From: Nomi H. <no...@us...> - 2004-11-18 00:38:37
|
Update of /cvsroot/gmod/apollo/doc In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv10381 Modified Files: todo fixed-bugs Log Message: Index: todo =================================================================== RCS file: /cvsroot/gmod/apollo/doc/todo,v retrieving revision 1.658 retrieving revision 1.659 diff -C2 -d -r1.658 -r1.659 *** todo 17 Nov 2004 19:07:28 -0000 1.658 --- todo 18 Nov 2004 00:38:27 -0000 1.659 *************** *** 5,70 **** ---------------------------------------------------- ! changing result tag to incomplete cds has a repaint issue ! ! Add GAME version number in GAME output (but where to get version #?). ! ! on changing the range of an exon in Hph (annot-types.xml) in the EDE ! it seemed to erroneously split the gene and get all wack - need to ! look into this more. - happens on making intron on highest exon (3' ! forward, 3' reverse) ! ! sequencing error should cause a recalc of translation start - doesnt ! seem to - where translation stop does get recalculated. ! ! Add ability to add publication to synonyms. need in gui and game. ! ! When a comment is changed (via the comment editor), change the author and ! datestamp. ! ! Release 1.4.5 doesn't make--it is missing src/java/apollo/dataadapter/ensj19 ! (which I added) but it still is having trouble compiling ! src/java/apollo/dataadapter/ensj19/AnnotationEnsJAdapterGUI.java ! src/java/apollo/dataadapter/ensj19/AnnotationEnsJAdapter.java ! Make a new release 1.4.6 (might as well) tar for sourceforge. ! These are my thoughts on redoing the gui for comments (yes i know we did a ! which really helped but I think we need to go one step further) ! In main annot info window have comment list - of comment text not author/date ! or maybe both but minimally comment text (maybe just 1st 30 chars?) - anyways ! there would be an add,delete and edit buttons. Edit and delete buttons would be ! greyed out unless a comment was selected. If add comment was pressed the ! commentator box would come up (minus the comment list, add button, and delete ! buttons). The comment would then be constructed via the dropdown list and editing. ! and close would be hit - I suppose there could also be a cancel button to cancel ! the add. If a comment was selected in the comment list (in the main window NOT ! the commentator) edit and delete would be enabled (ungreyed). Delete would just ! delete it. Edit would popup the commentator as with add. With comments in this ! list in main window comments could be taken out of "comments" window. Comments ! window could turn into a "properties" window and show all the other properties ! as it does now. ! unless/until the annot info gui is changed have to make sure empty ! comments get wiped out on commentator window clsing - only get wiped ! on close button at the moment. - no biggie ! ! minor bug: after changing type from gene to something non-coding like ! snoRNA the start and stop are still drawn in the main window. ! Don't let users create new transcript by dragging a collapsed tier. This ! should be popping up a warning message, but Mark isn't getting one--why? ! serialized files seem to fail to read in and only get vague error ! popup message: "Exception from java.lang.NullPointerException". Bug ! should be fixed and error message should be made more informative ! (a stack trace would be good). ! get the installed version to tell you how much ! memory it's allocating (could grep from Apollo.lax) ! synonyms, dbxrefs, and comments are being copied on just moving one ! exon (not a transcript merge - and its going from target transcript to ! source transcript). On transcript merge the new transcript does not ! have the deleted transcripts stuff - i suspect that this is because ! the merged stuff went to the deleted and not vice versa. Cant delete synonyms in info editor. can only add. if you delete one --- 5,42 ---- ---------------------------------------------------- ! New bugs introduced since r1.4.6: ! -------------------------------- ! Load tiny.xml; try to save. ! Warning: couldn't find file conf/null anywhere ! GAMEAdapter.saveTransactions(): java.lang.IllegalStateException: ChadoTransactionXML.commitTransactions(): Cannot find xml template for chado transaction. ! java.lang.IllegalStateException: ChadoTransactionXML.commitTransactions(): Cannot find xml template for chado transaction. ! at apollo.dataadapter.chadoxml.ChadoTransactionXMLWriter.commitTransactions(ChadoTransactionXMLWriter.java:73) ! at apollo.dataadapter.TransactionOutputAdapter.commit(TransactionOutputAdapter.java:87) ! at apollo.dataadapter.gamexml.GAMEAdapter.saveTransactions(GAMEAdapter.java:334) ! at apollo.dataadapter.gamexml.GAMEAdapter.commitChanges(GAMEAdapter.java:298) ! Reading from backup file