[#58] first draft

Dockerfile

@@ -0,0 +1,35 @@
+#
+# Oracle Java 8 Dockerfile
+#
+# https://github.com/dockerfile/java
+# https://github.com/dockerfile/java/tree/master/oracle-java8
+#
+
+# Pull base image.
+FROM ubuntu:16.04
+
+RUN apt-get -y update && apt-get install -y software-properties-common python-software-properties
+
+# Install Java.
+RUN \
+  echo oracle-java8-installer shared/accepted-oracle-license-v1-1 select true | debconf-set-selections && \
+  add-apt-repository -y ppa:webupd8team/java && \
+  apt-get update && \
+  apt-get install -y oracle-java8-installer && \
+  rm -rf /var/lib/apt/lists/* && \
+  rm -rf /var/cache/oracle-jdk8-installer
+
+# Define commonly used JAVA_HOME variable
+ENV JAVA_HOME /usr/lib/jvm/java-8-oracle
+
+# Define working directory.
+WORKDIR /data
+ADD owlsim-services/target/owlsim-services-3.0-SNAPSHOT.jar /data/
+ADD configuration-samples/configuration-all.yaml /data/configuration.yaml
+
+ADD http://ci.monarchinitiative.org/view/dev/job/create-owlsim-files-on-dev/lastSuccessfulBuild/artifact/server/all.owl /data/
+ADD https://raw.githubusercontent.com/monarch-initiative/monarch-owlsim-data/master/data/Homo_sapiens/Hs_disease_phenotype.txt /data/
+
+CMD java -jar /data/owlsim-services-3.0-SNAPSHOT.jar server /data/configuration.yaml
+
+EXPOSE 8080

README.md

@@ -56,8 +56,18 @@ Paths:
 or for some versions of dropwizard:
 
  * http://localhost:8080/api/docs/
- * http://localhost:8080/api/match/matchers 
+ * http://localhost:8080/api/match/matchers
 
 Example query using default config:
 
 http://localhost:8080/api/match/jaccard?id=X:heart-morphology&id=X:brain-morphology
+
+## Build with Docker
+
+Run those commands from the root directory:
+
+```
+mvn package
+docker build -t owlsim .
+docker run -p 8080:8080 owlsim
+```

