cdmbl 0.11.1 → 0.12.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: c35f23365189a85dd22882bc8fb7c4c07bf41e08
-  data.tar.gz: d80ba997c1cb74b63e40dbed80466407ed41d275
+  metadata.gz: 43e5cf62c413f7fc69f95751f5beccea22da21c6
+  data.tar.gz: 6c9d1fa56a87cfd5f48ff21dd5a91f5cfaff0fb7
 SHA512:
-  metadata.gz: 8fd6c40c05f25c803ce6d2e26500e4d9082c0006b9f835b92710208b0a3c3fa8369987ea02052cf5fc9a451038b3da1a9411328d12df60b28230bef628312fce
-  data.tar.gz: 2217bee7f887ea291b3b6fdfc842f0db063ff058340e4ff8af654708f4330ff64a32cb2e650a8793bb4e33cb03aed119a8c22947a68f8a9d3d3a3360f2a2acc8
+  metadata.gz: b147ab7d388f3f06c1ac47b66825fee6216c24972eff496440e576946d1946d23b9707d1dd419b8c2c4bbfb76ed7ae099c9f5d60accdb46d91ff307134583f1d
+  data.tar.gz: 7ae01db4bc7d3e3fbd8593f9cfb0f6e4f13ed09d98bad66741a89b555de68f4c22900a55a6ac17e296b92e117f77d8f04009ef0bfd28925e12892672e28ece3a

data/lib/cdmbl/regex_filter_callback.rb

@@ -0,0 +1,19 @@
+module CDMBL
+  # Search an OAI ListSets field using a regular expression
+  class RegexFilterCallback
+    attr_reader :field, :pattern, :inclusive
+    def initialize(field: 'setName', pattern: /.*/, inclusive: true)
+      @field     = field
+      @pattern   = pattern
+      @inclusive = inclusive
+    end
+
+    def valid?(set: {})
+      inclusive ? matches?(set) : !matches?(set)
+    end
+
+    def matches?(set)
+      pattern.match?(set[field])
+    end
+  end
+end

data/lib/cdmbl/tasks/etl.rake

@@ -30,20 +30,17 @@ namespace :cdmbl do
     :inclusive,
     :batch_size
   ] do |t, args|
-    # Required args
     oai_endpoint = args.fetch(:oai_endpoint)
-    solr_url = args.fetch(:solr_url)
-    cdm_endpoint = args.fetch(:cdm_endpoint)
     # Optional args
-    pattern    = args.fetch(:set_spec_pattern, false)
-    inclusive  = args.fetch(:inclusive, true)
-    batch_size = args.fetch(:batch_size, 5)
-
+    pattern = args.fetch(:set_spec_pattern, false)
+    inclusive = args.fetch(:inclusive, 'true') == 'true'
     # Define your own callback if you want to use other set related fields
-    # Use the SetSpecFilterCallback as an example of how to build your own filter
+    # Use the RegexFilterCallback as an example of how to build your own filter
     set_specs =
       if pattern
-        filter = CDMBL::SetSpecFilterCallback.new(pattern: Regexp.new(pattern))
+        filter = CDMBL::RegexFilterCallback.new(field: 'setName',
+                                                pattern: Regexp.new(pattern),
+                                                inclusive: inclusive)
         CDMBL::FilteredSetSpecs.new(oai_base_url: oai_endpoint,
                                     callback: filter).set_specs
       else
@@ -56,7 +53,7 @@ namespace :cdmbl do
       solr_config: { url: args.fetch(:solr_url) },
       oai_endpoint: args.fetch(:oai_endpoint),
       cdm_endpoint: args.fetch(:cdm_endpoint),
-      batch_size: args.fetch(:batch_size, 10),
+      batch_size: args.fetch(:batch_size, 5),
       max_compounds: args.fetch(:max_compounds, 10)
     }
 

data/lib/cdmbl/version.rb

@@ -1,3 +1,3 @@
 module CDMBL
-  VERSION = "0.11.1"
+  VERSION = "0.12.0"
 end

data/lib/cdmbl.rb

@@ -28,4 +28,4 @@ require 'cdmbl/load_worker'
 require 'cdmbl/transform_worker'
 require 'cdmbl/filtered_set_specs'
 require 'cdmbl/etl_by_set_specs'
-require 'cdmbl/set_spec_filter_callback'
+require 'cdmbl/regex_filter_callback'

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: cdmbl
 version: !ruby/object:Gem::Version
-  version: 0.11.1
+  version: 0.12.0
 platform: ruby
 authors:
 - chadfennell
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2018-04-23 00:00:00.000000000 Z
+date: 2018-04-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: hash_at_path
@@ -238,7 +238,7 @@ files:
 - lib/cdmbl/oai_set_lookup.rb
 - lib/cdmbl/rake_task.rb
 - lib/cdmbl/record_transformer.rb
-- lib/cdmbl/set_spec_filter_callback.rb
+- lib/cdmbl/regex_filter_callback.rb
 - lib/cdmbl/tasks/delete.rake
 - lib/cdmbl/tasks/etl.rake
 - lib/cdmbl/transform_worker.rb

data/lib/cdmbl/set_spec_filter_callback.rb

@@ -1,17 +0,0 @@
-module CDMBL
-  class SetSpecFilterCallback
-    attr_reader :pattern, :inclusive
-    def initialize(pattern: /.*/, inclusive: true)
-      @pattern   = pattern
-      @inclusive = inclusive
-    end
-
-    def valid?(set: {})
-      (inclusive) ? matches?(set) : !matches?(set)
-    end
-
-    def matches?(set)
-      pattern.match?(set['setSpec'])
-    end
-  end
-end