Added DatabaseAdapter#map_urls and #map_docs

lib/wgit/database/adapters/in_memory.rb

@@ -54,7 +54,8 @@ def size
     #   DB.
     # @return [Array<Wgit::Document>] The search results obtained from the DB.
     def search(
-      query, case_sensitive: false, whole_sentence: true, limit: 10, skip: 0
+      query, case_sensitive: false, whole_sentence: true,
+      limit: 10, skip: 0, &block
     )
       regex = if query.is_a?(Regexp)
                 query
@@ -78,11 +79,7 @@ def search(
       return [] unless results
 
       results = results[0...limit] if limit.positive?
-      results.map do |doc|
-        doc = Wgit::Document.new(doc)
-        yield(doc) if block_given?
-        doc
-      end
+      map_documents(results, &block)
     end
 
     # Deletes everything in the urls and documents collections.
@@ -101,17 +98,13 @@ def empty
     # @param skip [Integer] Skip n amount of Url's.
     # @yield [url] Given each Url object (Wgit::Url) returned from the DB.
     # @return [Array<Wgit::Url>] The uncrawled Urls obtained from the DB.
-    def uncrawled_urls(limit: 0, skip: 0)
+    def uncrawled_urls(limit: 0, skip: 0, &block)
       uncrawled = @urls.reject { |url| url['crawled'] }
       uncrawled = uncrawled[skip..]
       return [] unless uncrawled
 
       uncrawled = uncrawled[0...limit] if limit.positive?
-      uncrawled.map do |url_doc|
-        url = Wgit::Url.new(url_doc)
-        yield url if block_given?
-        url
-      end
+      map_urls(uncrawled, &block)
     end
 
     # Inserts or updates the object in the in-memory database.

lib/wgit/database/adapters/mongo_db.rb

@@ -240,14 +240,10 @@ def bulk_upsert(objs)
     # @return [Array<Wgit::Document>] The Documents obtained from the DB.
     def docs(limit: 0, skip: 0, &block)
       results = retrieve(DOCUMENTS_COLLECTION, {},
-                        sort: { date_added: 1 }, limit:, skip:)
+                         sort: { date_added: 1 }, limit:, skip:)
       return [] if results.count < 1 # results#empty? doesn't exist.
 
-      # results.respond_to? :map! is false so we use map and overwrite the var.
-      results = results.map { |doc_hash| Wgit::Document.new(doc_hash) }
-      results.each(&block) if block_given?
-
-      results
+      map_documents(results, &block)
     end
 
     # Returns all Url records from the DB.
@@ -267,11 +263,7 @@ def urls(crawled: nil, limit: 0, skip: 0, &block)
       results = retrieve(URLS_COLLECTION, query, sort:, limit:, skip:)
       return [] if results.count < 1 # results#empty? doesn't exist.
 
-      # results.respond_to? :map! is false so we use map and overwrite the var.
-      results = results.map { |url_doc| Wgit::Url.new(url_doc) }
-      results.each(&block) if block_given?
-
-      results
+      map_urls(results, &block)
     end
 
     # Returns Url records that have been crawled.
@@ -315,7 +307,8 @@ def uncrawled_urls(limit: 0, skip: 0, &block)
     #   DB.
     # @return [Array<Wgit::Document>] The search results obtained from the DB.
     def search(
-      query, case_sensitive: false, whole_sentence: true, limit: 10, skip: 0
+      query, case_sensitive: false, whole_sentence: true,
+      limit: 10, skip: 0, &block
     )
       query = query.to_s.strip
       query.replace("\"#{query}\"") if whole_sentence
@@ -333,12 +326,7 @@ def search(
       results = retrieve(DOCUMENTS_COLLECTION, query,
                         sort: sort_proj, projection: sort_proj,
                         limit:, skip:)
-
-      results.map do |mongo_doc|
-        doc = Wgit::Document.new(mongo_doc)
-        yield(doc) if block_given?
-        doc
-      end
+      map_documents(results, &block)
     end
 