broken (null pointer--no informative message). ! If you load the example GFF file and then load a GAME file, none of the ! GAME results show up (only the annotations). (If you load the GAME file ! first, or load it after first loading a GenBank file, it works fine--it ! only seems to be a problem after loading a GFF file.) ! Also, if you load a GAME file (e.g. annot-types.xml) and then load the ! sample GFF file, it comes up with all the results tiny and gray. If you ! load the GFF file first, it's fine. ! In both cases, it seems that the tiers are coming up collapsed to the ! point of (near-)invisibility; if you say "Expand all tiers," they show up. ! Annot info editor: Undo doesn't work for symbols, ID, synonyms, or ! changes to comments. [Mark says these issues are already on his list ! (doc/chado/transaction-todo)] ! -------------------------------- ! (End of new bugs introduced since r1.4.6) ! Add GAME version number in GAME output (but where to get version #?). ! Add ability to add publications (database literature IDs--e.g. FBrf or ! PubmedID) to synonyms. need in gui and game. Cant delete synonyms in info editor. can only add. if you delete one *************** *** 72,75 **** --- 44,50 ---- changing selection. + get the installed version to tell you how much + memory it's allocating (could grep from Apollo.lax) + <arm> tag is not always written out (it should be)--this is because Apollo expects the arm to have a name different from the seqid, and the *************** *** 80,84 **** using the old method and is writing the seq.getDisplayId ! Can't click "select" boxes in restriction enzyme table. produces_seq field being lost from <feature_set> records after roundtrip --- 55,61 ---- using the old method and is writing the seq.getDisplayId ! Can't click "select" boxes in restriction enzyme table. (Broke between ! r1.4.4 and r1.4.5.) ! http://cvs.sourceforge.net/viewcvs.py/gmod/apollo/src/java/apollo/gui/tweeker/RestrictionEnzymeSelector.java?r1=1.14&r2=1.15 produces_seq field being lost from <feature_set> records after roundtrip *************** *** 91,98 **** I have a feeling this was a problem that was fixed before that has now returned. - Peptide spacing can be buggy (most of the time gets it right). - RE54776 DCG cdna dark green feature in AE003689 result seq has peptide - spacing in its zoomed in main panel feat though it's nucleotide seq. - i just noticed a rather obscure and not-very-important bug that possibly hints at other as-yet-undiscovered issues. i tried reading from --- 68,71 ---- *************** *** 107,131 **** properly in the backup (although i didn't see any other bad symptoms). ! Sima reported problems editing gene symbols and IDs in annot info ! editor. Can't replicate. Sima wrote: ! I'm in internal mode, and looking at genes in the "gadflyst" database. ! For example, if I query for "gene" "HP1679" from gadflyst, and then select ! "ST:X:15590323:15589211:GENSCAN" and I open up the Annotation Info editor, ! and then I select some text at the end of the "Symbol", delete it, then go ! to the middle, and try to delete the X, the cursor goes the end of the ! string. Then if I go the beginning of the string and try to type ! "CG:temp", the cursor goes to the end of the string after every 1 or 2 ! strokes I type (e.g. if I type "te" really fast it lets me get those 2 in ! before the cursor ends up at the end of the string). ! [Sima sees this on Linux and on Mac.] ! ! Inappropriately asking if you want to split genes you weren't working on ! (see Kathryn Campbell's message of Tue, 16 Mar 2004 18:20:27). ! to see this behavior, select gene CG12147 in the old (jan 30) example (i ! have it saved as ~nomi/apollo/data/example.old.xml), click somewhere in ! the middle of it, and say "split exon". now select the first exon and ! say "delete selection". you are asked, "Are you sure you want gene ! CG12163 to be split"? CG12163 is clear on the other side of the rev ! strand. Show axis in view menu needs a repaint. --- 80,85 ---- properly in the backup (although i didn't see any other bad symptoms). ! Don't let users create new transcript by dragging a collapsed tier. This ! should pop up a warning message (it seems to do so only sporadically). Show axis in view menu needs a repaint. *************** *** 133,145 **** Annotation labels don't slide along as you scroll if you're revcomped. - Put translational exceptions (selenocysteine for stop, +/-1 frameshift, - one base pair genomic sequencing errors) into the transaction records. - [They are already there, but just as FeatureChangeEvent.REPLACE or LIMIT - events on TRANSLATION--is that not sufficient for now?] - - After much fiddling with editing I managed to get in a state where - sequencing errors i was doing in EDE were not coming up in annot - info. Sporatic - dont have repeatable. - URLs for "Show gene report" need to be updated. internal: ProjectRefURL "http://www.fruitfly.org/cgi-bin/annot/mini-search.pl?query=" --- 87,90 ---- *************** *** 221,227 **** worse; it clears the main window, leaving it blank!] Split up gui from non-gui config - so bop can run without apollo gui classes. - Minor bug: Cant move feature to other strand if there exists no feature of that type on other strand - will throw exception - no model being made for --- 166,173 ---- worse; it clears the main window, leaving it blank!] + changing result tag to incomplete cds has a repaint issue + Split up gui from non-gui config - so bop can run without apollo gui classes. Cant move feature to other strand if there exists no feature of that type on other strand - will throw exception - no model being made for *************** *** 265,268 **** --- 211,226 ---- - If user changes annotation ID, make it show up in a different color in the annot info window. + - Sima reported problems editing gene symbols and IDs in annot info + editor. Can't replicate. Sima wrote: + I'm in internal mode, and looking at genes in the "gadflyst" database. + For example, if I query for "gene" "HP1679" from gadflyst, and then select + "ST:X:15590323:15589211:GENSCAN" and I open up the Annotation Info editor, + and then I select some text at the end of the "Symbol", delete it, then go + to the middle, and try to delete the X, the cursor goes the end of the + string. Then if I go the beginning of the string and try to type + "CG:temp", the cursor goes to the end of the string after every 1 or 2 + strokes I type (e.g. if I type "te" really fast it lets me get those 2 in + before the cursor ends up at the end of the string). + [Sima sees this on Linux and on Mac.] InstallAnywhere issues: *************** *** 418,421 **** --- 376,383 ---- JV118) If just viewing protein alignments, genomic sequence gets protein seq - should be genomic. + JV119) If we have only cigars in the input, and not alignments, the + genomic sequence appears only as nucleotides, even when aligned with a + peptide--it should be using the cigar to fill in the genomic translation, + but apparently it's not. -------------- *************** *** 461,464 **** --- 423,428 ---- - When type changes, right detail panel doesn't change (though left one does!) + - After changing type from gene to something non-coding like + snoRNA the start and stop are still drawn in the main window. - What options should be turned off for non-gene annots--e.g. Exon Detail Editor, calculate longest ORF, ...? *************** *** 733,736 **** --- 697,716 ---- at apollo.gui.AnnotationEditor.addGeneOrTranscript(AnnotationEditor.java:1701) at apollo.gui.AnnotationEditor.addGeneOrTranscript(AnnotationEditor.java:1687) + AE51) Inappropriately asking if you want to split genes you weren't working on + (see Kathryn Campbell's message of Tue, 16 Mar 2004 18:20:27). + to see this behavior, select gene CG12147 in the old (jan 30) example (i + have it saved as ~nomi/apollo/data/example.old.xml), click somewhere in + the middle of it, and say "split exon". now select the first exon and + say "delete selection". you are asked, "Are you sure you want gene + CG12163 to be split"? CG12163 is clear on the other side of the rev + strand. + AE52) sequencing error should cause a recalc of translation start - doesnt + seem to - where translation stop does get recalculated. + AE53) synonyms, dbxrefs, and comments are being copied on just moving one + exon (not a transcript merge - and its going from target transcript to + source transcript). On transcript merge the new transcript does not + have the deleted transcripts stuff - i suspect that this is because + the merged stuff went to the deleted and not vice versa. + --------------------- *************** *** 750,757 **** --- 730,761 ---- annotation info window even appears? - Should add variable (GOURL?) to style instead of just hardcoding it in GeneEditPanel. + + Annotation comments: + - When a comment is changed (via the comment editor), change the author and + datestamp. + - unless/until the annot info gui is changed have to make sure empty + comments get wiped out on commentator window clsing - only get wiped + on close button at the moment. - no biggie - It's weird to have the list of comments be by author name and date. Should have the comment text in that list. Panel to right is really just a panel for adding/editing comments, which should be a popup that comes up when you want to add or edit rather than living there full-time. + - These are my [Mark's] thoughts on redoing the gui for comments (yes i + know we did a which really helped but I think we need to go one step + further) In main annot info window have comment list - of comment text + not author/date