configuration-samples/configuration-all.yaml

@@ -0,0 +1,197 @@
+ontologyUris:
+  - /data/all.owl
+ontologyDataUris: []
+dataTsvs: []
+curies:
+  # Skolemize Blank Nodes
+  # overwrite iri fragment with '' for unresovable bnodes
+  '_':    'https://monarchinitiative.org/.well-known/genid/'
+
+  # Monarch-specific
+  '': 'https://monarchinitiative.org/'
+  'MONARCH': 'https://monarchinitiative.org/MONARCH_'
+
+  'MonarchData': 'http://data.monarchinitiative.org/ttl/'
+  'MonarchArchive': 'http://archive.monarchinitiative.org/ttl/'
+
+  # other semantic-web items
+  'Annotation': 'http://www.w3.org/ns/oa#Annotation'  # FIXME - i don't think we're using this
+  'dc': 'http://purl.org/dc/elements/1.1/'
+  'foaf': 'http://xmlns.com/foaf/0.1/'
+
+  # ontologies
+  # [y] indicates those that the monarch team contributes to
+  'AQTLTrait': 'http://identifiers.org/animalqtltrait/'  # FIXME - should get integrated into Upheno
+  'BFO': 'http://purl.obolibrary.org/obo/BFO_' # BFO: Basic Formal Ontology
+  'CHEBI' : 'http://purl.obolibrary.org/obo/CHEBI_' # ChEBI: Chemicals of Biological Interest
+  'CHR' : 'http://purl.obolibrary.org/obo/CHR_'  # CHR: Chromosome Ontology
+  'CL' : 'http://purl.obolibrary.org/obo/CL_'  # CL: Cell Ontology (cell types)  [y]
+  'CLO' : 'http://purl.obolibrary.org/obo/CLO_'  # CLO: Cell Line Ontology  [y]
+  'CMO' : 'http://purl.obolibrary.org/obo/CMO_'  # CMO: Clinical Measurements Ontology
+  'DATA' : 'http://edamontology.org/data_'  # EDAM: Data and Methods Ontology (data artifacts)
+  'DC' : 'http://purl.obolibrary.org/obo/DC_'  # TODO
+  'DECIPHER' : 'http://purl.obolibrary.org/obo/DECIPHER_'  # DECIPHER: Deciphering Developmental Disease
+  'DOID': 'http://purl.obolibrary.org/obo/DOID_'  # DOID: Human Disease Ontology  [y]
+  'ECO': 'http://purl.obolibrary.org/obo/ECO_'  # ECO: Evidence Code Ontology [y]
+  'EFO' : 'http://www.ebi.ac.uk/efo/EFO_'  # EFO: Experimental Factor Ontology (all kinds of stuff) [y]
+  'ENVO' : 'http://purl.obolibrary.org/obo/ENVO_'  # ENVO: Environment Ontology
+  'EOM' : 'http://purl.obolibrary.org/obo/EOM_'  # elements of morphology phentoypes
+  'ERO' : 'http://purl.obolibrary.org/obo/ERO_' # ERO: eagle-i resource ontology  [y]
+  'faldo' : 'http://biohackathon.org/resource/faldo#'  # FALDO: Feature Annotation Location Description Ontology (genomic feature properties)  [y]
+  'FBcv' : 'http://purl.obolibrary.org/obo/FBcv_'  # FBcv: flybase CV (includes phenotypes)
+  'FBbt': 'http://purl.obolibrary.org/obo/FBbt_'  # FBbt: flybase anatomy
+  'FBdv': 'http://purl.obolibrary.org/obo/FBdv_'  # FBdv: flybase developmental stages
+  'GENO': 'http://purl.obolibrary.org/obo/GENO_'  # GENO: Genotype Partonomy Ontology [y]
+  'GO' : 'http://purl.obolibrary.org/obo/GO_'  # GO: Gene Ontology [y]
+  'HP': 'http://purl.obolibrary.org/obo/HP_'  # HP: Human Phenotype Ontology [y]
+  'IAO': 'http://purl.obolibrary.org/obo/IAO_'  # IAO: Information Artifact Ontology [y]
+  'KEGG-ds' : 'http://purl.obolibrary.org/KEGG-ds_'  # KEGG-ds: KEGG Disease Ontology
+  'LPT': 'http://purl.obolibrary.org/obo/LPT_'  # LPT: Livestock Phenotypic Trait Ontology
