#! /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))


(define-transformer gn:molecular-trait->gn:dataset
  (tables (Tissue))
  (schema-triples
   (gnc:molecular_trait a owl:Class)
   (gnc:molecular_trait a skos:Concept)
   (gnc:molecular_trait rdfs:subClassOf obo:UBERON_0000479)
   (gnc:molecular_trait rdfs:label "Molecular Trait.   This describes a melecular trait of a given species.  We combine the species name and the tissue name in order to differentiate the traits across different inbredset groups."))
  (triples (string->identifier "trait" (field Tissue Short_Name) #:separator "_")
    (set rdf:type 'gnc:molecular_trait)
    (set skos:prefLabel (field Tissue Name))
    (set skos:altLabel (field Tissue Short_Name))))

(define-transformer gnc:molecular_trait->gn:molecular_trait
  (tables (Tissue))
  (triples "gnc:molecular_trait"
    (set skos:member (string->identifier "trait" (field Tissue Short_Name) #:separator "_"))))

(define-transformer gn:set->gn:dataset
  (tables (Species
           (inner-join InbredSet "ON InbredSet.SpeciesId = Species.Id")
           (inner-join ProbeFreeze "ON ProbeFreeze.InbredSetId = InbredSet.Id")
           (inner-join ProbeSetFreeze "ON ProbeSetFreeze.ProbeFreezeId = ProbeFreeze.Id")
           (inner-join Tissue "ON ProbeFreeze.TissueId = Tissue.Id"))
          "WHERE ProbeSetFreeze.public > 0 AND Species.Name != 'monkey' GROUP BY Species.Name, Tissue.Short_Name")
  (schema-triples
   (gnt:has_probeset_data rdf:type owl:ObjectProperty)
   (gnt:has_probeset_data rdfs:label "this resources has this probeset data.")
   (gnt:has_probeset_data rdfs:comment "Associates a resource with this probeset data.")
   (gnt:has_probeset_data rdfs:domain gnc:set)
   (gnt:has_probeset_data rdfs:range gnc:molecular_trait)
   (gnt:has_probeset_data rdfs:subPropertyOf dct:relation))
  (triples (string->identifier "set" (field InbredSet Name InbredSetName) #:separator "_")
    (multiset gnt:has_probeset_data
              (map (cut string->identifier "dataset" <> #:separator "_")
                   (string-split
                    (field ("GROUP_CONCAT(ProbeSetFreeze.Name SEPARATOR ',')"
                            dataset_name))
                    #\,)))))

(define-transformer gn:dataset->metadata
  (tables (ProbeSetFreeze
           (inner-join ProbeFreeze "ON ProbeSetFreeze.ProbeFreezeId = ProbeFreeze.Id")
           (inner-join InbredSet "ON InbredSet.Id = ProbeFreeze.InbredSetId")
           (inner-join Species "ON InbredSet.SpeciesId = Species.Id")
           (inner-join Tissue "ON ProbeFreeze.TissueId = Tissue.Id")
           (inner-join AvgMethod "ON AvgMethod.AvgMethodId = ProbeSetFreeze.AvgID")
           (inner-join InfoFiles "ON InfoFiles.InfoPageName = ProbeSetFreeze.Name")
           (inner-join Datasets "ON InfoFiles.DatasetId = Datasets.DatasetId")
           (left-join GeneChip "ON GeneChip.Id =  InfoFiles.GeneChipId"))
          "WHERE ProbeSetFreeze.public > 0 AND Species.Name != 'monkey'")
  (schema-triples
   (gnt:has_molecular_trait rdf:type owl:ObjectProperty)
   (gnt:has_molecular_trait rdfs:domain gnc:set)
   (gnt:has_molecular_trait rdfs:range gnc:molecular_trait)
   (gnt:has_molecular_trait rdfs:label "has molecular trait")
   (gnt:uses_genechip a owl:ObjectProperty)
   (gnt:uses_genechip rdfs:domain dcat:Dataset)
   (gnt:uses_genechip skos:definition "The Platform this resource uses..")
   (gnt:uses_normalization_method rdfs:comment "The normalization method used for the molecular traits in this dataset")
   (gnt:uses_normalization_method rdfs:domain dcat:Dataset)
   (gnt:uses_normalization_method rdfs:label "Averaging method used for the molecular traits in this dataset.")
   (gnt:uses_normalization_method rdfs:range gnc:avg_method))
  (triples (string->identifier "dataset" (field ProbeSetFreeze Name) #:separator "_")
    (set dct:created (annotate-field (field ProbeSetFreeze CreateTime) '^^xsd:datetime))
    (set gnt:uses_normalization_method
         (string->identifier "avg_method" (field AvgMethod Name AvgMethodName) #:separator "_"))
    (set gnt:has_molecular_trait
         (string->identifier "trait" (field Tissue Short_Name) #:separator "_"))
    (set gnt:uses_genechip
         (string->identifier "platform" (field GeneChip Name) #:separator "_"))))


(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 "Molecular Traits")
   (connection %connection-settings)
   (table-metadata? #f)
   (prefixes
    '(("dcat:" "<http://www.w3.org/ns/dcat#>")
      ("gn:" "<http://rdf.genenetwork.org/v1/id/>")
      ("obo:" "<http://purl.obolibrary.org/obo/>")
      ("owl:" "<http://www.w3.org/2002/07/owl#>")
      ("xsd:" "<http://www.w3.org/2001/XMLSchema#>")
      ("dct:" "<http://purl.org/dc/terms/>")
      ("xkos:" "<http://rdf-vocabulary.ddialliance.org/xkos#>")
      ("gnt:" "<http://rdf.genenetwork.org/v1/term/>")
      ("skos:" "<http://www.w3.org/2004/02/skos/core#>")
      ("gnc:" "<http://rdf.genenetwork.org/v1/category/>")
      ("rdf:" "<http://www.w3.org/1999/02/22-rdf-syntax-ns#>")
      ("rdfs:" "<http://www.w3.org/2000/01/rdf-schema#>")))
   (inputs
    (list
     gn:dataset->metadata
     gn:molecular-trait->gn:dataset
     gn:set->gn:dataset
     gnc:molecular_trait->gn:molecular_trait))
   (outputs
    `(#:documentation ,documentation
      #:rdf ,output))))