Merge pull request #37 from biolink/biolink-issue-48

Added configurability of remove services

Makefile

@@ -10,6 +10,15 @@ subpackage_tests: $(patsubst %,test-%,$(PACKAGES))
 test:
 	pytest tests/*.py
 
+debug_test:
+	pytest -s -vvvv tests/*.py
+
+t-%:
+	pytest tests/test_$*.py
+
+foo:
+	which pytest
+
 # only run local tests
 travis_test:
 	pytest tests/test_*local*.py tests/test_*parser*.py
@@ -25,5 +34,5 @@ release: cleandist
 nb:
 	PYTHONPATH=.. jupyter notebook
 
-foo:
+mm:
 	./bin/flask2marshmallow.pl ../biolink-api/biolink/datamodel/serializers.py 

bin/ontobio-assoc.py

@@ -12,16 +12,22 @@
 Examples:
 
 ```
-ogr-assoc -v -r go -T NCBITaxon:9606 -C gene function enrichment -q GO:1903010 
+ontobio-assoc.py -v -r go -T NCBITaxon:9606 -C gene function query -q GO:1903010 
 
-ogr-assoc -v -r go -T NCBITaxon:10090 -C gene function dendrogram GO:0003700 GO:0005215 GO:0005634 GO:0005737 GO:0005739 GO:0005694 GO:0005730  GO:0000228 GO:0000262 
+Query planteome:
 
-ogr-assoc -v -r go -T NCBITaxon:10090 -C gene function simmatrix MGI:1890081 MGI:97487 MGI:106593 MGI:97250 MGI:2151057 MGI:1347473
+ontobio-assoc.py -v -y conf/planteome-config.yaml -r po -T NCBITaxon:3702 -C gene anatomy query -q PO:0025034
 
-ogr-assoc -C gene function -T pombe -r go -f tests/resources/truncated-pombase.gaf query -q GO:0005622
+ontobio-assoc.py -v -r go -T NCBITaxon:9606 -C gene function enrichment -q GO:1903010 
+
+ontobio-assoc.py -v -r go -T NCBITaxon:10090 -C gene function dendrogram GO:0003700 GO:0005215 GO:0005634 GO:0005737 GO:0005739 GO:0005694 GO:0005730  GO:0000228 GO:0000262 
+
+ontobio-assoc.py -v -r go -T NCBITaxon:10090 -C gene function simmatrix MGI:1890081 MGI:97487 MGI:106593 MGI:97250 MGI:2151057 MGI:1347473
+
+ontobio-assoc.py -C gene function -T pombe -r go -f tests/resources/truncated-pombase.gaf query -q GO:0005622
 
 # requires files from ftp://ftp.rgd.mcw.edu/pub/ontology/annotated_rgd_objects_by_ontology
-ogr-assoc  -T . -C gene disease -r doid -f homo_genes_do phenolog -R pw -F homo_genes_pw > PAIRS.txt
+ontobio-assoc.py  -T . -C gene disease -r doid -f homo_genes_do phenolog -R pw -F homo_genes_pw > PAIRS.txt
 ```
 
 """
@@ -61,8 +67,8 @@ def main():
                         help='ECO class')
     parser.add_argument('-p', '--properties', nargs='*', type=str, required=False,
                         help='Properties')
-    parser.add_argument('-s', '--search', type=str, default='', required=False,
-                        help='Search type. p=partial, r=regex')
+    parser.add_argument('-y', '--yamlconfig', type=str, required=False,
+                        help='Path to setup/configuration yaml file')
     parser.add_argument('-S', '--slim', type=str, default='', required=False,
                         help='Slim type. m=minimal')
     parser.add_argument('-c', '--container_properties', nargs='*', type=str, required=False,
@@ -136,6 +142,14 @@ def main():
     if args.verbosity == 1:
         logging.basicConfig(level=logging.INFO)
     logging.info("Welcome!")
+
+    if args.yamlconfig is not None:
+        logging.info("Setting config from: {}".format(args.yamlconfig))
+        # note this sets a global:
+        # we would not do this outside the context of a standalone script
+        from ontobio.config import set_config
+        set_config(args.yamlconfig)
+
 