+  'MA': 'http://purl.obolibrary.org/obo/MA_'  # MA: Mouse Anatomy Ontology [y]
+  'MedGen' : 'http://www.ncbi.nlm.nih.gov/medgen/'  # a vocabulary - should this be in purl?
+  'MESH': 'http://purl.obolibrary.org/obo/MESH_'  # MeSH: Medical Subject Headings (medical diseases, phenotypes, and drugs)
+  'MP': 'http://purl.obolibrary.org/obo/MP_'  # MP: Mammalian Phenotype Ontology [y]
+  'MPATH': 'http://purl.obolibrary.org/obo/MPATH_'  # MPATH: Mammalian Pathology Ontology
+  'NBO': 'http://purl.obolibrary.org/obo/NBO_'  # NBO: NeuroBehavior Ontology [y]
+  'OBA': 'http://purl.obolibrary.org/obo/OBA_'  # OBA: Ontology of Biological Attributes (traits)
+  'OBAN': 'http://purl.org/oban/'  # OBAN: Open Biomedical Annotation Model [y]
+  'OBI': 'http://purl.obolibrary.org/obo/OBI_'  # OBI: Ontology of Biomedical Investigations [y]
+  'OBO': 'http://purl.obolibrary.org/obo/'  # all ontologies in the OBO namespace (this is not itself an ontology)
+  'OIO': 'http://www.geneontology.org/formats/oboInOwl#'  # oboInOwl: obo-specific annotation properties, like synonym types
+  'OMIA' : 'http://purl.obolibrary.org/obo/OMIA_'  # OMIA: Online Mendelian Inheritance in Animals (animal diseases)
+  'OMIM' : 'http://purl.obolibrary.org/obo/OMIM_'  # OMIM: Online Mendelian Inheritance in Man (human disease and variants)
+  'Orphanet' : 'http://www.orpha.net/ORDO/Orphanet_'  # Orphanet: rare diseases and orphan drugs
+  'PATO': 'http://purl.obolibrary.org/obo/PATO_'  # PATO: Phenotypic Quality Ontology [y]
+  'PCO': 'http://purl.obolibrary.org/obo/PCO_'  # PCO: Population and Community Ontology [y]
+  'PR': 'http://purl.obolibrary.org/obo/PR_'  # PRO: protein ontology
+  'PW' : 'http://purl.obolibrary.org/obo/PW_'  # PW: pathway ontology
+  'RO': 'http://purl.obolibrary.org/obo/RO_'  # RO: Relationship Ontology [y]
+  'SIO' : 'http://semanticscience.org/resource/SIO_'  # SIO: SemanticScience Integrated Ontology (information artifacts)
+  'SNOMED' : 'http://purl.obolibrary.org/obo/SNOMED_'  # SNOMED:diseases and phenotypes
+  'SO' : 'http://purl.obolibrary.org/obo/SO_'  # SO: Sequence Ontology [y]
+  'STATO': 'http://purl.obolibrary.org/obo/STATO_'  # Statistics Ontology
+  'UBERON' : 'http://purl.obolibrary.org/obo/UBERON_'  # UBERON: integrated anatomy ontology (metazoans, mostly) [y]
+  'UPHENO' : 'http://purl.obolibrary.org/obo/UPHENO_'  # UPHENO: integrated phenotype ontology, and normal traits [y]
+  'UMLS' : 'http://purl.obolibrary.org/obo/UMLS_'  # UMLS: unified medical language system
+  'UO' : 'http://purl.obolibrary.org/obo/UO_'  # UO: units of measurements
+  'VT' : 'http://purl.obolibrary.org/obo/VT_'  # VT: Vertebrate Trait Ontology
+  'WBPhenotype': 'http://purl.obolibrary.org/obo/WBPhenotype_' # WBPhenotype: WormBase phenotypes (nematode) [y]
+  'XCO' : 'http://purl.obolibrary.org/obo/XCO_'  # XCO: Experimental Conditions Ontology
+  'ZFA': 'http://purl.obolibrary.org/obo/ZFA_'  # ZFA: Zebrafish Anatomy Ontology [y]
+  'ZFS': 'http://purl.obolibrary.org/obo/ZFS_'  # ZFS: Zebrafish Staging [y]
+  'ZP': 'http://purl.obolibrary.org/obo/ZP_'  # ZP: Zebrafish Phenotype Ontology [y]
