OE IRC 8Feb06: Difference between revisions
Jump to navigation
Jump to search
(New page: <pre> gwg: it's just no one's gotten around to obsoleting it yet Harold: we scared jen off! gwg: again! pascale: that terms is a bit confusing pascale: again!! jen_clark has quit the serv...) |
No edit summary |
||
Line 1: | Line 1: | ||
[[Category:OBO-Edit working group]] | |||
<pre> | <pre> | ||
gwg: it's just no one's gotten around to obsoleting it yet | gwg: it's just no one's gotten around to obsoleting it yet |
Latest revision as of 18:08, 30 June 2014
gwg: it's just no one's gotten around to obsoleting it yet Harold: we scared jen off! gwg: again! pascale: that terms is a bit confusing pascale: again!! jen_clark has quit the server saying: Quit: jen_clark jen_erika has joined the channel gwg: ha, no one dares say anything now Midori has arrived to bring some discipline to proceedings! pascale: very true gwg: Jen should call herself "jen_erik" midori: oh ... thought I'd just put everyone to sleep gwg: altho that would be a very ordinary name... midori: how would we knwo it was Jen and not just anyone? Harold: groan gwg: lol! 8:50 PM J-Lo has joined the channel gwg: Ah, it's Jenny from the block! J-Lo: don't be fooled by the rocks... gwg: ... the cell block, that is! J-Lo: is Gizmo on yet? midori: which number are we supposed to use this time? Harold: use the 3348 number midori: sounds like a thunderstorm on that line ... Harold: in the absence of a mic, I'll use the "blast sounds to comment" J-Lo: someone hasn't muted! midori: I hear voices, but I don't know whose ... jen_erika: beep! tberardini: They're the ones in Amelia's head. 8:55 PM jen_erika: We don't hear people talking. should we? gwg: if they're saying "kill, kill, kill!", Tanya's right tberardini: I hear someone talking. But that could just be the voices in my head. tberardini: I definitely hear Jen. carol: I hear OBO_EDIT 15 is broken... J-Lo: yes - use beta14 jen_erika: no really it isn't. carol: I meant in a conversation on Gizmo jen_erika: i never said that tberardini: Well, whoever it was is gone noe. tberardini: now. jen_erika: that's not because I unplugged my mic midori: 15 is indeed badly broken. and all is quiet on the gizmo front. Alex_MGI has joined the channel tberardini: I can hear three of you now. Harold: this is the voice... of the lord! tberardini: Allelujah! deweybaby: Hi there, david from Cambridge here. Pretty lousy phone line huh? Harold: I hear frogs carol: I hear nothing now on Gizmo. deweybaby: So do I midori: I've had thunder twice now ... jen_erika: I hear interference 9:00 PM deweybaby: its kind like trying to tune in to police radio J-Lo: Jen - that's definitely you! midori: I hear Jen again Harold: yes Alex_MGI: I can here three people now. gwg: I hear you call my name, and it feels like home! deweybaby: yep midori: John - yes, but you're very faint deweybaby: OK muted now jen_erika: we're all muted now gwg: prize for anyone who gets that lyric reference... midori: yes deweybaby: yep Alex_MGI: yes Harold: yes J-Lo: Madonna carol: yes, with interference. J-Lo: Like a Prayer rama has joined the channel Alex_MGI: I'll try it. gwg: well done, J-Lo! midori: sure ... MelissaH has joined the channel J-Lo: what's my prize? deweybaby: haven't tried the last one yet! gwg: er... carol: I'll try it too. MelissaH: sure gwg: will get back to you on that! deweybaby: Hi Melissa. David here jen_erika: what's he uploading? Harold: oboedit -beta 16 I guess jen_erika: ooooooo midori: or maybe it's still alpha ... Harold: now cut that out midori: why, is it all greek to you? tberardini: can someone hit midori for that one? MelissaH: Hi David, glad you could make it deweybaby: left work specially for it (!) tberardini: Suparna's not here today. Stomach flu has felled her. 9:05 PM jen_erika: which david is that? deweybaby: David Sutherland Cambridge deweybaby: Perhaps silly alias was a bad idea jen_erika: where in cambridge deweybaby: FlyBAse kchris has joined the channel tberardini: Can someone signal when Gizmo is back on? deweybaby: Et tu MelissaH: hey, any other of you work on anatomy ontologies besides David and I? jrichter: Okay, you can download the new version from http://www.fruitfly.org/oboedit_webinar/oboedit_1_000-beta16-pre-release.tar.gz pascale: me kinda- although a very small ontology, dicty jen_erika: some of us work on development rama: John-Boy or Girl? jrichter: This new version fixes the link creation problem, AND should solve the problem of saving new categories in the Category Manager (and if that works, I'll now how to fix all the managers) jrichter: Rama - We aren't finding out. jen_erika: should we hear talking? Harold: new version will be able to add children? J-Lo: congratulations by the way John! MelissaH: I can't quite read the end of the link on