Kepler: Issueshttps://projects.ecoinformatics.org/ecoinfo/https://projects.ecoinformatics.org/ecoinfo/ecoinfo/favicon.ico?14691340362010-03-09T23:53:44ZEcoinformatics Redmine
Redmine Bug #4873 (Resolved): Actors in repository library subtree are crammed togetherhttps://projects.ecoinformatics.org/ecoinfo/issues/48732010-03-09T23:53:44ZChad Berkleyberkley@nceas.ucsb.edu
<p>On Windows XP (maybe on other versions of windows) the actor icons are crammed tightly together when you open up Actors-2_0/CoreActors,kar or the other yellow repositories. They should be spaced similarly to the mac.</p> Bug #4629 (Resolved): Undo doesn't roll LSID revisionhttps://projects.ecoinformatics.org/ecoinfo/issues/46292009-12-14T21:48:26ZAaron Aaronaschultz@nceas.ucsb.edu
<p>When undoing changes in a workflow the LSID of the workflow does not get incremented.</p> Bug #4581 (Resolved): Handle multiple revisions of components in the Component Libraryhttps://projects.ecoinformatics.org/ecoinfo/issues/45812009-11-25T05:23:19ZAaron Aaronaschultz@nceas.ucsb.edu
<p>It is possible for many different revisions of an object to contain the same semantic types. The question is how to handle this in the Component Library.</p>
<p>Imagine you have two different KARS<br />one with a component, urn:lsid:someauth:namespace:24:10<br />and the other with a component, urn:lsid:someauth:namespace:24:11<br />both of which have equivalent semantic types.</p>
<p>These two components should show up properly in the Ontology tree in the Component Library but since they would be in the same place there is a collision. For this reason, LSID cannot be used as the identifier in the Component Library. The new Library Index has a unique Library Index ID (LIID) for each item that shows up in the Component Library tree. With this LIID we can associated multiple LSIDs with any given LIID (see LIBRARY_LSIDS table in the sql).</p>
<p>This problem is not limited to LSID revisions. But also objects that have the same name and the same semantic type.</p> Bug #4477 (Resolved): Grey out "undraggables"https://projects.ecoinformatics.org/ecoinfo/issues/44772009-10-16T19:17:34Zben leinfelderleinfelder@nceas.ucsb.edu
<p>In the Components tree we now have non-workflow items (items contained in KARs). These should be marked as different so that it's clear that we are mixing metaphors.</p> Bug #4475 (Resolved): Order context menu items consistentlyhttps://projects.ecoinformatics.org/ecoinfo/issues/44752009-10-16T19:15:36Zben leinfelderleinfelder@nceas.ucsb.edu
<p>There is a mix of options, sure, but options that appear across different context menus (ie "Open") should appear in similar locations on the list.</p> Bug #4380 (Resolved): Removing LocalRepository does not remove actors from Ontologieshttps://projects.ecoinformatics.org/ecoinfo/issues/43802009-09-11T04:28:03ZAaron Aaronaschultz@nceas.ucsb.edu
<p>scenario:<br />place a kar containing a component that has a semantic type in a folder<br />add that folder as a local repository<br />see that the component has shown up in the appropriate ontology<br />now remove that local repository<br />see that the folder is no longer in the library<br />but the component still shows up in the ontology</p> Bug #4174 (Resolved): Actor library not building correctlyhttps://projects.ecoinformatics.org/ecoinfo/issues/41742009-06-19T04:19:15ZAaron Aaronaschultz@nceas.ucsb.edu
<p>Many of the actors are not appearing in the actor library when using the WRP suite. No visible errors are given...</p> Bug #4008 (Resolved): EML 2 Dataset throws errors when accessing public data from mixed public/pr...https://projects.ecoinformatics.org/ecoinfo/issues/40082009-04-21T21:21:41ZOliver Soongsoong@nceas.ucsb.edu
<p>When a dataset on KNB has some tables that are public and some tables that are private, EML 2 Dataset attempts to load and cache both the public and the private tables, even though only public tables are requested in the workflow. Because the actor cannot access the private data, it throws errors.</p> Bug #3898 (Resolved): update .kepler instead of removing it between versionshttps://projects.ecoinformatics.org/ecoinfo/issues/38982009-03-17T16:59:32ZChristopher Brookscxh@eecs.berkeley.edu
