#! /usr/bin/env guile
!#

(use-modules (srfi srfi-1)
             (srfi srfi-26)
             (ice-9 getopt-long)
             (ice-9 match)
             (ice-9 regex)
             (transform strings)
             (transform sql)
             (transform triples)
             (transform special-forms))



;; Classification Scheme
(define-transformer classification-scheme-species
  (tables (Species))
  (schema-triples
   (gnc:resource_classification_scheme a skos:ConceptScheme)
   (gnc:resource_classification_scheme skos:prefLabel "GeneNetwork Resource Classification Scheme")
   (gnc:resource_classification_scheme skos:definition "A hierarchical classification scheme for organizing GeneNetwork resources by dataset type, resource set (inbredset group), or species.")
   (gnc:resource_classification_scheme xkos:numberOfLevels "3")
   (gnc:resource_classification_scheme xkos:levels gnc:dataset_type)
   (gnc:resource_classification_scheme xkos:levels gnc:set)
   (gnc:resource_classification_scheme xkos:levels gnc:species)
   (gnc:dataset_type a xkos:ClassificationLevel)
   (gnc:dataset_type skos:inScheme gnc:resource_classification_scheme)
   (gnc:dataset_type skos:prefLabel "Dataset Type")
   (gnc:dataset_type skos:definition "Classification level describing the biological or experimental nature of a dataset.   A dataset can either be a probeSet, a genotype or a phenotype.")
   (gnc:dataset_type xkos:depth "1")
   (gnc:dataset_type xkos:nextLevel gnc:set)
   (gnc:dataset_type skos:member gnc:probeset)
   (gnc:dataset_type skos:member gnc:genotype)
   (gnc:dataset_type skos:member gnc:phenotype)
   (gnc:probeset a skos:Concept)
   (gnc:probeset skos:inScheme gnc:dataset_type)
   (gnc:probeset skos:prefLabel "Transcriptomic Datasets")
   (gnc:probeset skos:altLabel "ProbeSet")
   (gnc:probeset skos:definition "A category representing microarray or sequencing probe sets that measure gene expression or other molecular traits.")
   (gnc:probeset skos:note "Individual probe sets are too numerous to list explicitly in this ontology but are available through the GeneNetwork API.")
   (gnc:genotype a skos:Concept)
   (gnc:genotype skos:inScheme gnc:dataset_type)
   (gnc:genotype skos:prefLabel "Genotype Datasets")
   (gnc:genotype skos:altLabel "Genotype")
   (gnc:genotype skos:definition "A category representing genetic marker or variant datasets used for genetic mapping.")
   (gnc:phenotype a skos:Concept)
   (gnc:phenotype skos:inScheme gnc:dataset_type)
   (gnc:phenotype skos:prefLabel "Phenotype Datasets")
   (gnc:phenotype skos:altLabel "Phenotype")
   (gnc:phenotype skos:definition "A category representing measured traits or phenotypes for genetic analysis.")
   (gnc:species a xkos:ClassificationLevel)
   (gnc:species skos:inScheme gnc:resource_classification_scheme)
   (gnc:species xkos:previousLevel gnc:set)
   (gnc:species skos:prefLabel "Species")
   (gnc:species skos:definition "A classification level that that associates a given resource to a species in GeneNetwork.")
   (gnc:species xkos:depth "3"))
  (triples "gnc:species"
    (set skos:member
         (string->identifier "" (remap-species-identifiers (field Species Fullname))))))

(define-transformer classification-scheme-set
  (tables (InbredSet)
          "WHERE public > 0")
  (schema-triples
   (gnc:set a xkos:ClassificationLevel)
   (gnc:set skos:inScheme gnc:resource_classification_scheme)
   (gnc:set xkos:nextLevel gnc:species)
   (gnc:set xkos:previousLevel gnc:dataset_type)
   (gnc:set skos:prefLabel "InbredSet Group")
   (gnc:set skos:definition "A category representing groups of genetically related strains or individuals (inbred sets, recombinant inbred lines, etc.).")
   (gnc:set xkos:depth "2"))
  (triples "gnc:set"
    (set skos:member
         (string->identifier
          "set" (field InbredSet Name InbredSetName) #:separator "_"))))

(define-transformer species
  (tables (Species))
  (schema-triples
   (gnt:has_uniprot_taxon_id a owl:ObjectProperty)
   (gnt:has_uniprot_taxon_id rdfs:label "has uniprot taxonomic id")
   (gnt:has_family a owl:ObjectProperty)
   (gnt:has_family rdfs:label "has family")
   (gnt:has_family skos:definition "Links a species to its taxonomic family")
   (gnt:has_family schema:domainIncludes gnc:species)
   (gnt:has_family schema:domainIncludes gnc:set)
   (gnt:short_name a owl:DatatypeProperty)
   (gnt:short_name rdfs:label "has short name")
   (gnt:short_name rdfs:domain gnc:species)
   (gnt:short_name skos:definition "The short name of a given resource")
   (gnt:belongs_to_species a owl:ObjectProperty)
   (gnt:belongs_to_species rdf:comment "This resource belongs to this species")
   (gnt:belongs_to_species rdfs:label "belongs to species")
   (gnt:belongs_to_species rdfs:range gnc:species))
  (triples
      (string->identifier "" (remap-species-identifiers (field Species Fullname)))
    (set rdf:type 'gnc:species)
    (set rdfs:label (remap-species-identifiers (field Species Fullname)))
    (set skos:prefLabel (field Species MenuName))
    (set skos:altLabel (field Species SpeciesName))
    (set gnt:short_name (field Species Name))
    (set gnt:has_family (field Species Family))
    (set gnt:has_uniprot_taxon_id (ontology
                                   'taxon:
                                   (field Species TaxonomyId)))))