my IRC J-Lo has left the channel tberardini: John, the URL you sent appears truncated. Is it on SF? jrichter: No. J-Lo has joined the channel rama: i was able to download from that URL pascale: it appreared truncated to me too jrichter: It's at the webinar site, but instead of index.html, get... jrichter: oboedit_1_000-beta16-pre-release.tar.gz Harold: I didn't get a url jrichter: http://www.fruitfly.org/oboedit_webinar/ jrichter: oboedit_1_000-beta16-pre-release.tar.gz jrichter: But all one big thing Alex_MGI: I get the error message oboedit.jar won't launch (mac 10.4.4 today) 9:10 PM jrichter: Launch it from a console; this is a platform independant version, and I removed the launcher jar when I fixed the Mac installation. jrichter: Also: Let's move to the 9 number on Gizmo; our current conference room seemed to have a lot of static. Alex_MGI: pardon my ignorance MelissaH: got it open deweybaby: & mine - 9? jrichter: 1-222-626-3349 pascale: can we call the 9 number now? kchris: the nine number doesn't seem any better midori: 9 number is staticky too Alex_MGI: I mean how do you launch it? jrichter: If you installed it in ~/oboedit: Alex_MGI: In the terminal? jrichter: Start a terminal window and type: jrichter: cd ~/oboedit jrichter: ./oboedit tberardini: How do I launch it from a windows machine? Harold: alex, you'l need to cd to the directory that you installed oboedit into jrichter: Tanya, it's easiest if you wait for the official release with the installers. pascale: click on the .jar file J-Lo: actually, there does seem to be a .jar file and it launched mine okay... jrichter: Yeah, but your memory settings will be ignored. Bad news. midori: ok ... preliminary report: I can add a term, but changes won't commit (reverts to <new term>) J-Lo: oh pascale: I amanged to install 16 and load the GO! jen_erika: I can't launch on windows either tberardini: So should I just use beta14, John? Alex_MGI: It appears to have launched successfully -- Thanks! jen_erika: shalll we just use 14? Harold: john, I still can't find the url; the obo-edit link on http://www.fruitlfly.org/oboedit_webinar takes me to Soreforge kchris: I seem to have 16 launched and loaded as well MelissaH: added child and it won't commit midori: changes to existing terms won't commit either kchris: Is anyone else hearing anything but static on Gizmo with the 9 number? Alex_MGI: By the way, are you also talking? midori: static on 9 here 9:15 PM tberardini: Nope, just hear static. pascale: I only hear static and beeps, no voice MelissaH: static here too deweybaby: I hung up & called again - & now I just have jungle noises! jrichter: Yeah, Midori. I got that too. jrichter: Ah well. jrichter: Let's launch beta14 and get down to business. jrichter: (Copy works in the new version, though) jrichter: Also, I'm talking on Gizmo and no one can hear. jrichter: Looks like it's text only today. jrichter: Today's lesson: The ID Manager. jrichter: Open up the ID Manager Plugin. jrichter: You'll see 4 components: A drop down menu, a configure button, an add button, and a delete button. jrichter: The dropdown menu is the profile selector. MelissaH: are we back to using 14? jrichter: (Any version will work for this) You can use a number of different id jrichter: generation profiles jrichter: If you work on different ontologies with different id generation rules, you can use a different id profile for each ontology. jen_erika: are we meant to hear stuff on gizmo? jrichter: Text only today, I'm afraid. jen_erika: cool jrichter: The configure button allows you to modify the currently selected id profile. jrichter: Click it now. jrichter: The top field is the profile name. Pretty easy. jrichter: The next field is the default id generation rule. pfey has joined the channel jrichter: This is the rule that's used if no other rule matches (I'll explain what that means in a moment) jrichter: Usually, you'll only need to setup a default rule. jrichter: When you create a new term, OBO-Edit evaluates the id generation rule to create a new id. jrichter: It uses the text of the rule verbatim, until it comes to anything enclosed in $ characters. jrichter: My rule says GO:$sequence(7,0,9999999)$ jrichter: That means when I create an ID, it will look like jrichter: GO:<some autogenerated string> 9:20 PM jrichter: There are two variables that could appear in between $ characters. rama: some autogenerated string between 7 and 99999999? jrichter: I'll explain the $sequence()$ variable in a moment. jrichter: The first (which you should NEVER use) is $parentid$ jrichter: This is replaced with the id of the new term's parent jrichter: It's always a bad idea to do this kind of thing, because you're encoding structural information in the id. jrichter: But some ontologies need it. jrichter: The other (more useful) variable