+  'WBbt': 'http://purl.obolibrary.org/obo/WBbt_' #WBbt:  C. elegans gross anatomy
+  'EMAPA': 'http://purl.obolibrary.org/obo/EMAPA_' # EMAPA: Mouse gross anatomy and development, timed
+  'XAO': 'http://purl.obolibrary.org/obo/XAO_' # XAO: Xenopus anatomy and development
+
+  # publication/reference sources
+  'DOI' : 'http://dx.doi.org/'
+  'GeneReviews' : 'http://www.ncbi.nlm.nih.gov/books/'  # diseases too
+  'ISBN': 'https://monarchinitiative.org/ISBN_'
+  'ISBN-10': 'https://monarchinitiative.org/ISBN10_'
+  'ISBN-13': 'https://monarchinitiative.org/ISBN13_'
+  'ISBN-15': 'https://monarchinitiative.org/ISBN15_'
+  'J' : 'http://www.informatics.jax.org/reference/J:'  # MGI-internal identifiers for pubs
+  'MPD':  'http://phenome.jax.org/'
+  'MPD-assay': 'http://phenome.jax.org/db/qp?rtn=views/catlines&keymeas='
+  'PMID': 'http://www.ncbi.nlm.nih.gov/pubmed/'
+  'PMCID' : 'http://www.ncbi.nlm.nih.gov/pmc/'
+  'AQTLPub' : 'http://www.animalgenome.org/cgi-bin/QTLdb/BT/qabstract?PUBMED_ID='
+  'GO_REF' : 'http://www.geneontology.org/cgi-bin/references.cgi#GO_REF:'
+  'HPO' : 'http://human-phenotype-ontology.org/'  # to be used for persons, though they don't resolve with this
+
+  # strains, lines, or organismal reagents
+  'APB': 'http://pb.apf.edu.au/phenbank/strain.html?id='
+  'CMMR': 'http://www.cmmr.ca/order.php?t=m&id='
+  'Coriell' : 'https://catalog.coriell.org/0/Sections/Search/Sample_Detail.aspx?Ref='
+  'CoriellCollection' : 'https://catalog.coriell.org/1/'
+  'CoriellFamily' : 'https://catalog.coriell.org/0/Sections/BrowseCatalog/FamilyTypeSubDetail.aspx?fam='
+  'CoriellIndividual' : 'https://catalog.coriell.org/Search?q='
+  'dbSNPIndividual' : 'http://www.ncbi.nlm.nih.gov/SNP/snp_ind.cgi?ind_id='  # FIXME
+  'EMMA' : 'https://www.infrafrontier.eu/search?keyword=EM:'
+  'JAX' : 'http://jaxmice.jax.org/strain/'
+  'MMRRC' : 'https://www.mmrrc.org/catalog/sds.php?mmrrc_id='
+  'MPD-strain':  'http://phenome.jax.org/db/q?rtn=strains/details&strainid='
+  'MUGEN': 'http://bioit.fleming.gr/mugen/Controller?workflow=ViewModel&expand_all=true&name_begins=model.block&eid='
+  'NCIMR': 'https://mouse.ncifcrf.gov/available_details.asp?ID='
+  'RBRC': 'http://www2.brc.riken.jp/lab/animal/detail.php?brc_no=RBRC'
+
+  # organisms and genome builds  (also NCBITaxon)
+  'NCBIAssembly': 'http://www.ncbi.nlm.nih.gov/assembly/'
+  'NCBIGenome' : 'http://www.ncbi.nlm.nih.gov/genome/'
+  'NCBITaxon' : 'http://purl.obolibrary.org/obo/NCBITaxon_'
+  'OMIA-breed': 'https://monarchinitiative.org/model/OMIA-breed:'
+  'UCSC' : 'ftp://hgdownload.cse.ucsc.edu/goldenPath/'
+
+  # homology
+  'HOMOLOGENE' : 'http://www.ncbi.nlm.nih.gov/homologene/'
+  'KEGG-ko' : 'http://www.kegg.jp/dbget-bin/www_bget?ko:'
+  'PANTHER' : 'http://www.pantherdb.org/panther/family.do?clsAccession='   # protein/orthologous families
+
+  # variants
+  'AQTL' : 'http://identifiers.org/animalqtl/'  # FIXME temporary  # traits
+  'CGD' : 'http://ohsu.edu/cgd/'  # diseases, variant instances
+  'ClinVar' : 'http://www.ncbi.nlm.nih.gov/clinvar/'  # variant+condition
+  'ClinVarVariant' : 'http://www.ncbi.nlm.nih.gov/clinvar/variation/'
