goals

understand masymos ——————- * *.main batch stuff * *.extractor extractors for SBML, CellML, SedML, Owl, XML (=just throws not possible) * owl to neo4j is old rebekka stuff -> can go * *.data : wrapper for fields (to eg. generate JSON?) * fixed structure for info ron wants to have * persons and publication * *.database actual database administration * singleton instance (only one instance) * Manager: * “thinks about it” comment? * for ever index and every feature/filed there is an analyser, see *.analyser * execution engine (runs cypher stuff) * ModelInserter` / `ModelLookup: connection to MartinP, crawling other model repos and import into db * *.database.traverse go from search hit to model node etc * for models and sedml * *configuration * Config: basic static stuff, saved config files * Properties`, `Relation`, `NodeLabel: enums for all labels, properties, etc * *analyser * all have a name, * analyse search terms * eg. lowercasekeywordanalyser: * info missing? * preprocessing before goes into index: split by whitespace, transform into lower case * *annotation * resolver: query miriam web service -> resolve URNs to URLs * fetch: reads text content on the url and stores it in db * *grouping (hehe, was done by rebekka -> ron) * clustering of ontology terms * *query * enumerator: just enumerating the “strings” * results: for each index -> how the results are handled * sedml has problems? nothing in there, written by rebekka * how to handle results sets? * important equals function -> compare result sets * structure is something ron is trying -> skipping * interfaces: all information that is needed if you want to implement an interface * types, eg person * analyser, fields, index defined * lucene specific formats * retrieve/ModelResultByPerson * queryadapter * they changed something in lucene code…? -> should we implement that natively? * util: stuff that fits nowhere else * assign -> cant remember -> was for rebekka -> hungarian? * cmeta -> done by tommy * hungarian -> fancy algorithm -> wikiedia * identifier converter * etc * ontology factory -> include owl * ontology util -> distance, shorted paths etc * resultsetcomparator -> mariam is working on -> thats useful for the diff * structure query -> actually only comments

“much creepy code” (cite r. henkel 2015)

understand diff

• *diff:
• diffexecuter: commandozeilen tool
• executer: sucher doc ohne diff, 5 threads in parallel
• diffjob: eigentliche arbeit: kriegt 2 knoten, erzeugt diff, speichert diff in datenbank:
• iteriere ueber diffs
• hole old/new xpath -> doc nodes
• inherit: go up in tree to find a doc node that has representation in db
• diffutil traversiert graph
• data unspannend, only id kram

TODO: * HAS_DIFF in eine richtung umbenennen

optimize

split into modules

modules: - CellML - - server plugin CORE - SBML - bundle - cli - SedML -

create a roadmap