Merge pull request #80 from ncbo/develop

Develop to master merge, release v5.30.0
ncbo · May 6, 2024 · cd44c1d · cd44c1d
2 parents f239b34 + 98dc1a9
commit cd44c1d
Show file tree

Hide file tree

Showing 5 changed files with 52 additions and 75 deletions.
diff --git a/.gitignore b/.gitignore
@@ -6,6 +6,7 @@ config/*.p12
 config/*.json
 data/
 projectFilesBackup/
+.bundle
 .ruby-version
 repo*
 *.turtle

diff --git a/Gemfile b/Gemfile
@@ -24,10 +24,10 @@ gem 'sys-proctable'
 gem 'cube-ruby', require: 'cube'
 
 # NCBO
-gem 'goo', github: 'ncbo/goo', branch: 'master'
-gem 'ncbo_annotator', github: 'ncbo/ncbo_annotator', branch: 'master'
-gem 'ontologies_linked_data', github: 'ncbo/ontologies_linked_data', branch: 'master'
-gem 'sparql-client', github: 'ncbo/sparql-client', branch: 'master'
+gem 'goo', github: 'ncbo/goo', branch: 'develop'
+gem 'ncbo_annotator', github: 'ncbo/ncbo_annotator', branch: 'develop'
+gem 'ontologies_linked_data', github: 'ncbo/ontologies_linked_data', branch: 'develop'
+gem 'sparql-client', github: 'ncbo/sparql-client', branch: 'develop'
 
 group :test do
   gem 'email_spec'

diff --git a/Gemfile.lock b/Gemfile.lock
@@ -1,22 +1,23 @@
 GIT
   remote: https://github.com/ncbo/goo.git
-  revision: f01386fe55ffba39cbf5e0bf95c8e2dc18767711
-  branch: master
+  revision: 33583fd1c1d72b449cd6bb91815e3422b8448983
+  branch: develop
   specs:
     goo (0.0.2)
       addressable (~> 2.8)
       pry
       rdf (= 1.0.8)
       redis
+      request_store
       rest-client
       rsolr
       sparql-client
       uuid
 
 GIT
   remote: https://github.com/ncbo/ncbo_annotator.git
-  revision: 63c986880aa88c9384043e6611a682434a14aba7
-  branch: master
+  revision: 7531e10ad55ac66e925c099d7fc05a5a3ceae67e
+  branch: develop
   specs:
     ncbo_annotator (0.0.1)
       goo
@@ -26,8 +27,8 @@ GIT
 
 GIT
   remote: https://github.com/ncbo/ontologies_linked_data.git
-  revision: 4dc3b1123e871b53ba9eb51983ff69c5d2c75c18
-  branch: master
+  revision: a3f6cf0493bb595011f98acf95efda33111c537b
+  branch: develop
   specs:
     ontologies_linked_data (0.0.1)
       activesupport
@@ -46,8 +47,8 @@ GIT
 
 GIT
   remote: https://github.com/ncbo/sparql-client.git
-  revision: d418d56a6c9ff5692f925b45739a2a1c66bca851
-  branch: master
+  revision: 1657f0dd69fd4b522d3549a6848670175f5e98cc
+  branch: develop
   specs:
     sparql-client (1.0.1)
       json_pure (>= 1.4)
@@ -59,6 +60,8 @@ PATH
   specs:
     ncbo_cron (0.0.1)
       dante
+      faraday (~> 2)
+      faraday-follow_redirects (~> 0)
       goo
       google-analytics-data
       mlanett-redis-lock
@@ -97,8 +100,10 @@ GEM
       base64
       faraday-net_http (>= 2.0, < 3.1)
       ruby2_keywords (>= 0.0.4)
+    faraday-follow_redirects (0.3.0)
+      faraday (>= 1, < 3)
     faraday-net_http (3.0.2)
-    faraday-retry (2.2.0)
+    faraday-retry (2.2.1)
       faraday (~> 2.0)
     ffi (1.16.3)
     gapic-common (0.21.1)
@@ -146,10 +151,10 @@ GEM
       multi_json (~> 1.11)
       os (>= 0.9, < 2.0)
       signet (>= 0.16, < 2.a)
-    grpc (1.62.0)
+    grpc (1.63.0)
       google-protobuf (~> 3.25)
       googleapis-common-protos-types (~> 1.0)
-    grpc (1.62.0-x86_64-linux)
+    grpc (1.63.0-x86_64-linux)
       google-protobuf (~> 3.25)
       googleapis-common-protos-types (~> 1.0)
     htmlentities (4.3.4)
@@ -159,8 +164,8 @@ GEM
     httpclient (2.8.3)
     i18n (0.9.5)
       concurrent-ruby (~> 1.0)
-    json (2.7.1)
-    json_pure (2.7.1)
+    json (2.7.2)
+    json_pure (2.7.2)
     jwt (2.8.1)
       base64
     launchy (2.5.2)
@@ -174,7 +179,7 @@ GEM
       net-imap
       net-pop
       net-smtp
-    method_source (1.0.0)
+    method_source (1.1.0)
     mime-types (3.5.2)
       mime-types-data (~> 3.2015)
     mime-types-data (3.2024.0305)
@@ -205,21 +210,23 @@ GEM
     pry (0.14.2)
       coderay (~> 1.1)
       method_source (~> 1.0)
-    public_suffix (5.0.4)
+    public_suffix (5.0.5)
     rack (3.0.10)
     rack-test (2.1.0)
       rack (>= 1.3)
-    rake (13.1.0)
+    rake (13.2.1)
     rdf (1.0.8)
       addressable (>= 2.2)
-    redis (5.1.0)
-      redis-client (>= 0.17.0)
-    redis-client (0.21.1)
+    redis (5.2.0)
+      redis-client (>= 0.22.0)
+    redis-client (0.22.1)
       connection_pool
     representable (3.2.0)
       declarative (< 0.1.0)
       trailblazer-option (>= 0.1.1, < 0.2.0)
       uber (< 0.2.0)
+    request_store (1.7.0)
+      rack (>= 1.4)
     rest-client (2.1.0)
       http-accept (>= 1.7.0, < 2.0)
       http-cookie (>= 1.0.2, < 2.0)