is $sequence()$ jrichter: The format is $sequence(STRING_LENGTH, MIN_VAL, MAX_VAL)$ jrichter: So my sequence string will generate a 7 digit sequence between zero and 9999999. jrichter: When a new term is created, OBO-Edit will use the first sequence number that is not already in use. jrichter: This is why it's a bad idea to use the Destroy command in a published ontology; the id goes back into circulation, because it's no longer in use. jrichter: Questions so far? jen_erika: no rama: do we always have to specify sequence? deweybaby: no pascale: can you just explain the "parent" vs "sequence" again midori: no, but I'm eagerly awaiting an intro to the other box jrichter: Not necessarily, Rama. But if you don't you'll only have one id available. jrichter: For example, my rule could just be "johns_id" jrichter: But if I use that rule, I can only create one term before I'm out of ids. 9:25 PM rama: I see jrichter: The $sequence()$ section allows you to have an autogenerated numeric component to your id. jrichter: Also, the rule can have more than one $sequence()$ in it, and as long of a prefix or suffix as you want. tberardini: so if I wanted TAIR's range that would be $sequence(7,9501,11000)? pascale: so the "sequence" rule says to use consecutive IDs from the range you give it jrichter: That's right, Tanya and Pascale. jrichter: Also, you don't need to specify all the parameters in sequence. jrichter: "MYRULE$sequence(7)$---$sequence(3)$is_awesome" is perfectly fine. jrichter: That rule would generate the id "MYRULE0000000---000isawesome" the first time you used it. tberardini: For some reason, I'm not seeing all the text that you're typing John. Does anyone have the same problem? MelissaH: me too midori: how would we know? pascale: me too , just for that line Alex_MGI: I can see everything, I think. Harold: me neither MelissaH: links are cut off, other text seems to wrap jen_erika: we're fine pascale: I think when there is no spaces sometimes we dont see eveything jrichter: Maybe IRC is doing something weird because of the special characters. midori: I don't think I'm missing anything ... at least, what I do see makes sense deweybaby: perhaps doesn't like spec characters kchris: what I see makes sense too rama: yeap, what i see makes sense. jrichter: Is anyone confused, or should I move on? kchris: I'm good pascale: I am okay deweybaby: OK here MelissaH: good to go tberardini: MOve on, I'll pipe up if I'm lost. jrichter: Okay, moving on. jrichter: The box underneath is the matched rule section. Click the add button to add a new matched rule (do it now). jrichter: Then click the new rule (called "<your rule>") jrichter: Look familiar? Alex_MGI: yes jrichter: You specify a filter that matches the PARENT of your new term. midori: it's everywhere! 9:30 PM jrichter: When you create a new term, the first rule that matches the parent term is used. jrichter: If no rules match, the default rule is used. MelissaH: you lost me, where is this box? part of the ID manager? Alex_MGI: Can someone give an example of the utility of this? jrichter: Hold on Alex, I'm getting to it... jrichter: Melissa: There's a box on the bottom left of the ID Manager with "Add" and "Del" buttons beneath. jrichter: Click "Add" to create a new rule, and then select it from the list. jrichter: Alex: jrichter: Let's say you work with both the GO and the Fly Anatomy, and you are creating cross products. MelissaH: when I open the IDmanager plugin I see 4 things all in a row, the last two are a + and _. jrichter: You can create a rule that says: "Use the ID FLY:$sequence(7)$ if the parent is a fly term, but use GO:$sequence(7)$ if the parent is a GO term" jrichter: Then you can easily create new terms in each ontology. pascale: that's great! jrichter: Melissa: Click the "Configure" button to see the configuration screen I've been talking about. jrichter: (It's the first button with the gear icons) MelissaH: got it Alex_MGI: But doesn;t the cross product have parents in both ontologies. jrichter: It uses whatever parent you have selected when you choose "Create new child" from the right-click menu. Alex_MGI: I see. deweybaby: Not so clear on the practicality here. Is it possible to have both otnologies loaded at once ?? jrichter: You can load as many ontologies as you want at once. carol: How?? jrichter: If we're done here, I'll move on to load/save. deweybaby: Ahhh - must be a bit behind w/ thse features kchris: Would you want to give a cross product an ID from either of the parent ontologies? or would you want it to get some other ID? jrichter: (Actually, we've been able to load/save multiple ontologies since the early days of DAG-Edit) 9:35 PM jrichter: Karen: Depends on the ontology. jrichter: If you were making a cross product like "heart development" in GO between "development" and "heart" in some generic anatomy, it would probably go in GO. deweybaby: Oh - guess I used to do that with GO flat files... jrichter: But if it were between "development" and "mouse heart", you'd probably want the id to go in some 3rd id space like MOUSEGO. jrichter: One last thing about ids: deweybaby: So - I set the rules to create the ID specific for each ontology as I switch between them? kchris: Can you set up a rule where it looks at both types of parents to know? I guess I'm not quite following what you said about it taking the parentage from the term where you are when you say add new term jrichter: Right, d. deweybaby: OK , cool pascale: what happens to MOUSEGO if GO changes? II know this is off topic, maybe we can talk about that later) jrichter: Karen: When you add a new term, you always add it under some parent. OBO-Edit uses that initial parent to figure out the id. jrichter: Pascale - better to talk about that later... pascale: ok kchris: If it only looks at one parent to figure out the ID, how would it ever know to give it some third ID space like MOUSEGO? Alex_MGI: Can these rules be set up and last between sessions? jrichter: If you're uncomfortable basing your ID creation on the initial parent (which is just a shorthand more than anything else), you can manually switch between different ID generation profiles. jrichter: Let's create a new one now. jrichter: Close the configuration window. jrichter: Click the + button. jrichter: <new id profile> has appeared. jrichter: Click the configure button. jrichter: Karen: You'd create the new term under some existing term in the new id space. jrichter: The root of the new id space would need to be created with a new id creation profile. jen_erika: hi john. we just loaded the go ontologies and the cell type ontology as well and the go ontology disapeared and was replaced by the cell type ontology. Can you talk through at some point how we load two together? jrichter: Alex: The rules last between sessions, and can even be saved in your OBO files. jrichter: jen_erika: We'll talk about that in a second. Alex_MGI: great kchris: Do the rules last if you upgrade to the next version of OBO-Edit? jrichter: (Quick answer: You can't add a new ontology to an existing session, you need to load them together) 9:40 PM jen_erika: cool jrichter: Karen: Yes. I've moved to a more robust way of saving component configurations, so these should last between versions, unless I do something drastic to the way filters work. jrichter: Back to the demo (please hold other questions for a sec): jrichter: We've got our configuration window open for our new profile. jrichter: Give it a name, and set the rule to "SO:$sequence(3)$-there" jrichter: Click commit. jrichter: Your rule should have the name you gave it. jrichter: Create a new term and behold the beauty of id creation. jrichter: Now: questions/issues? rama: wow! midori: it worked ... but one small oddity: jrichter has a bad feeling about this MelissaH: So, this is how one would set up multiple IDsets, whereas if you were only working with one you could just use the configuration manager? midori: the ID manager showed the <default profile> and newname choices (as expected), but also had two extra copies of <new id profile> midori: it doesn't seem like a terrible problem, it's just, well, odd deweybaby: Same here jrichter: Midori - That's weird. You can get rid of the spares by selecting them and hitting the minus button. pascale: I used to use the configuration manager to do this. I suppose this way is more strict, because depending on the rule, you can generate differnt ID types in the same otology? jrichter: Those spares might be a bug, or they might have come from the bad id profile that was floating around in the OBO file for a while. 9:45 PM MelissaH: I got extra's too. jrichter: Yes, pascale. Plus, the configuration manager way doesn't work anymore. That's a vestigial control. kchris: I didn't have any spares, just the two profiles I deliberately created jen_erika: is it meant to say <your rule> in the place where it usually puts the GO:id? midori: OK, tried that -- and (this is bad) all except <default profile> disappeared. jrichter: jen_erika: If it says that, it means that the default rule wasn't set up right. Harold: but now, it appears that if you only have one specified in the configuration manager, the ID range coming in prevail; You need to have one set up as default in the id manager deweybaby: As Midori - mine all went too pascale: I was under the impression that the space IDs got created even though we didn't commit anything pascale: (srare IDs, sorry) pascale: (spare) jrichter: Harold: If OBO-Edit doesn't find any id profiles, it creates a new one based on your old Configuration Manager settings. jrichter: But then the configuration