<p>Thanks to Qiao Huijie for pointing this out.<br />When switching between Kepler-1.0.0 and the Subversion development tree,<br />the ~/.kepler directory must be removed.<br />If it is not removed, then no actors are present in the actor browser pane.</p>
<p>I'm not sure why there is an incompatibility, but the code should detect<br />this incompatibility and prompt the user.</p>
<p>In addition, upon startup, the code should check to see if the actor browser<br />pane is empty and, if it is, prompt the user to delete their ~/.kepler <br />directory</p>
<p>Also, the developer instructions should be updated to mention removing<br />the ~/.kepler directory. I'll update the instructions.</p> Bug #3277 (Resolved): Create reusable jython-based SelectionDialog actor to collect runtime param...https://projects.ecoinformatics.org/ecoinfo/issues/32772008-05-06T21:42:44Zben leinfelderleinfelder@nceas.ucsb.edu
<p>Dan (Higgins) did a prototype of this use of the PythonScript actor to create Swing interfaces for displaying drop down lists and then sending the selected value out on a port.<br />This will be a useful feature when workflows need to provide data-driven parameters (select date ranges or sites from a list of possible values).<br />The prototype SelectionDialog actor I have drafted so far includes the following ports:<br />"choices" (list)<br />"prompt" (message for prompt as a string)<br />output (the value selected)</p> Bug #3186 (Resolved): commit RExpression code changes, but leave actor openhttps://projects.ecoinformatics.org/ecoinfo/issues/31862008-03-26T00:18:05ZKevin Drurydrury@nceas.ucsb.edu
<p>R-script development within an RExpression actor can get pretty tedious. For example, when you add a line of code and want to test its effects, you push the RExpression actor "commit" button and this closes the RExpression actor. You can then run the workflow. Then, to make the next small change, you re-open the RExpression actor, alter the code, push "commit", etc. I have repeated this procedure 741 times in a single day (yes, I counted :). It would be REALLY nice if, when the text within the R-code window of the RExpression actor changed, the "submit" button would become activated (i.e., not grayed out). Then, when an analyist has added the next bit of code, they can just push submit and "run" the workflow. In other words, "submit" does not close the RExpression actor, but does make R-code changes visible to the workflow. Is this possible with Kepler?</p> Bug #3185 (Resolved): Ctrl-z does not "undo" in RExpression actorhttps://projects.ecoinformatics.org/ecoinfo/issues/31852008-03-26T00:06:14ZKevin Drurydrury@nceas.ucsb.edu
<p>Ctrl-c (copy) and Ctrl-v (paste) work as expected in the RExpression actor. Is it possible to also make Ctrl-z (undo) work within the R-code window?</p> Bug #3154 (Resolved): Documentation missing for DarwinCoreDataSourcehttps://projects.ecoinformatics.org/ecoinfo/issues/31542008-02-21T19:47:38ZDan Higginshiggins@nceas.ucsb.edu
<p>When a DarwinCoreDataSource is dragged to the work area from the actor list, there is no documentation available. The same problem occurs when the actor is the result of a search.</p> Bug #2964 (New): Change RExpression plotting actors to allow custom axis and title labelshttps://projects.ecoinformatics.org/ecoinfo/issues/29642007-09-12T00:53:57Zben leinfelderleinfelder@nceas.ucsb.edu
<p>If/When metadata is being passed around with workflow tokens, use the information contained within that to set the labels for chart axis, title etc...</p> Bug #2960 (Resolved): Change RExpression actor to use meaningful graphic file output nameshttps://projects.ecoinformatics.org/ecoinfo/issues/29602007-09-12T00:29:16Zben leinfelderleinfelder@nceas.ucsb.edu
<p>Possibilities include:<br />-nested workflow.composite_actor.actor.incremental_number naming<br />-use a similar approach to above, but somehow track subsequent runs of the workflow (like versioning 1.1, 1.2, 1.3 would indicate three different runs of the same workflow/actor)<br />-have user set prefix for output object (either as parameter or port) in the actor<br />-create new output directory (likely timestamp) for each workflow execution to group output files.</p>
<p>probably end up using a combination of the choices here.<br />directory + actor_name being a likely candidate</p>