+  'ClinVarSubmitters' : 'http://www.ncbi.nlm.nih.gov/clinvar/submitters/'
+  'COSMIC' : 'http://cancer.sanger.ac.uk/cosmic/mutation/overview?id='
+  'HGMD' : 'http://identifiers.org/hgmd/'
+  'dbSNP' : 'http://www.ncbi.nlm.nih.gov/projects/SNP/snp_ref.cgi?rs='
+  'dbVar' : 'http://www.ncbi.nlm.nih.gov/dbvar/'
+
+  # pathways
+  'KEGG-path' : 'http://www.kegg.jp/dbget-bin/www_bget?path:'
+  'REACT' : 'http://www.reactome.org/PathwayBrowser/#/'
+
+  # genes (and RNAs and transcripts)
+  'BIOGRID' : 'http://thebiogrid.org/'  # also interactions
+  'CCDS' : 'http://www.ncbi.nlm.nih.gov/CCDS/CcdsBrowse.cgi?REQUEST=CCDS&DATA='  # transcripty things
+  'dictyBase' : 'http://dictybase.org/gene/'
+  'EcoGene' : 'http://ecogene.org/gene/'
+  'ENSEMBL' : 'http://identifiers.org/ensembl/'
+  'FlyBase' : 'http://flybase.org/reports/'  # also variants, pubs, genotypes, strains
+  'GenBank' : 'http://www.ncbi.nlm.nih.gov/nuccore/'
+  'HGNC' : 'http://identifiers.org/hgnc/HGNC:'
+  'IMPC' : 'http://www.mousephenotype.org/data/genes/'  # FIXME
+  'KEGG-hsa' : 'http://www.kegg.jp/dbget-bin/www_bget?hsa:'
+  'MGI': 'http://www.informatics.jax.org/accession/MGI:'  # also variants, pubs, genotypes
+  'miRBase' : 'http://identifiers.org/mirbase/'  # microRNA genes
+  'NCBIGene' : 'http://www.ncbi.nlm.nih.gov/gene/'
+  'PomBase' : 'http://identifiers.org/PomBase:'
+  'RefSeq' : 'http://www.ncbi.nlm.nih.gov/refseq/?term='
+  'RGD' : 'http://rgd.mcw.edu/rgdweb/report/gene/main.html?id='
+  'SGD' : 'http://identifiers.org/SGD:'
+  'TAIR' : 'http://identifiers.org/TAIR:'
+  'WormBase' : 'http://identifiers.org/wormbase/'  # also variants, pubs, genotypes
+  'Xenbase' : 'http://identifiers.org/xenbase/'
+  'ZFIN' : 'http://zfin.org/'  # also variants, pubs, genotypes
+
+  # proteins
+  'EC' : 'http://identifiers.org/ec-code/'
+  'HPRD' : 'http://www.hprd.org/protein/'
+  'NCBIProtein' : 'http://www.ncbi.nlm.nih.gov/protein/'
+  'PDB' : 'http://identifiers.org/PDB:'
+  'SwissProt' : 'http://identifiers.org/SwissProt:'
+  'TrEMBL' : 'http://www.uniprot.org/uniprot/'
+  'UniProtKB' : 'http://identifiers.org/uniprot/'
+
+  # SEPIO: Scientific Evidence and Provenance Information Ontology
+  'SEPIO': 'http://purl.obolibrary.org/obo/SEPIO_'
+  'VIVO': 'http://vivoweb.org/ontology/core#'
+
+  #Procedures/protocols
+  'IMPRESS-procedure' : 'https://www.mousephenotype.org/impress/procedures/'
+  'IMPRESS-protocol' : 'https://www.mousephenotype.org/impress/protocol/'
+  'IMPRESS-parameter' : 'https://www.mousephenotype.org/impress/parameterontologies/'
+
+  #Drugs, chemicals, compounds
+  'CID' : 'http://pubchem.ncbi.nlm.nih.gov/compound/'
+  'DrugBank' : 'http://www.drugbank.ca/drugs/'
+  'SIO': 'http://semanticscience.org/resource/SIO_'
+  'OAE': 'http://purl.obolibrary.org/obo/OAE_'
+  'RXCUI': 'http://purl.bioontology.org/ontology/RXNORM/'
+  'MEDDRA': 'http://purl.bioontology.org/ontology/MEDDRA/'
+  'FDADrug': 'http://www.fda.gov/Drugs/InformationOnDrugs/'
+  'BT': 'http://c.biothings.io/#'
+  'UNII': 'http://fdasis.nlm.nih.gov/srs/unii/'
+  'GINAS' : 'http://tripod.nih.gov/ginas/app/substance#'