manager is ignored forever. jrichter: I don't really understand your question, Pascale. jrichter: But, time is short, so unless there are showstopper issues here, we should move on. jrichter: s'okay? kchris: Hmmm, I can't seem to delete an ID profile that I don't want. deweybaby: Except for not being able to delete one without deleting all deweybaby: Midori & I seem to habve to same ID manager phenotype jrichter: I'm having a little trouble with the delete button too. On to the list it goes! MelissaH: I could delete two of the three extra new Id profiles, but not the third midori: glad it's not just me pascale: I was just trying to see where the spare IDs came from. They were'nt there when I first opened the ID manager, so it seems as if the empty "new id_profiles" gotcreated by just opening the ID manager Alex_MGI: They were created on mine too, and disappeared when I clicked on the minus key. jrichter: Hmm... I'll look into that as well, Pascale. jrichter: Now that people will use this feature, I can start debugging it better. jrichter: Anyway - on to load/save! pascale: it's just a wild guess midori: it'll go on our test list too pascale: but I will use that feature!! jrichter: Hooray! jrichter: Moving on- jrichter: Choose Load from the File menu. 9:50 PM jrichter: If you are in advanced mode, click the "Basic..." button to move to basic mode. jrichter: Even in basic mode, you can load as many ontologies as you want. jrichter: To load several ontologies, click the browse button. jrichter: You can Ctrl+select or Shift+select several ontologies here. jrichter: Click the Open button to list them in the "Load" dialog. kchris: Do they have to be in the same directory to select them at the same time? jrichter: The paths to the ontologies appear in the "Load files" box, enclosed in quotes, and seperated by a space. jrichter: Yes. jrichter: But you can load many ontologies from different directories by typing in the locations, or using the advanced options. tberardini: Got it. jrichter: An introduction to the load/save dialog for beginners: jrichter: The dropdown box at the top allows you to select a data adapter. jrichter: Data adapters are little plugins that know how to load/save different file types. jrichter: If you wanted to open a GO flat file, you would choose "GO Flat File Adapter" from that dropdown. jrichter: The next button over is the "Advanced/Basic" button. Some adapters have advanced options; clicking this button displays the advanced settings. jrichter: The little triangle is the profile manager button. jrichter: Click it now. pascale: Suggestion: it would be nice if the default would be obo rather than GO jrichter: Pascale: agreed. MelissaH: mine comes up as obo jrichter: Clicking the triangle reveals the profile manager for this adapter. You can have several profiles for an adapter. jrichter: Profiles remember a collection of load/save settings. jrichter: This is especially useful when you're doing complicated saves to several files. jrichter: Let's look at the advanced load options. Click the advanced button. 9:55 PM jrichter: You can load as many files as you want here. jrichter: To load a file, click the "Add" button. jrichter: A new entry will appear that says "<new path>" jrichter: Choose a file to load by clicking the "Browse..." button. jrichter: You can also type in a file URL here, like http://www.geneontology.org/ontology/gene_ontology.obo jrichter: You can remove an item from the file list by selecting it and pressing "Del". deweybaby: On the obo flat file adaptor, but I don;'t see and add button jrichter: Did you click the Advanced button? deweybaby: yep jrichter: What do you see? midori: try resizing the load window tberardini: How do I delete a file that I've added on? tberardini: Ah, resizing made the del box appear. jrichter: (Tanya: Click the file you added in the list, and press the "Del" button) Harold: *hjd needs to say adios deweybaby: i remeber when resizing this window showed more buttons - but doesn;t seem to now Harold has quit the server saying: Quit: Harold MelissaH: resizing added the Del button for me too. jrichter: d: There is a little resize bar in the middle of the panel. Try pulling that to the right. MelissaH: Now when I resize, the Del button stays. jrichter: In the meantime: There's one special advanced option: "Allow dangling references" deweybaby: No luck jrichter: d: We may need to handle this offline. I suggest trying to resize the window and various components. deweybaby: fair enough jrichter: The "Allow dangling references" checkbox lets you load an ontology that contains links to terms that haven't been loaded. 