Camelia, the Perl 6 bug

IRC log for #bioclipse, 2008-10-26

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

All times shown according to UTC.

Time Nick Message
00:15 CIA-52 bioclipse: edrin_t * r7426 /bioclipse2/trunk/plugins/net​.bioclipse.xws4j/.classpath:
00:16 CIA-52 bioclipse: edrin_t * r7427 /bioclipse2/trunk/plugins/net.bioclipse.xws4j/sr​c/net/bioclipse/xws4j/business/IXwsManager.java:
00:18 CIA-52 bioclipse: edrin_t * r7428 /bioclipse2/trunk/plugins/net.bioclipse.xws4j/sr​c/net/bioclipse/xws4j/business/XwsManager.java:
00:21 CIA-52 bioclipse: edrin_t * r7429 /xws/trunk/xws4j-binding/ (.classpath .project):
00:37 CIA-52 bioclipse: edrin_t * r7430 /bioclipse2/trunk/plugins/net.bioclipse.xws4j/ (.classpath jars/xmlbeans-2.4.0/):
01:12 CIA-52 bioclipse: edrin_t * r7431 /xws/trunk/xws4j-binding/src​/net/bioclipse/xws/binding/ (BindingDefinitions.java BindingManager.java):
01:13 CIA-52 bioclipse: edrin_t * r7432 /bioclipse2/trunk/plugins/net.bioclipse.xws4j/ja​rs/xws4j-current/xws4j-binding-26.10.08-1_5.jar:
01:41 CIA-52 bioclipse: edrin_t * r7433 /bioclipse2/trunk/plugins/net.bioclipse.xws4j/ (6 files in 4 dirs):
02:45 CIA-52 bioclipse: edrin_t * r7434 /bioclipse2/trunk/plugins/net.bioclipse.xws4j/ (6 files in 4 dirs):
02:53 CIA-52 bioclipse: edrin_t * r7435 /xws/trunk/xws4j/src/net/biocli​pse/xws/client/disco/Info.java: desc sjhould not be null
03:05 CIA-52 bioclipse: edrin_t * r7436 /xws/trunk/xws4j/src/net/b​ioclipse/xws/client/disco/ (Functions.java Info.java):
03:07 CIA-52 bioclipse: edrin_t * r7437 /bioclipse2/trunk/plugins/net.bioclipse.xws4j/ (.classpath jars/xws4j-current/xws4j-26.10.08-1_5.jar):
03:08 CIA-52 bioclipse: edrin_t * r7438 /bioclipse2/trunk/plugins/net.bioclipse.xws4j/ (.classpath jars/xws4j-current/xws4j-26.10.08-1_5.jar):
09:45 edrin joined #bioclipse
12:30 CIA-52 bioclipse: edrin_t * r7439 /xws/trunk/xws4j-binding/src/net/bioclips​e/xws/binding/tools/BindingCompiler.java: put the strings for batch compiler cmd in ""
12:33 CIA-52 bioclipse: edrin_t * r7440 /bioclipse2/trunk/plugins/net.bioclipse.xws4j/ja​rs/xws4j-current/xws4j-binding-26.10.08-1_5.jar:
12:39 egonw joined #bioclipse
12:39 edrin hi egonw
12:40 egonw hi
12:40 zarah hi egonw
12:41 edrin egonw: the binding tool
12:41 edrin it
12:41 CIA-52 bioclipse: edrin_t * r7441 /bioclipse2/trunk/plugins/net.bioclipse.xws4j/ (5 files in 4 dirs):
12:42 edrin it's current limitations are: does only support xml schemata that allow only one defined root element type...
12:42 egonw only one element to be root?
12:43 egonw or, just one root element?
12:44 edrin egonw: well if you have a xsd that defines two "allowed" differnt root element types (for example "proteinseq" and "dnaseq") this does not work
12:44 edrin if you have a xsd that allows only "proteinseq" or "dnaseq" to be root it is ok
12:45 egonw mmm... that would hamper the reuse of existing schemata's, not?
12:45 edrin egonw: maybe, i dont know... does it happen often that a schema allows different root element types?
12:46 egonw no idea, but I would not even know how to actually restrict only one element to be used as root
12:46 egonw it mostly defines child-parent relations
12:46 egonw not which parent rules them all
12:47 edrin hm, i thought it defines the list of parents that rules them all...
12:47 edrin well... but i think I could add this support, too
12:47 edrin but it would be really complex then and not nice to use from a console...
12:48 edrin egonw: on the other hand so far it seems to work :)
12:48 egonw good :)
12:48 edrin with:
12:48 egonw it's a start :)
12:48 edrin a little example script:
12:48 edrin (in bc2 console!)
12:49 edrin f = xws.getFunction("xws.ayena.de", "getCoffeeCanStatus");
12:49 edrin ios = f.getIoSchemataSync(20000);
12:49 edrin iof = xws.getIoFactory(ios);
12:50 egonw 20seconds?
12:51 egonw that is... getIoSchemataSync takes an timeout in ms?
12:51 edrin .// this autorgenerates a binding and a IoFactory class within a defined directory, compiles them with the eclipse compiler, and load them to a classloader. then it spawns an instance of IoFactory();
12:51 egonw cool :)
12:51 edrin egonw: yes, 20 seconds ... but it takes only 1 or so, depending on xmpp server..
12:51 edrin right now you can then use iof like this:
12:52 egonw oh, I don't worry about that...
12:52 egonw because it really is for building a local data structure...
12:52 egonw which one does not do often, I'd say...
12:52 edrin input = iof.createInputObject();
12:53 edrin eand input can then be used in a very cool way:
12:53 edrin .//of course depending on XML Schemata!
12:53 egonw sure :)
12:53 edrin input.setInputProteinSequence("hhllaacc");
12:53 edrin and then...
12:54 edrin r = f.invokeSync(input, 20000);
12:54 edrin and to get the output object:
12:54 edrin output = iof.getOutputObject(r);
12:54 edrin thus:
12:55 edrin the_result_value = output.getPredictionValue();
12:55 egonw hh?
12:56 edrin hh?
12:56 egonw what residue is that?
12:56 edrin histidine =
