This page outlines the structure and behavior of the FP-Akka quality control workflow as of 19Feb15.  Links are provided (in blue) to the source code for each key class comprising the workflow.  Indentation is used to indicated class dependencies (has-a and uses relationships).  For example, akka.fp.Loader is dependent on akka.fp.MongoDBReader.  Maven modules packaging each class are given in grey. 

This description is incomplete and includes deep dependencies only for the akka.fp.NewScientificNameValidator actor.

akka.fp.Loader  [FP-Akka module]

• Instantiates a FP-Akka QC Workflow comprising five Akka actors:  akka.fp.MongoDBReader, akka.fp.NewScientificNameValidator, akka.fp.InternalDateValidator, akka.fp.GEORefValidator, and akka.fp.MongoSummaryWriter.  (See below for descriptions of each).
• Configures akka.fp.MongoDBReader and akka.fp.MongoSummaryWriter actors using parsed command-line options.
• Injects fp.services.COLService into the scientificNameService field of the akka.fp.NewScientificName actor.
• Injects fp.services.InternalDateValidationService into the singleDateValidationService field of the akka.fp.InternalDateValidator actor.
• Injects fp.services.GeoLocate3 into the geoRefValidationService field of the akka.fp.GEORefValidator actor.
• Executes the workflow.

akka.fp.MongoDBReader  [FP-Akka module]

• • First actor in the FP-Akka QC workflow.

akka.fp.NewScientificNameValidator  [FP-Akka module]

• • Second actor in the FP-Akka QC workflow.
  • Akka actor for validating scientific name and authorship fields.
  • Uses the class injected into the scientificNameService field (fp.services.COLService and its parent fp.services.SciNameServiceParent) to carry out validation tasks.
  • Receives individual SpecimenRecord instances from upstream MongoDBReader.
  • For each specimen record:
    • Disassembles the specimen record into fields.
    • Calls the validateScientificName() method on the scientificNameService (implemented by fp.services.SciNameServiceParent) passing the individual field values extracted from the record.
    • Calls getters getCurationStatus(), getCorrectedScientificName(), getCorrectedAuthor(), getLSID(), getComment(), getServiceName() on the scientificNameService to extract validation results.
    • If the result returned by getCurationStatus() is CURATED or Filled_In, replaces the scientificName and scientificNameAuthorship fields in the input specimen record with the results from getCorrectedScientificName() and getCorrectedAuthor() respectively.
    • Adds to the input specimen record three fields with labels scinComment, scinStatus, and scinSource using the results from getCurationStatus(), getComment(), and getServiceName() on the scientificNameService.
    • Forwards the updated specimen record to downstream actors in workflow.
    • Comment by T.M.  The NewScientificNameValidator actor overwrites the original scientificName and scientificNameAuthorship fields in each record that it updates.  Downstream actors in the workflow, including MongoDBWriter which saves the workflow results, do not have programmatic access to the original values in these fields.

fp.services.COLService [FP-KurationServices module]

• • • • Injected by akka.FP.Loader into the scientificNameService field of the akka.fp.NewScientificNameValidator actor.
      • Derived from fp.services.SciNameServiceParent (to which it defers most method calls from the actor).
      • Overrides nameSearchAgainstServices() to look up scientific name and author in the Catalog of Life using the web service at http://www.catalogueoflife.org/col/webservice.

fp.services.SciNameServiceParent [FP-KurationServices module]

• • • • Parent class of fp.Services.COLService.  Implements most of the methods called by akka.fp.NewScientificNameValidator actor on the scientificNameService.
      • Most of name validation logic is defined in validateScientificName() method which calls methods on other objects and services.
      • Begins by checking internal consistency of scientific name fields scientificName,  genus,  subgenus,  specificEpithet,  verbatimTaxonRank,  taxonRank, infraspecificEpithet using the static checkConsistencyToAtomicField() method defined in fp.util.SciNameServiceUtil.
      • Under some conditions calls GNISupportingService.resolveDataSourcesNameInLexicalGroupFromGNI(), SciNameServiceUtil.checklistBankNameSearch(), and nameSearchAgainstServices().
      • Comment by T.M.  The internal logic of fp.services.SciNameServiceParent is unclear to me.
      • Comment by T.M.  A comment on line 150 states that the call to nameSearchAgainstServices() uses the GNI search. The comment value returned to akka.FP.NewScientificNameValidator via getComment() appears to state the same (see line 161). However, as implemented in fp.Services.COLService, it is the Catalog of Life web service that is used.