10:00 PM Alex_MGI: I did find the DEL button. kchris: Not this is really something that I'd normally do, but just because it was the only other ontology file I had, I tried to load the GO OBO file and also a SO file that was almost surely in old flatfile format. It loaded the GO file just fine and ignored the SO file, but it might have been good to have some error message that the SO file was ignored. jrichter: If you load an ontology with dangling references, the references are hidden until the ontology that contains those terms is loaded. When it is, the danling references magically appear as real links. deweybaby: erms that haven't been loaded? jrichter: d: Let's say you have a file that contains cross products to another ontology, in another file. deweybaby: OK jrichter: Later on, you load your first ontology, but you don't load the second ontology. jrichter: You need to enable dangling references. Alex_MGI: Aren't all the ontologies loaded at the same time (or at least before the user can do anything)? deweybaby: OK - think I'm with you jrichter: Yes, Alex. But you can load an ontology with dangling references, save it out, and load it later with an ontology that resolves those references. Alex_MGI: I understand now. jrichter: Lets load some ontology, and then we'll look at saving. jrichter: Go ahead and load something, and then choose "Save As" from the file menu. jrichter: "Save As" has basic and advanced modes too. jrichter: In Basic mode, it saves everything you have loaded into a single file. jrichter: That's it. jrichter: In advanced mode, things are a lot messier... jrichter: Click the "Advanced" button now. MelissaH: I don't have advanced mode in version 14 rama: neither do i jrichter: Make sure you have "OBO Flat File Adapter" selected as your data adapter. MelissaH: aha! rama: ho, thanks. 10:05 PM pascale: is that how you save the cross-product ontology? deweybaby: Weird - seems to have fixed itself - am now getting all buttons for both load and save jrichter: Yes, Pascale. pascale: jrichter: In advanced mode, you'll see a file list (like in the advanced load screen) and a drop down at the bottom. jrichter: The drop down at the bottom lets you decide which OBO format to use. jrichter: OBO 1.2 supports more features (like synonym categories) but OBO 1.0 is more compatible with old tools. jrichter: Each item in the file list represents a file you should write out. jrichter: Add a file to that list now by pressing "Add" rama: are there only two output_types? jrichter: Rama: For now. jrichter: At the top is a box for the save path. This is the location where the file will be written. pascale: is this documented anywhere? I didnt realize there were two OBO formats jrichter: Pascale: The OBO site has a revolting amount of documentation on this. Right now, GO uses version 1.0, so the GO consortium itself has not yet been convulsed by this new format. jrichter: Anyway: jrichter: The next line under the save path is the filtering controls. jrichter: We'll come back to that. jrichter: Below the filtering controls is the "Save implied links" section. jrichter: It is grayed out because the reasoner is turned off. jrichter: This section is only available if the reasoner is on. jrichter: If the reasoner is on, this option allows you to save reasoner derived links in the file. rama: what happens if I select obo_1_0 and commit the file? jrichter: Rama: It will work as per usual. rama: cool jrichter: If you select obo_1_2 and commit, there will be some problems with the synonym lines and you'll get a big diff. 10:10 PM jrichter: (The reasoner derived links will be marked with a special flag in the OBO file if you save implied links. jrichter: Next box down is the id rules box. jrichter: You can write out a special header that contains id rules in OBO 1.2. deweybaby: "The reasoner derived links will be marked with a special flag" - good could be messy otherwise jrichter: It can only write out certain simple rules based on namespaces. jrichter: This is useful if your ontology has a special id generation rule, AND lots of people edit it, AND you want to make sure that everyone is using the same rules. deweybaby: Ahh - so sticking with 1_0 will be a problem for saving the ID rules we created earlier? jrichter: No. Your id rules are saved in the OBO-Edit config. deweybaby: Oh yeh - course... jrichter: You just won't be able to communicate those rules to other users using OBO 1.0. jrichter: Last simple thing: deweybaby: with you jrichter: The "Allow dangling parents box" jrichter: This box is used in conjunction with filtering. jrichter: You may create a filter where you write out a term that has links to terms that will not appear in the file you write. jrichter: (ie Term A matches your filter, but one of its parents, Term B, doesn't match) jrichter: If "Allow dangling parents" is turned on, a dangling parent link will be written to preserve the information that B is a parent of A, even though B isn't in the file. 