(define-transformer species-fan-out
  (tables (InbredSet
           (left-join Species "ON InbredSet.SpeciesId=Species.Id"))
          "WHERE public > 0")
  (schema-triples
   (gnt:has_strain a owl:ObjectProperty)
   (gnt:has_strain rdfs:range gnc:set)
   (gnt:has_strain rdfs:domain gnc:species)
   (gnt:has_strain owl:inverseOf gnt:belongs_to_species)
   (gnt:has_strain rdfs:label "this resource belongs to this strain.")
   (gnt:has_strain skos:definition "Lists all strains that belong to this resource."))
  (triples (string->identifier "" (remap-species-identifiers (field Species Fullname)))
    (set gnt:has_strain
         (string->identifier "set" (field InbredSet Name InbredSetName) #:separator "_"))))

(define-transformer inbred-set
  (tables (InbredSet
           (left-join Species "ON InbredSet.SpeciesId=Species.Id")
           (left-join MappingMethod
                      "ON InbredSet.MappingMethodId=MappingMethod.Id"))
          "WHERE public > 0")
  (schema-triples
   (gnt:genetic_type a owl:DatatypeProperty)
   (gnt:genetic_type rdfs:label "has genetic type")
   (gnt:genetic_type skos:definition "Describes the genetic architecture of a resource set (e.g., intercross, riset).")
   (gnt:genetic_type rdfs:domain gnc:set)
   (gnt:genetic_type rdfs:range xsd:string)
   (gnt:has_set_code a owl:DatatypeProperty)
   (gnt:has_set_code rdfs:label "has set code")
   (gnt:has_set_code skos:definition "Provides a unique identifier code for a resource set.")
   (gnt:has_set_code rdfs:domain gnc:set)
   (gnt:has_set_code rdfs:range xsd:string)
   ;; Already defined as an owl prop in species
   (gnt:has_family rdfs:domain gnc:set)
   (gnt:uses_mapping_method a owl:ObjectProperty)
   (gnt:uses_mapping_method rdfs:label "mapping method")
   (gnt:uses_mapping_method rdfs:domain gnc:set)
   (gnt:uses_mapping_method rdfs:range gnc:mapping_method)
   (gnt:uses_mapping_method rdfs:comment "The method used to map genetic or experimental data for this resource.")
   (gnt:has_strain a owl:ObjectProperty)
   (gnt:has_strain rdf:comment "Indicates the group the resources belongs to")
   (gnt:has_strain schema:domainIncludes dcat:Dataset)
   (gnt:has_strain schema:domainIncludes gnc:species)
   (gnt:has_strain rdfs:range gnc:set)
   (gnt:has_strain rdfs:label "belongs-to-group"))
  (triples (string->identifier "set" (field InbredSet Name InbredSetName) #:separator "_")
    (set rdf:type 'gnc:set)
    (set rdfs:label (field InbredSet FullName))
    (set skos:prefLabel (field InbredSet Name InbredSetName))
    (set gnt:genetic_type (field InbredSet GeneticType))
    (set gnt:uses_mapping_method
         (string->identifier "mapping_method" (field MappingMethod Name) #:separator "_"))
    (set gnt:has_set_code (field InbredSet InbredSetCode))
    (set gnt:belongs_to_species
         (string->identifier "" (remap-species-identifiers (field Species Fullname))))))



(let* ((option-spec
        '((settings (single-char #\s) (value #t))
          (output (single-char #\o) (value #t))
          (documentation (single-char #\d) (value #t))))
       (options (getopt-long (command-line) option-spec))
       (settings (option-ref options 'settings #f))
       (output (option-ref options 'output #f))
       (documentation (option-ref options 'documentation #f))
       (%connection-settings
        (call-with-input-file settings
          read)))

  (with-documentation
   (name "Species Metadata")
   (connection %connection-settings)
   (table-metadata? #f)
   (prefixes
    '(("dcat:" "<http://www.w3.org/ns/dcat#>")
      ("gn:" "<http://rdf.genenetwork.org/v1/id/>")
      ("gnc:" "<http://rdf.genenetwork.org/v1/category/>")
      ("owl:" "<http://www.w3.org/2002/07/owl#>")
      ("gnt:" "<http://rdf.genenetwork.org/v1/term/>")
      ("schema:" "<https://schema.org/>")
      ("skos:" "<http://www.w3.org/2004/02/skos/core#>")
      ("xkos:" "<http://rdf-vocabulary.ddialliance.org/xkos#>")
      ("xsd:" "<http://www.w3.org/2001/XMLSchema#>")
      ("rdf:" "<http://www.w3.org/1999/02/22-rdf-syntax-ns#>")
      ("rdfs:" "<http://www.w3.org/2000/01/rdf-schema#>")
      ("taxon:" "<http://purl.uniprot.org/taxonomy/>")))
   (inputs
    (list classification-scheme-species
          classification-scheme-set
          species
          species-fan-out
          inbred-set))
   (outputs
    `(#:documentation ,documentation
      #:rdf ,output))))