     # Searches the database's Documents for the given query and then searches

lib/wgit/database/database_adapter.rb

@@ -110,5 +110,25 @@ def build_model(obj)
         Wgit::Database::Model.document(obj)
       end
     end
+
+    # Map each DB hash object into a Wgit::Document. Each Document is yielded
+    # if a block is given before returning the mapped Array of Documents.
+    def map_documents(doc_hashes)
+      doc_hashes.map do |doc|
+        doc = Wgit::Document.new(doc)
+        yield(doc) if block_given?
+        doc
+      end
+    end
+
+    # Map each DB hash object into a Wgit::Url. Each Url is yielded
+    # if a block is given before returning the mapped Array of Urls.
+    def map_urls(url_hashes)
+      url_hashes.map do |url|
+        url = Wgit::Url.new(url)
+        yield(url) if block_given?
+        url
+      end
+    end
   end
 end

test/test_in_memory.rb

@@ -138,9 +138,8 @@ def test_search
     assert results.first.score > results.last.score
   end
 
-  def test_search__case_sensitive__whole_sentence
+  def test_search__case_sensitive
     @docs.last.text << 'Foo Bar'
-
     seed { docs @docs }
 
     # Test case_sensitive: false and block.
@@ -155,6 +154,11 @@ def test_search__case_sensitive__whole_sentence
 
     # Test case_sensitive: true.
     assert_empty db.search('foo bar', case_sensitive: true)
+  end
+
+  def test_search__whole_sentence
+    @docs.last.text << 'Foo Bar'
+    seed { docs @docs }
 
     # Test whole_sentence: false.
     results = db.search('bar foo', whole_sentence: false)
@@ -170,8 +174,11 @@ def test_search__case_sensitive__whole_sentence
     assert results.all? { |doc| doc.instance_of? Wgit::Document }
   end
 
-  def test_search__limit__skip
-    # All dev data docs contain the word 'Everest'.
+  def test_search__limit
+    # First doc has highest textScore and so on...
+    @docs.reverse.each_with_index do |doc, i|
+      i.times { doc.text << 'Everest' }
+    end
     seed { docs @docs }
 
     assert_equal 3, db.search('everest').length
@@ -185,6 +192,14 @@ def test_search__limit__skip
       assert_equal @docs[i], doc
       assert_equal @docs[i].url.to_h, doc.url.to_h
     end
+  end
+
+  def test_search__skip
+    # First doc has highest textScore and so on...
+    @docs.reverse.each_with_index do |doc, i|
+      i.times { doc.text << 'Everest' }
+    end
+    seed { docs @docs }
 
     # Test skip.
     results = db.search('everest', skip: 1)

test/test_mongo_db.rb

@@ -278,9 +278,8 @@ def test_urls__with_redirects
     assert_equal redirects_hash, urls.first.redirects
   end
 
-  def test_search__case_sensitive__whole_sentence
+  def test_search__case_sensitive
     @docs.last.text << 'Foo Bar'
-
     seed { docs @docs }
 
     # Test no results.
@@ -299,6 +298,11 @@ def test_search__case_sensitive__whole_sentence
 
     # Test case_sensitive: true.
     assert_empty db.search('foo bar', case_sensitive: true)
+  end
+
+  def test_search__whole_sentence
+    @docs.last.text << 'Foo Bar'
+    seed { docs @docs }
 
     # Test whole_sentence: false.
     results = db.search('bar foo', whole_sentence: false)
@@ -314,10 +318,14 @@ def test_search__case_sensitive__whole_sentence
     assert results.all? { |doc| doc.instance_of? Wgit::Document }
   end
 
-  def test_search__limit__skip
-    # All dev data docs contain the word 'Everest'.
+  def test_search__limit
+    # First doc has highest textScore and so on...
+    @docs.reverse.each_with_index do |doc, i|
+      i.times { doc.text << 'Everest' }
+    end
     seed { docs @docs }
 
+    # Test search.
     assert_equal 3, db.search('everest').length
     assert_equal 3, db.last_result&.count
 
@@ -331,6 +339,14 @@ def test_search__limit__skip
       assert_equal @docs[i], doc
       assert_equal @docs[i].url.to_h, doc.url.to_h
     end
+  end
+
+  def test_search__skip
+    # First doc has highest textScore and so on...
+    @docs.reverse.each_with_index do |doc, i|
+      i.times { doc.text << 'Everest' }
+    end
+    seed { docs @docs }
 
     # Test skip.
     results = db.search('everest', skip: 1)