Check why nub build produces imperfect taxonomic trees

11056
Reporter: mdoering
Assignee: mdoering
Type: Task
Summary: Check why nub build produces imperfect taxonomic trees
Priority: Major
Resolution: WontFix
Status: Resolved
Created: 2012-05-09 09:37:51.845
Updated: 2015-03-02 15:17:56.366
Resolved: 2015-03-02 15:17:56.336
        
Description: The last nub build from today has these logs which stated that synonyms of synonyms and also taxonomic hierarchy loops existed in the initial nub.
These are fixed/broken by the import routine, but they should not have existed in the first place. Investiagate!

---

17:38:15.177 Built nub with NubKeys with 3302365 canonical keys, 756117 canonical names being a potential homonym and 28228 true homonyms.
17:38:15.177 Persisting 4118840 nub usages to checklist bank ...
17:38:15.212 import schema import1 initialised.
17:38:15.212 import schema import1 initialised.
17:38:44.631 Finished Persisting nub usages to import schema in 29s)
17:38:59.038 Finished Persisting nub source ids to import schema in 14s)
17:38:59.043 Prepare temporary import classification import1.classification with usages from import1.name_usage ...
17:38:59.043 Updating classification in import1.classification with usages from import1.name_usage ...
17:38:59.043 Creating a new classification table import1.classification to update table import1.classification
17:38:59.054 Copy usages from import1.name_usage ...
17:40:22.798 Generating classification import1.classification
17:40:22.798 Adding required indices on table import1.classification ...
17:40:35.371 Normalising source adjacency data...
17:40:40.527 Resolved 0 accepted taxa pointing to synonyms which in turn point to the same accepted taxon
17:40:47.302 Resolved 928 synonyms of synonyms
17:41:01.718 Resolved 37897 accepted taxa pointing to synonyms
17:41:08.988 Broke 130 hierarchy loops at level 1
17:41:23.077 Broke 1 hierarchy loops at level 2
17:41:41.508 Broke 0 hierarchy loops at level 3
17:42:05.448 Broke 0 hierarchy loops at level 4
17:42:05.448 Normalize parent_fk to be 0 instead of pointing to itself on import1.classification ...
17:42:06.134 Set parent_kf=0 for 0 usages
17:42:08.896 16 distinct ranks found.
17:42:08.897 Start walking the taxonomy...]]>
    

Author: mdoering@gbif.org
Comment: the backbone generation code will be rewritten using neo4j
Created: 2015-03-02 15:17:56.364
Updated: 2015-03-02 15:17:56.364