12:56 egonw histidine
12:56 egonw sure
12:56 egonw looked weird...
12:56 edrin ;)
12:56 egonw must be the lower case :)
12:56 egonw don't work much with protein sequences...
12:57 edrin the xws4j-binding lib also has a BatchBinder you can use from console to create a binding for a function
12:58 edrin you just pass your jabber account data, classpath, target_directory, xws-service-jid and functionname... and it create all classes fron batch... can be used in user defined projects then
12:59 egonw nice
12:59 edrin but you see the problem with the iof.createInputObject() ? how can iof know what root element type it should create? currently it just uses the first described in XML Schema...
13:00 egonw ah... because it can only return one...
13:00 edrin of course i could also do it like ifo.create_The_Name_For_A_Defined_Input_Type...
13:00 edrin yes
13:01 edrin that I will probably have to do...
13:02 egonw you could consider iof.createInputObject("alt​ernativeRootElementName") ?
13:02 edrin same for output... currently the factory eats/creates only "the first described root type"
13:03 edrin egonw: for example... but for the java reflection of scripting console the ifo.createAlternativeRootElementInputObject(); might be better. same for using the created factory from java project directly...
13:03 edrin probably I should change this...
13:05 edrin egonw: example http://eugeneciurana.com/past​ebin/pastebin.php?show=35972
13:05 zarah edrin's link is also http://tinyurl.com/5wns4h
13:06 egonw anyway, looks terrific!
13:06 edrin egonw: for the xws4j bioclipse plugin only two things are still problematic: 1. the generated bindings must be stored in some directory (probably best within the workspace) and 2. somehow the "load class" thing did not work yet...
13:07 edrin meaning in bc2 it did not spawn the factory. strange. in bc1 javascripts it works nicely...
13:08 egonw ah, but bc2 uses spring too...
13:08 egonw which makes classloading quite different from bc1
13:08 edrin oh?
13:08 egonw I ran into that problem last week with the CDK plugins...
13:08 edrin maybe that's the reason...
13:09 egonw talk to Gpox and masak next week...
13:09 edrin i will point you to the class loading code...
13:09 egonw they are more into this...
13:09 egonw best to chat with jonalv, but he'll be on holiday...
13:09 egonw please put that pointer in a mail to bioclipse-devel...
13:09 edrin will do so, too...
13:10 egonw jonathan may have time to answer that email...
13:13 edrin I also wanted to add a getIoFactory(IFunction function); that does not generate new binding from schema but tries to load an existing binding from target directory according to the function name. binding names are unique, containing all info "parseable" + a unique timestamp in ms in its name: example "C:\target-dir\xws.ayena.de#get​CoffeeCanStatus%1225024595765"
13:14 edrin egonw: see : http://bioclipse.svn.sourceforge.net/vie​wvc/bioclipse/xws/trunk/xws4j-binding/sr​c/net/bioclipse/xws/binding/BindingManag​er.java?revision=7431&view=markup function loadIoFactory
13:14 zarah edrin's link is also http://tinyurl.com/5q7u9x
13:22 CIA-52 bioclipse: edrin_t * r7442 /xws/trunk/xws4j/src/net/bioclipse​/xws/client/adhoc/IoSchemata.java: should return IFunction and not Function
16:03 CIA-52 bioclipse: egonw * r7444 /bioclipse2/trunk/plugins/net.biocli​pse.cdk.business/src/net/bioclipse/c​dk/business/CDKManagerHelper.java: Cleaned up code for readability
16:10 CIA-52 bioclipse: egonw * r7445 /bioclipse2/trunk/plugins/net.bioclipse.cdk.busine​ss/src/net/bioclipse/cdk/business/CDKManager.java: Made the factory static
16:10 CIA-52 bioclipse: egonw * r7446 /bioclipse2/trunk/plugins/ (3 files in 2 dirs): Renamed method, to indicate it add the BC2 supported formats, and not all (see next commit)
16:11 CIA-52 bioclipse: egonw * r7447 /bioclipse2/trunk/plugins/net.biocli​pse.cdk.business/src/net/bioclipse/c​dk/business/CDKManagerHelper.java: Helper method to register all CDK known chemical formats
16:11 CIA-52 bioclipse: egonw * r7448 /bioclipse2/trunk/plugins/net.bioclipse.cd​k.business/src/net/bioclipse/cdk/business/ (CDKManager.java ICDKManager.java): Added script command to determine the (chemical) format of a file (see http://gist.github.com/19904)
16:11 CIA-52 bioclipse: egonw * r7449 /bioclipse2/trunk/plugins/net.biocli​pse.cdk.business/src/net/bioclipse/c​dk/business/CDKManagerHelper.java: Readded PDB format, which accidentally got lost
17:55 CIA-52 joined #bioclipse
20:09 egonw joined #bioclipse
23:33 CIA-52 bioclipse: edrin_t * r7450 /xws/trunk/xws4j-binding/src/net/bioclipse​/xws/binding/tools/IoFactoryBuilder.java: now the input output factory/IoFactory can work with schemata that offer multiple root element types!
23:33 edrin joined #bioclipse
23:51 CIA-52 bioclipse: edrin_t * r7451 /xws/trunk/xws4j-binding/src/net/bioclipse​/xws/binding/tools/IoFactoryBuilder.java: bug fix (missing return);

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