10:15 PM jrichter: If "Allow dangling parents" is turned off, the link won't be written at all. jrichter: Are we cool so far, before I go into filtering? deweybaby: yep midori: yup MelissaH: gotcha jrichter: Click the "Filter terms" checkbox. J-Lo: i think so... jen_erika: can you look at loading two ontologies some time? It didn't work for us jrichter: jen_erika: We'll go through an example of that at the end. jrichter: (end of the day, once others have left) jen_erika: ta jen_erika: that's a bit late for us jrichter: jen_erika: I can email you instructions then. jen_erika: another time is fine jrichter: Once you click the filter terms button, you should see the familiar filter interface. jrichter: When you set up a term filter here, only terms that match your filter will appear in the file. J-Lo has her hand up deweybaby: How would I use this to filter for name space - say a slime? jrichter acknowledges J-Lo kchris: What are the 3 icons to the right of the "Allow dangling parents" option? Though most of my tool tips show up, the tips for those three never show up for me. deweybaby: Dewey would put his hand up up if knew how! J-Lo: Does the filter save as greedy/strict - how does it know? jrichter: d: [Self] [Namespace] [equals] "slime" - no dangling parents jrichter: J-Lo - we'll get to that when we talk about "Filter links", but it's always Greedy. J-Lo: okay jrichter: Recall what I said about dangling parents... jrichter: That goes out the window when we filter links. jrichter: Click the "Filter links" button now. deweybaby: The list contains the choice 'name' - don't see namespace jrichter: d: Scroll down, it's under Definition. deweybaby: sory 10:20 PM deweybaby: sorry - godit jrichter: When you click "Filter links", you may need to make your window bigger, because the interface is pretty crowded. jen_erika: what's meant to be ticked in advanced options? jrichter: When "Filter links" and "Filter terms" are up together, only terms that match the term filter are written, and only links that match the link filter are written. jrichter: Right now we have "filter terms" and "filter links" clicked. jrichter: With both selected, you have the greatest possible control over what appears in your file. jrichter: BTW: If only "Filter links" is selected, all terms appear in the file, but only matching links appear. jrichter: Back to Karen's question: jrichter: The three little buttons allow you to save a reset/load/save a filter to a file on disk. jrichter: This can be very useful! jrichter: You can "beta-test" your save filter by using the normal filter controls, save the filter, and then load it up here when you save. deweybaby: List of choices in link filtering looks like it contains lots of things specific to term .../? deweybaby: terms jrichter: Remember from yesterday... deweybaby: OK jrichter: Link filters filter on either the link parent, link child, link type, OR attributes of the link itself. jrichter: The parent, child, and type aspects act just like term filters, but if you select "Self" from the first dropdown, you see link-specific stuff. jrichter: Okay. Questions? 10:25 PM jen_erika: did you say that if we put criteria in both filters then it uses both? midori: ok here deweybaby: Get the general pic - gonna have to play with it to be clear... jrichter: Yup, jen_erika. MelissaH: need to play too jrichter: Remember - the easiest way to use this is to use the OBO-Edit filtering system to test your filter first. jrichter: Make sure you can see every term and link you want to see. jrichter: Then, save your filter and re-use it on the save screen. deweybaby: cool jrichter: I'm ready to move on to the reasoner if we're okay. deweybaby: yep carol: yes midori: I'm gonna have to bow out shortly ... otherwise, fine J-Lo: okay jrichter: To use the reasoner, open up the Reasoner Plugin. MelissaH: ok jrichter: (I'll be done in < 15 mins) jrichter: Make the window a little bigger than it starts initially. jrichter: Then, click the "Use reasoner" checkbox. jrichter: Sit back and relax, this can take a while. jrichter: (Mine takes 13 seconds) midori: mine's done! jrichter: When it finishes, you'll see a report like this: midori: 24732 ms jrichter: Reasoning completed in 13207ms jrichter: * Discovered 211264 implied links jrichter: * Showing 0 implied links jrichter: * Found 14 redundant links jrichter: Or something similar. jrichter: Did anyone have zero redundant links? kchris: yes rama: still going for me midori: I had 14 tberardini: Mine says: Finished reasoning. O new links created by reasoner. MelissaH: I had too many ontologies loaded, had to quit 10:30 PM jrichter: Karen, what ontology are you using? jrichter: Tanya: Make your window bigger to see the report. tberardini: Window is big and application has crashed. Restarting jrichter: Any