell And implement only one, or both, in the Manager file?
15:28 jonalv you annotate the one takinga String with @publisehdMethod becasue that annoation is for JavaScript and in JavaScript we don't use IFiles but Strings.
15:28 samuell Ah, ok
15:28 jonalv then you implmeent only the one taking an IFile in the manager implmentation. You are free to implment both if you want but you don't need to
15:29 samuell Ok
15:29 samuell Now I think I get it. Thanks!
15:29 jonalv awesome
15:29 * jonalv helps masak with the same thing now.,.. :)
15:29 jonalv afk
15:32 samuell jonalv++
15:33 samuell Trying now...
15:36 samuell Hmm... it is complaining that "The type PelletManager must implement the inherited abstract method IPelletManager.createStore(String)"
15:36 samuell And when I implement it, that one is used instead of the IFile one ...
15:37 samuell Does the file/path specified in the string have to be an existing file?
15:43 samuell jonalv: Which plugin can I look at, where this is in action?
15:43 samuell There seems to be something I'm not getting right..
15:44 jonalv samuell: you manager should not implement you IManger but only IBioclipseManager
15:44 samuell jonalv: Sorry, I didn't get it really?
15:45 jonalv the PelletManger should not implment IPelletManager
15:46 samuell Hmm.... but isn't that how it is done all the time... ? Or do you mean the specific methods?
15:46 samuell ...that we were talking about?
15:47 jonalv samuell: the special thing about NWO is that he manager does not implement the managerinterface
15:47 samuell Hmm.. but "public class PelletManager implements IPelletManager" was there from before...
15:48 samuell Maybe I misunderstand?
15:49 jonalv samuell: I don't think so see first paragraph on this page: http://wiki.bioclipse.net/index.p​hp?title=How_to_make_a_manager_2
15:49 zarah jonalv's link is also http://tinyurl.com/ouun2r
15:50 samuell Hmm...
15:50 samuell IC... but there is some code still written according to the old way then ... ?
15:51 samuell like the pellet manager ...
15:51 jonalv samuell: yes there are a few managers laggin behind
15:51 samuell jonalv: Ok... things are getting clearer :)
16:10 samuell jonalv: Ok... slowly proceeding now. One question: In the so called "third world order", when another plugin is instantiating a manager object, should it be typed with the class itself, or some interface?
16:10 samuell (The pellet manager was instantiated with the IPelletManager, in the owl plugin)
16:10 samuell (IPelletManager as type, I mean)
16:11 samuell Got it working now, in fact :)))
16:11 samuell jonalv++
16:12 jonalv samuell: always program against interfaces
16:13 samuell jonalv: Ok... well, I guess I better ask egonw to participate in updating the plugins to the "3rd world order", so that we get everything right.
16:13 samuell (The rdf related plugins)
16:14 egonw samuell: please file me a bug report
16:14 egonw marked 2.2
16:14 egonw (I thought the rdf stuff was already NWO, but apparently not)
16:14 samuell egonw: So will I
16:14 samuell Ok!
16:16 samuell egonw: So, to get by an error in the owl plugin, I temporarily instantiated it with the class as type (not an interface).
16:28 samuell I have a hard time understanding the bugzilla interface :(
16:28 samuell Is there any way to just browse open issues, without explicitly searching for them?
16:31 samuell egonw: Can't find 2.2 as version in Bugzilla
16:32 samuell egonw: Never mind... found it in "Target milestone"
16:36 samuell egonw: Bug filed. (Feedback on how to write bug reports is highly welcome)
16:36 * samuell goes home for dinner. BBL.
16:42 CIA-57 bioclipse.core: Carl Masak 2.2.x * rc5921ed / (9 files in 5 dirs): moved JsConsoleManager over to 3WO - http://bit.ly/6B5IF2
16:42 CIA-57 bioclipse.core: Carl Masak 2.2.x * rbd752fb / (3 files in 2 dirs): rewired the JsEditor to use hairdresser progress bar - http://bit.ly/6j0Vqc
16:50 jonalv masak++
16:50 masak jonalv++ # thanks for the help!
17:02 egonw masak: why not move it to 3-CPO directly ?
18:27 CIA-57 bioclipse.rdf: Samuel Lampa master * re8cf010 / (4 files in 2 dirs): Changed PelletManager to NWO use of interfaces and renamed variable tdbDirectory to tripleStoreDirectoryPath - http://bit.ly/6F1xVr
18:39 egonw samuell: ping
18:39 samuell egonw: pong
18:40 egonw does the file based model for tdb only work for pellet? or also with just jena?
18:41 samuell It is Jena specific,
18:41 samuell so I guess it is not in any way restricted to Pellet
18:42 egonw that's what I thought...