     handle = args.resource
 

conf/config.yaml

@@ -1,12 +1,28 @@
-solr_assocs_url: "https://solr.monarchinitiative.org/solr/golr"
-amigo_solr_assocs_url: "http://golr.berkeleybop.org"
-solr_search_url: "https://solr-dev.monarchinitiative.org/solr/search"
-amigo_solr_search_url: "http://golr.berkeleybop.org"
-sparql_url: "http://sparql.hegroup.org/sparql"
-scigraph_ontology_url: "https://scigraph-ontology.monarchinitiative.org/scigraph/"
-scigraph_data_url: "https://scigraph-ontology.monarchinitiative.org/scigraph/"
-use_amigo_for: function
+solr_assocs:
+  url: "https://solr.monarchinitiative.org/solr/golr"
+  timeout: 5
+amigo_solr_assocs:
+  url: "http://golr.berkeleybop.org"
+  timeout: 4
+solr_search:
+  url: "https://solr.monarchinitiative.org/solr/search"
+  timeout: 2
+amigo_solr_search:
+  url: "http://golr.berkeleybop.org"
+  timeout: 2
+sparql:
+  url: "http://sparql.hegroup.org/sparql"
+  timeout: 2
+scigraph_ontology:
+  url: "https://scigraph-ontology.monarchinitiative.org/scigraph/"
+  timeout: 2
+scigraph_data:
+  url: "https://scigraph-ontology.monarchinitiative.org/scigraph/"
+  timeout: 2
+use_amigo_for:
+  - function
 ontologies:
   - id: magic1234
-    handle: ceph
+    handle: pato
     pre_load: true
+

conf/planteome-config.yaml

@@ -1,4 +1,34 @@
-amigo_solr_assocs_url: http://browser.planteome.org/solr
-amigo_solr_search_url: http://browser.planteome.org/solr
-sparql_url: http://sparql.hegroup.org/sparql
-use_amigo_for: *
+solr_assocs:
+  url: "http://browser.planteome.org/solr"
+  timeout: 5
+amigo_solr_assocs:
+  url: "http://browser.planteome.org/solr"
+  timeout: 8
+solr_search:
+  url: "http://browser.planteome.org/solr"
+  timeout: 2
+amigo_solr_search:
+  url: "http://browser.planteome.org/solr"
+  timeout: 2
+sparql:
+  url: "http://sparql.hegroup.org/sparql"
+  timeout: 2
+scigraph_ontology:
+  url: "https://scigraph-ontology.monarchinitiative.org/scigraph/"
+  timeout: 2
+scigraph_data:
+  url: "https://scigraph-ontology.monarchinitiative.org/scigraph/"
+  timeout: 2
+categories:
+  - id: anatomy
+    superclass: PO:0025131
+  - id: phenotype
+    superclass: TO:0000387
+  - id: environment
+    superclass: TO:0000387
+default_solr_schema: amigo
+ontologies:
+  - id: magic1234
+    handle: pato
+    pre_load: true
+

ontobio/config.py

@@ -10,24 +10,67 @@ class OntologyConfigSchema(Schema):
     handle = fields.Str(description="ontology handle")
     pre_load = fields.Bool(description="if true, load this ontology at startup")
 
+class EndpointSchema(Schema):
+    """
+    Configuration for a REST or RESTish endpoint
+    """
+    url = fields.Url()
+    timeout = fields.Int()
+
+    @post_load
+    def make_object(self, data):
+        return Endpoint(**data)
+
+class CategorySchema(Schema):
+    """
+    Maps a category label to a root ontology class
+    """
+    id = fields.Str()
+    superclass = fields.Str()
+
+    @post_load
+    def make_object(self, data):
+        return Category(**data)
+
 class ConfigSchema(Schema):
     """
     Marshmallow schema for configuration objects.
     """
-    sparql_url = fields.Url(description="SPARQL URL to use for ontology queries")
-    solr_url = fields.Url()
-    solr_search_url = fields.Url()
-    go_solr_url = fields.Url()
-    go_solr_search_url = fields.Url()
-    scigraph_ontology_url = fields.Url()
-    scigraph_data_url = fields.Url()
+    sparql = fields.Nested(EndpointSchema, description="SPARQL URL to use for ontology queries")
+    solr_assocs = fields.Nested(EndpointSchema)
+    solr_search = fields.Nested(EndpointSchema)
+    amigo_solr_assocs = fields.Nested(EndpointSchema)
+    amigo_solr_search = fields.Nested(EndpointSchema)
+    scigraph_ontology = fields.Nested(EndpointSchema)
+    scigraph_data = fields.Nested(EndpointSchema)
+    default_solr_schema = fields.Str()
     ontologies = fields.List(fields.Nested(OntologyConfigSchema))
+    categories = fields.List(fields.Nested(CategorySchema))
+    use_amigo_for = fields.List(fields.Str(description="category to use amigo for"))
 
     @post_load
     def make_object(self, data):
-        logging.info("POST-LOAD: {}".format(data))
         return Config(**data)
 