non-experimental ontology is going to create zero new links. kchris: the official version of GO-OBO that I was using yesterday. I'll check the version # jrichter: If you're crashing, you may need to up your memory to to this. MelissaH: 8 redundant links jrichter: Karen: I find it hard to believe that GO now contains 0 redundant links... jrichter: If you found 1 or more redundant links, lets see where they are... gwg: Oh ye of little faith John! Alex_MGI: I found 0 redundant links with both GO and cell ontologies loaded. J-Lo: is there any way to get it to list the redundant links? jrichter: Jane: Yup. Set up your filtering system to do a link filter. jrichter: Search for [Self] [Self] [Is redundant] kchris: I'm using revision number 3.1358, which appears to be two revision numbers out of date from the most current J-Lo: great! rama: mine is still going. wonder if i should start again. jrichter: Rama: You probably should up your memory, restart OBO-Edit and try again. kchris: I also see that the GO-OBO file header doesn't seem to contain its revision number any more jrichter: Karen: I'm loading the version of GO from the website now. MelissaH: I found 8 in revision 3.1355 Alex_MGI: How do you up the memory on OBO-Edit? gwg removed the topic deweybaby: Mine spazzed. Fine with a nice small FB qualifier ontology gwg removed the topic tberardini has quit the server saying: Ping timeout midori: I've got 3.1358 here too ... 14 redundant links (GO and Cell loaded); GO file has CVS revision line in header ... 10:35 PM jrichter: Increase your memory by going to Configuration Manager Plugin. MelissaH: I filtered and now I only see the roots jrichter: Type in a new memory value, hit "Save Configuration" and restart OBO-Edit. jrichter: Melissa: Turn off the filter. midori: some of the redundant relationships are in CL, others in GO Alex_MGI: The vestigial one? kchris: about revision number in header, I see it now, it's just too far down to get it with the unix command 'head' now jrichter: Set up the link filter, but hit "Search" not save. midori: yep, it's line 12, and head defaults to 10 lines tberardini has joined the channel Alex_MGI: I'm working with an older version of GO, by a few weeks. jrichter: If you got some redundant links, use the link search I described to Jane to find them. tberardini: Very, very bad crashing on this end. Killed Firefox, killed OBO edit, killed IE. jrichter: Windows trouble. tberardini: Aye. jrichter: Make sure you've got enough memory allocated to OBO-Edit. Reasoning requires a lot of it. jrichter: If you've done the search for redundant links, select one of the matching links in the search results panel. tberardini: Hmmm. This time it says completed in 7610 ms and there are 7 redundant links. Sorry if I'm going over old ground. I can't see the conversation I missed when I got kicked off. jrichter: Good Tanya, that's what you want. rama: i have 10 redundant links. I have version 3.1358. Only GO loaded jrichter: Karen: you're right! No redundant links in the new GO. 10:40 PM kchris: When I set up the link filter to search for redundant links in GO-OBO 3.1358, with no other ontologies loaded, I get 0 results, but if I hit the filter icon instead of the search icon, I get lots of things, way more than 14... I'm not sure I understand what the filter is doing midori: Amelia did say she'd committed some fixes today ... tberardini: sorry again if you've repeated this alot but how to I look at the redundant links. jrichter: I'm going to put a version of the GO online that HAS redundant links so we can look at this. jrichter: In the meantime, everyone should take a moment to give OBO-Edit more memory and restart it. midori: just get the cell type ontology -- it's got a few kchris: how do we give it more memory midori: http://cvs.sourceforge.net/viewcvs.py/*checkout*/obo/obo/ontology/anatomy/cell_type/cell.obo?rev=1.16 pascale: and how much? MelissaH: I can't find the memory allocation pascale: it's in the configuration manager MelissaH: there is only one DAG-edit configuration manager, right? rama: yes MelissaH: I see 5 tabs... jrichter: It's in the first tab, near the bottom. jrichter: It says 256M initially. MelissaH: needed a resize..... jrichter: You can download the new file from http://www.fruitfly.org/oboedit_webinar/go_redundant.obo 10:45 PM pascale: how much memory would you say is enough? kchris: apologies, I've got to go to another appointment, I'll read the transcript of the chat to see how this works jrichter: (midori - I lied about how long this would take - you can sleep if you want) midori: Well, I have to make a move now. John, thanks you very much! Everyone have fun ... jrichter: Pascale - as much as you can handle. jrichter: By Karen. deweybaby: So - 78 redundant links in Fb anatomy (not suprising) jrichter: bye