or maybe both but minimally comment text (maybe just 1st + 30 chars?) - anyways there would be an add,delete and edit buttons. Edit + and delete buttons would be greyed out unless a comment was selected. If + add comment was pressed the commentator box would come up (minus the + comment list, add button, and delete buttons). The comment would then be + constructed via the dropdown list and editing. and close would be hit - + I suppose there could also be a cancel button to cancel the add. If a + comment was selected in the comment list (in the main window NOT the + commentator) edit and delete would be enabled (ungreyed). Delete would + just delete it. Edit would popup the commentator as with add. With + comments in this list in main window comments could be taken out of + "comments" window. Comments window could turn into a "properties" window + and show all the other properties as it does now. + ----------------- *************** *** 786,789 **** --- 790,796 ---- like C's Maybe choose a different font? Or maybe we shouldn't show the bases until the user zooms in a teensy bit more? (See ScaleView:drawSequence) + - Peptide spacing can be buggy (most of the time gets it right). + RE54776 DCG cdna dark green feature in AE003689 result seq has peptide + spacing in its zoomed in main panel feat though it's nucleotide seq. Annotation creator colors and names: *************** *** 862,865 **** --- 869,876 ---- at java.awt.EventDispatchThread.run(EventDispatchThread.java:80) + Right-menu option sometimes says "Set owner to FlyBase (taking from + FlyBase)". It doesn't make sense to offer a set-owner change when the + new value is the same as the old. (Very minor issue, though.) + ---- Exon Detail Editor (BOFE): *************** *** 986,990 **** change, not for just any gene. But if I remove the deletion, I still can't see the black rectangle when I don't show introns for the sdk gene. ! EDE81) creating exons (in introns) seems to repeatably produce this message: DEBUG: Range.isContainedByRefSeq() feature has null reference seq Returning true even though we dont really know --- 997,1001 ---- change, not for just any gene. But if I remove the deletion, I still can't see the black rectangle when I don't show introns for the sdk gene. ! EDE82) creating exons (in introns) seems to repeatably produce this message: DEBUG: Range.isContainedByRefSeq() feature has null reference seq Returning true even though we dont really know *************** *** 998,1001 **** --- 1009,1024 ---- at javax.swing.AbstractButton.fireActionPerformed(AbstractButton.java:1786) at javax.swing.AbstractButton$ForwardActionEvents.actionPerformed(AbstractButton.java:1 + EDE83) on changing the range of an exon in Hph (annot-types.xml) in the EDE + it seemed to erroneously split the gene and get all wack - need to + look into this more. - happens on making intron on highest exon (3' + forward, 3' reverse) + EDE84) After much fiddling with editing I managed to get in a state where + sequencing errors i was doing in EDE were not coming up in annot + info. Sporatic - dont have repeatable. + EDE85) it seems like the "adjust for sequencing error" functionality in + the EDE does not work right. (this is not a new problem--it seems to be + the same in r1.4.4.) if i try to add an insertion, it asks me which base + i want to add, but then it frequently adds a different base--i asked for + an N and got a T; asked for a G and got an A. -------------------- Index: fixed-bugs =================================================================== RCS file: /cvsroot/gmod/apollo/doc/fixed-bugs,v retrieving revision 1.388 retrieving revision 1.389 diff -C2 -d -r1.388 -r1.389 *** fixed-bugs 11 Nov 2004 00:43:49 -0000 1.388 --- fixed-bugs 18 Nov 2004 00:38:27 -0000 1.389 *************** *** 5528,5531 **** --- 5528,5532 ---- being re-entity-ized when being written out. Sequence names weren't; now they are, too.] + [Also fixed this in current version.] - Guanming's buffered save fix (These fixes were added on top of the June 2004 patches to v1.4.4.) *************** *** 5536,5537 **** --- 5537,5550 ---- as db for prediction analysis and like, and xml has empty database element. (See /users/sshu/tmp/CG1234_chado.game.xml) + + serialized files seem to fail to read in and only get vague error + popup message: "Exception from java.lang.NullPointerException". Bug + should be fixed and error message should be made more informative + (a stack trace would be good). + + Release 1.4.5 doesn't make--it is missing src/java/apollo/dataadapter/ensj19 + (which I added) but it still is having trouble compiling + src/java/apollo/dataadapter/ensj19/AnnotationEnsJAdapterGUI.java + src/java/apollo/dataadapter/ensj19/AnnotationEnsJAdapter.java + Make a new release 1.4.6 (might as well) tar for sourceforge. + |