18:42 samuell Would it be better placed in the rdf feature?
18:42 samuell The file-based store?
18:42 egonw I think we need duplicate stuff here...
18:43 egonw pellet needs a pellet model
18:43 egonw which extends a Jena model or so...
18:43 samuell Ah, ok
18:43 egonw this is possibly also the case for the file backend
18:43 samuell Ok
18:46 samuell brb
18:55 egonw OK, seems to compile fine here...
18:55 egonw samuell: going to merge your patches
18:55 egonw samuell: btw, you were commiting to your master branch, not your 2.2.x branch
19:21 egonw samuell: please do add your name in the copyright lines for classes you worked on
19:38 sneumann joined #bioclipse
19:39 edrin left #bioclipse
19:40 CIA-57 bioclipse.rdf: Egon Willighagen 2.2.x * rfd7c8af / (12 files in 3 dirs): Added test suite for the Pellet manager - http://bit.ly/4CGd1g
19:40 CIA-57 bioclipse.rdf: Egon Willighagen 2.2.x * rbc4c4cb / plugins/net.bioclipse.pellet/src/net/bioc​lipse/pellet/business/IPelletManager.java : Added unit testing annotation - http://bit.ly/5atn32
19:45 samuell back now, uhm
19:46 samuell egonw: My changes will break the owl plugin, unless one line is changed in it (one moment)
19:47 egonw no worries
19:47 egonw the OWL plugin is not finished yet anyway...
19:47 egonw it's not part of the release...
19:47 samuell Ok
19:47 samuell In OWLManager the following line has to be changed:
19:47 samuell PelletManager rdf = new PelletManager();
19:47 egonw moreover, there remains the problem anyway, of using one manager from another
19:47 samuell IPelletManager rdf = new PelletManager();
19:48 samuell I mean... I changed it to
19:48 samuell PelletManager rdf = new PelletManager();
19:48 egonw I think you commited that
19:48 samuell (for the moment)
19:48 egonw yeah, that's the right change
19:49 samuell Ah, yes of course... owl is in the rdf package (or what to call it?)
19:49 samuell Should have realized that of course
19:49 egonw feature
19:49 egonw it should not
19:49 egonw actually
19:50 egonw it's in the pellet feature
19:50 egonw which is fine
19:50 sneumann_ joined #bioclipse
19:50 samuell Ok...
19:50 samuell But the bioclipse.rdf is a .... ?
19:50 samuell Just a git repo?
19:51 egonw yes
19:51 samuell (includes both the rdf and pellet features)
19:51 samuell Ok
19:51 egonw do you know how to rebase?
19:51 samuell Hmm.... not sure
19:51 egonw samuell: yes, indeed
19:51 egonw ok, try with: git pull
19:53 samuell What am I supposed to do, explained in natural language?
19:53 samuell :)
19:53 egonw update your branches (aka forks) for the changes in my branches
19:54 samuell Ah
19:54 samuell So, I should pull from origin then, (which is your repo)?
19:56 egonw yes
19:59 samuell Ok, done
20:09 samuell egonw: Why didn't I have a 2.2.x branch locally, after checking out the bioclipse.rdf repo?
20:09 egonw I think you did
20:09 samuell Is it added rather recently?
20:09 egonw but it did not automatically make a local clone
20:10 samuell Hmm...
20:10 samuell Anyway... so, basically most changes right now go both to the 2.2.x, and master branches?
20:12 * samuell has too little experience with advanced version control :(
20:13 samuell So, thinking about the general strategy you have, for working with branches
20:15 egonw yeah, people regularly merge in changes in 2.2.x branches into master
20:16 samuell Ok!
20:17 CIA-57 bioclipse.rdf: Egon Willighagen 2.2.x * re47d8c3 / (5 files in 3 dirs): Added two missing unit tests - http://bit.ly/86zryX
20:22 CIA-57 bioclipse.rdf: Egon Willighagen 2.2.x * r8efca7c / plugins/net.bioclipse.pellet/src/net/bioc​lipse/pellet/business/IPelletManager.java : Added missing test method annotation - http://bit.ly/8OuoN2
20:35 egonw enough for today
20:35 egonw cu tomorrow
20:37 CIA-57 bioclipse.rdf: Samuel Lampa 2.2.x * re7e16df / plugins/net.bioclipse.pellet/src/net/bioc​lipse/pellet/business/IPelletManager.java : Improved (R2) help message for manager method createStore(String tripleStoreDirectoryPath) - http://bit.ly/8d3Goz
20:38 samuell Ỳes, cu
21:19 egonw joined #bioclipse
21:57 vmrosario joined #bioclipse

| Channels | #bioclipse index | Today | | Search | Google Search | Plain-Text | summary