+class Endpoint():
+    """
+    RESTish endpoint
+    """
+    def __init__(self,
+                 url = None,
+                 timeout = None):
+        self.url = url
+        self.timeout = timeout
+
+class Category():
+    """
+    Maps category to class
+    """
+    def __init__(self,
+                 id = None,
+                 superclass = None):
+        self.id = id
+        self.superclass = superclass
 
 class Config():
     """
@@ -36,26 +79,42 @@ class Config():
 
     """
     def __init__(self,
-                 solr_assocs_url = "https://solr.monarchinitiative.org/solr/golr",
-                 amigo_solr_assocs_url = "http://golr.berkeleybop.org",
-                 solr_search_url = "https://solr-dev.monarchinitiative.org/solr/search",
-                 amigo_solr_search_url = "http://golr.berkeleybop.org",
-                 sparql_url = "http://sparql.hegroup.org/sparql",
-                 scigraph_ontology_url = None,
-                 scigraph_data_url = None,
-                 ontologies = {},
+                 solr_assocs = None,
+                 amigo_solr_assocs = None,
+                 solr_search = None,
+                 amigo_solr_search = None,
+                 sparql = None,
+                 scigraph_ontology = None,
+                 scigraph_data = None,
+                 ontologies = [],
+                 categories = [],
+                 default_solr_schema = None,
                  use_amigo_for = "function"):
-        self.solr_assocs_url = solr_assocs_url
-        self.amigo_solr_assocs_url = amigo_solr_assocs_url
-        self.solr_search_url = solr_assocs_url
-        self.amigo_solr_search_url = amigo_solr_assocs_url
-        self.sparql_url = sparql_url
-        self.scigraph_ontology_url = scigraph_ontology_url
-        self.scigraph_data_url = scigraph_data_url
+        self.solr_assocs = solr_assocs
+        self.amigo_solr_assocs = amigo_solr_assocs
+        self.solr_search = solr_search
+        self.amigo_solr_search = amigo_solr_search
+        self.sparql = sparql
+        self.scigraph_ontology = scigraph_ontology
+        self.scigraph_data = scigraph_data
         self.ontologies = ontologies
+        self.categories = categories
+        self.default_solr_schema = default_solr_schema
         self.use_amigo_for = use_amigo_for
 
-    def get_solr_search_url(use_amigo=False):
+    def endpoint_url(self, endpoint):
+        if endpoint is None:
+            return None
+        else:
+            return endpoint.url
+
+    def get_category_class(self, categ):
+        matches = [c.superclass for c in self.categories if c.id == categ]
+        if len(matches) > 0:
+            return matches[0]
+        return None
+
+    def get_solr_search_url(self, use_amigo=False):
         """
         Return solr URL to be used for lexical entity searches
 
@@ -66,9 +125,9 @@ def get_solr_search_url(use_amigo=False):
         use_amigo : bool
             If true, get the URL for the GO/AmiGO instance of GOlr. This is typically used for category='function' queries
         """
-        url = self.solr_search_url
+        url = self.endpoint_url(self.solr_search)
         if use_amigo:
-            url = self.amigo_solr_search_url
+            url = self.endpoint_url(self.amigo_solr_search)
         return url
 
     def get_solr_assocs_url(use_amigo=False):
@@ -80,17 +139,18 @@ def get_solr_assocs_url(use_amigo=False):
         There are two possible schemas: Monarch and AmiGO. The AmiGO schema is used for
         querying the GO and Planteome Golr instances
         """
-        url = self.solr_assocs_url
+        url = self.endpoint_url(self.solr_assocs)
         if use_amigo:
-            url = self.amigo_solr_assocs_url
+            url = self.endpoint_url(self.amigo_solr_assocs)
         return url
 
 class Session():
     """
     Configuration for current session
     """
-    default_config_path = 'conf/config.yaml'
-    config = None
+    def __init__(self):
+        self.default_config_path = 'conf/config.yaml'
+        self.config = None
 
 """
     Current session
@@ -111,14 +171,28 @@ def get_config():
         logging.info("Using pre-loaded object: {}".format(session.config))
     return session.config
 
+def set_config(path):
+    """
+    Set configuration for current session.
+    """
+    logging.info("LOADING FROM: {}".format(path))
+    session.config = load_config(path)
+    return session.config
+
+def reset_config():
+    """
+    Reset currrent session configuration
+    """
+    session.config = None
+
 def load_config(path):
     f = open(path,'r')
     obj = yaml.load(f)
     schema = ConfigSchema()
     config = schema.load(obj).data
     errs = schema.validate(obj)
     if len(errs) > 0:
-        logging.error("ERRS: {}".format(errs))
+        logging.error("CONFIG ERRS: {}".format(errs))
         raise ValueError('Error loading '+path)
     #config = Config()
     return config