diff --git a/lib/ncbo_cron/obofoundry_sync.rb b/lib/ncbo_cron/obofoundry_sync.rb
@@ -1,42 +1,39 @@
-require 'base64'
-require 'json'
-require 'net/http'
-require 'uri'
+require 'faraday'
+require 'faraday/follow_redirects'
+require 'multi_json'
 
 module NcboCron
   module Models
     class OBOFoundrySync
 
       def initialize
         @logger = Logger.new(STDOUT)
-        @oauth_token = Base64.decode64(NcboCron.settings.git_repo_access_token)
-        @graphql_uri = URI.parse("https://api.github.com/graphql")
-        @request_options = { use_ssl: @graphql_uri.scheme == "https" }
       end
 
       def run
         # Get a map of OBO ID spaces to BioPortal acronyms
         map = get_ids_to_acronyms_map
 
         onts = get_obofoundry_ontologies
-        @logger.info("Found #{onts.size} OBO Library ontologies")
+        @logger.info("Found #{onts.size} OBO Foundry ontologies")
 
-        # Are any OBO Library ontologies missing from BioPortal?
+        # Are any OBO Foundry ontologies missing from BioPortal?
         missing_onts = []
-        active_onts = onts.reject { |ont| ont.key?("is_obsolete") }
+        active_onts = onts.reject { |ont| ont.key?('is_obsolete') }
+        @logger.info("#{active_onts.size} OBO Foundry ontologies are currently active")
         active_onts.each do |ont|
-          if not map.key?(ont["id"])
+          if not map.key?(ont['id'])
             missing_onts << ont
-            @logger.info("Missing OBO Library ontology: #{ont['title']} (#{ont['id']})")
+            @logger.info("Missing OBO Foundry ontology: #{ont['title']} (#{ont['id']})")
           end
         end
 
-        # Have any of the OBO Library ontologies that BioPortal hosts become obsolete?
+        # Have any of the OBO Foundry ontologies that BioPortal hosts become obsolete?
         obsolete_onts = []
-        ids = active_onts.map{ |ont| ont["id"] }
+        ids = active_onts.map{ |ont| ont['id'] }
         obsolete_ids = map.keys - ids
         obsolete_ids.each do |id|
-          ont = onts.find{ |ont| ont["id"] == id }
+          ont = onts.find{ |ont| ont['id'] == id }
           @logger.info("Deprecated OBO Library ontology: #{ont['title']} (#{ont['id']})")
           obsolete_onts << ont
         end        
@@ -45,49 +42,19 @@ def run
       end
 
       def get_ids_to_acronyms_map
-        query = "query { 
-                  repository(name: \"ncbo.github.io\", owner: \"ncbo\") {
-                    object(expression: \"master:oboids_to_bpacronyms.json\") {
-                      ... on Blob {
-                        text
-                      }
-                    }
-                  }
-                }"
-
-        response = issue_request(query)
-        JSON.parse(response)
+        response = Faraday.get('https://ncbo.github.io/oboids_to_bpacronyms.json')
+        MultiJson.load(response.body)
       end
 
       def get_obofoundry_ontologies
-        query = "query { 
-                  repository(name: \"OBOFoundry.github.io\", owner: \"OBOFoundry\") {
-                    object(expression: \"master:registry/ontologies.jsonld\") {
-                      ... on Blob {
-                        text
-                      }
-                    }
-                  }
-                }"
-
-        response = issue_request(query)
-        ont_registry = JSON.parse(response)
-        ont_registry["ontologies"].to_a
-      end
-
-      def issue_request(query)
-        request = Net::HTTP::Post.new(@graphql_uri)
-        request["Authorization"] = "bearer #{@oauth_token}"
-        request.body = JSON.dump({"query" => query})
-
-        response = Net::HTTP.start(@graphql_uri.hostname, @graphql_uri.port, @request_options) do |http|
-          http.request(request)
+        conn = Faraday.new do |faraday|
+          faraday.response :follow_redirects
+          faraday.adapter Faraday.default_adapter
         end
+        response = conn.get('http://purl.obolibrary.org/meta/ontologies.jsonld')
 
-        parsed = JSON.parse(response.body)
-        parsed.dig("data", "repository", "object", "text")
+        MultiJson.load(response.body)['ontologies']
       end
-
     end
   end
 end
diff --git a/ncbo_cron.gemspec b/ncbo_cron.gemspec
@@ -15,6 +15,8 @@ Gem::Specification.new do |gem|
   gem.require_paths = ["lib"]
 
   gem.add_dependency("dante")
+  gem.add_dependency('faraday', '~> 2')
+  gem.add_dependency('faraday-follow_redirects', '~> 0')
   gem.add_dependency("goo")
   gem.add_dependency("google-analytics-data")
   gem.add_dependency("mlanett-redis-lock")