fp.services.GNISupportingService  [FP-KurationServices module]

• • • • • Implements resolveDataSourcesNameInLexicalGroupFromGNI() called by validateScientificNameAgainstServices() method in fp.services.ScinameServiceParent.

fp.util.SciNameServiceUtil  [FP-KurationServices module]

• • • • • Implements the checkConsistencyToAtomicField() method invoked by fp.Services.SciNameServiceParent.
          • Comment by T.M.  checkConsistencyToAtomicField() appears only to construct a scientific name from atomic fields for comparison with provided scientificName only if there is content in the genus, specificEpithet, and infraspecificEpithet fields.  Otherwise it it reports UNABLE_DETERMINE_VALIDITY, "can't construct sciName from atomic fields." Is this the correct behavior?
        • Implements the checkMisspelling() method invoked by fp.Services.SciNameServiceParent.
        • Implements the checklistBankNameSearch() method invoked by fp.Services.SciNameServiceParent.

edu.harvard.mcz.nametools.NameUsage  [FP-KurationServices module]

• • • • • • • Used by fp.util.SciNameServiceUtil during calls to checklistBankNameSearch().fp.services.InternalDateValidationService

org.gbif.nameparser.NameParser

• • • • • • • Used by fp.util.SciNameServiceUtil during calls to checkConsistencyToAtomicField().

org.gbif.api.model.checklistbank.ParsedName

• • • • • • • Used by fp.util.SciNameServiceUtil during calls to checkConsistencyToAtomicField().

akka.fp.InternalDateValidator  [FP-Akka module]

• • Third actor in the FP-Akka QC workflow.
  • Akka actor for validating the specimen collector an collection date fields.
  • Uses the class injected into the singleDateValidationService field (fp.services.InternalDateValidationService) to carry out validation tasks.
  • Receives individual SpecimenRecord instances from upstream MongoDBReader.
  • For each specimen record:
    • Disassembles the specimen record into fields.
    • Calls the validateDate() method on the singleDateValidationService (implemented by fp.services.InternalDateValidationService) passing the individual field values extracted from the record.
    • Calls getters getCurationStatus(), getCorrectedDate(), getComment(), getServiceName() on the singleDateValidationService to extract validation results.
    • If the result returned by getCurationStatus() is CURATED or Filled_In, replaces the eventDate field in the input specimen record with the results from getCorrectedDate().
    • Adds to the input specimen record three fields with labels dateComment, dateStatus, and dateSource using the results from getCurationStatus(), getComment(), and getServiceName() on the singleDateValidationService.
    • Forwards the updated specimen record to downstream actors in workflow.

fp.services.InternalDateValidationService  [FP-KurationServices module]

• • • • Implements date validation methods called by akka.fp.InternalDateValidator.
      • Most of name validation logic is defined in validateDate() method which calls methods on other objects and services.
      • validateDate() calls the private parseDate() method to check internal consistency of fields.
      • Then uses the checkWithAuthorSolr() method to to validate the collector and and collection date with the Filteredpush entomologists list (in a Solr server) and sets curationStatus to UNABLED_CURATED if collection date is not within the life span of the collector.
      • Although there there is a checkWithAuthorHarvard() method that validates collector and collection date against the Harvard List of Botanists, the code is hardwired only to use the  checkWithAuthorSolr().
      • It appears that validation of collector with respect to collection date is meaningfully performed only for entomologists--not for botanists or other collectors.

akka.fp.GEORefValidator  [FP-Akka module] 

• • Fourth actor in the FP-Akka QC workflow.

akka.fp.MongoSummaryWriter  [FP-Akka module]

• • Fifth and final actor in the FP-Akka QC workflow.