From: <rv...@us...> - 2009-11-25 12:41:26
|
Revision: 328 http://treebase.svn.sourceforge.net/treebase/?rev=328&view=rev Author: rvos Date: 2009-11-25 12:41:19 +0000 (Wed, 25 Nov 2009) Log Message: ----------- Added section explaining warning messages Modified Paths: -------------- trunk/treebase-core/src/main/perl/check/check Modified: trunk/treebase-core/src/main/perl/check/check =================================================================== --- trunk/treebase-core/src/main/perl/check/check 2009-11-25 12:36:02 UTC (rev 327) +++ trunk/treebase-core/src/main/perl/check/check 2009-11-25 12:41:19 UTC (rev 328) @@ -62,6 +62,97 @@ specified class. For example, "check Study 1234" prints out study #1234 and all of its trees, matrices, analyses, etc., and their contents. +=head1 WARNING MESSAGES + +=head2 *** Analysis # contains data with the following taxonlabelsets: ## + +This means that the taxonlables in one analysis belong to more than +one taxonlabelset. I had an idea at one point tha teach analysis's +taxonlabels should belong to only one TLS. I think Bill explained to +me that this was mistaken. Better double-check. + +=head2 *** Analysis # contains no matrices + +=head2 *** Analysis # contains no taxonlabelsets + +=head2 *** Analysis # contains no trees + +An analysis consists of a step that turns a matrix into one or more +trees. An analysis with no matrices is missing its matrix. I believe +that an analysis with no trees is similarly defective. The +taxonlabels in the matrices and trees should be organized into +taxonlabelsets. + +=head2 *** Matrix # has # rows but nchar= + +The actual number of columns doesn't match the NCHAR in the matrix +table. ("Rows" should be "columns".) + +=head2 *** Matrix # has # rows but ntax= + +The actual number of rows doesn't match the NTAX in the matrix +table. + +=head2 *** Object Matrix ## should have Study = #, but it is missing. + +The study contains an analysis which involces the specified matrix, +but the matrix's STUDY field doesn't point back. + +=head2 *** Object TaxonLabel ## should have Study = #, but instead has # + +=head2 *** Study # has analyzed data with taxon label # which doesn't point back. + +Similarly, this taxonlabel is included in some tree or matrix that is +part of the indicated study, but its STUDY field doesn't point back to +the study. I think these are essentially the same, but I'm not +certain; I'll try to remember to double-check. + +=head2 *** PhyloTree # has null title + +=head2 *** PhyloTree # has null tree_kind + +=head2 *** PhyloTree # has null tree_quality + +=head2 *** PhyloTree # has null tree_type + +Self-explanatory, I think. + +=head2 *** PhyloTreeNode # is a leaf but has no TaxonLabel + +Leaf nodes represent taxa and should have labels. Internal nodes do +not necessarily. + +=head2 *** Study # has no citation + +Should be clear. + +=head2 *** Study # in # submissions. + +Each study should belong to exactly one submission. + +=head2 *** Submission # refers to matrix # which has no study. + +=head2 *** Submission # refers to taxonlabel # which doesn't refer back. + +=head2 *** Submission # refers to tree # which has no study. + +Submissions are linked to matrices, TLs, and trees via the +submission_matrix table, etc. But if the submission is linked to a +matrix, then the matrix should also point back to the study with which +the submission is associated. That is, + + for all submissions s, + for all m in s.matrices, + m.study.submission should be equal to s + +Similarly for trees and TLs. + +=head2 *** tree # references TLS #, but its node # contains TL # which is not in the set. + +"TLS" is taxonlabelset. The tree is associated with a TLS, which +ought to contain all the taxonlabels that are attached to nodes in the +tree. But the tree has a node with a taxonlabel that is not in the TLS. + =head1 AUTHOR Mark Jason Dominus (mj...@pl...) This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site. |