gitlab-elasticsearch-git 0.0.13 → 0.0.14

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: c17d708d0668ff43d5497e36a96feb66787903b3
-  data.tar.gz: 5f0c4d7aa9e70b69644cc9ee7775f7d53caee20d
+  metadata.gz: ba4d3958d1aa1aa5ebc18dd268708607ea22cf59
+  data.tar.gz: a40a0d48a99aaf2d876ed4d733acf41e2b78ee8e
 SHA512:
-  metadata.gz: 836b2f456c9e68ea424ef76ce9d5c74cb9c0434a835575250c161c6898ddc9a837fa71e0c258bb5db11b92477a495b383b4182896c8b2abad7f8c4de1197f993
-  data.tar.gz: 135521c6f5cde51751ec7b0fbb561ff88016e2d584127e3c98bdbd0bd94ff0096ce3b91024816b104901f8a7ea12869ec7dd6db9da0cb92f4f8c03ff037db40b
+  metadata.gz: 90799948082455a0d6cde26b83b3f6b65664b3223ebd8e8edf8e4fbf6af8570c11d041874c088f2eac5d9608f0f7298f47ee1027482e6ad0000b8336e9dab6e7
+  data.tar.gz: 5f7c8eec4c6b73be317eeb394c6f6f928d4976290f1824934d311b4c638eb7c8e445f0fabe5740d7b99e8a84a5ba0602ae99c9f1f874a1fb1d69f626a2729812

data/CHANGELOG

@@ -1,3 +1,8 @@
+0.0.14
+ - Improve code search analyzer (custom word delimeter of source code)
+ - Fix encoding issues
+ - Replace deprecated "filtered" query with "bool" query
+
 0.0.13
  - Improve code search analyzer
 

data/lib/elasticsearch/git/model.rb

@@ -42,7 +42,8 @@ module Elasticsearch
               code_analyzer: {
                 type: 'custom',
                 tokenizer: 'standard',
-                filter: %w(lowercase asciifolding word_delimiter code_stemmer)
+                filter: %w(lowercase asciifolding code_stemmer),
+                char_filter: ["code_mapping"]
               }
             },
             tokenizer: {
@@ -68,7 +69,15 @@ module Elasticsearch
                 type: "stemmer",
                 name: "minimal_english"
               }
-            }
+            },
+            char_filter: {
+              code_mapping: {
+                type: "mapping",
+                mappings: [
+                  ". => ' '"
+                ]
+              }
+            },
           }
         }
       end

data/lib/elasticsearch/git/repository.rb

@@ -184,13 +184,13 @@ module Elasticsearch
                   rid: repository_id,
                   sha: commit.oid,
                   author: {
-                    name: commit.author[:name],
-                    email: commit.author[:email],
+                    name: encode!(commit.author[:name]),
+                    email: encode!(commit.author[:email]),
                     time: commit.author[:time].strftime('%Y%m%dT%H%M%S%z'),
                   },
                   committer: {
-                    name: commit.committer[:name],
-                    email: commit.committer[:email],
+                    name: encode!(commit.committer[:name]),
+                    email: encode!(commit.committer[:email]),
                     time: commit.committer[:time].strftime('%Y%m%dT%H%M%S%z'),
                   },
                   message: encode!(commit.message)
@@ -312,7 +312,7 @@ module Elasticsearch
 
         # Repository id used for identity data from different repositories
         # Update this value if need
-        def set_repository_id id = nil
+        def set_repository_id(id = nil)
           @repository_id = id || path_to_repo
         end
 
@@ -375,15 +375,15 @@ module Elasticsearch
 
           query_hash = {
             query: {
-              filtered: {
-                query: {
+              bool: {
+                must: [{
                   multi_match: {
                     fields: fields,
                     query: "#{query}",
                     operator: :or
                   }
-                },
-              },
+                }]
+              }
             },
             aggs: {
               commitRepositoryFaset: {
@@ -399,15 +399,14 @@ module Elasticsearch
           }
 
           if query.blank?
-            query_hash[:query][:filtered][:query] = { match_all: {}}
+            query_hash[:query][:bool][:must] = { match_all: {}}
             query_hash[:track_scores] = true
           end
 
           if options[:repository_id]
-            query_hash[:query][:filtered][:filter] ||= { and: [] }
-            query_hash[:query][:filtered][:filter][:and] << {
+            query_hash[:query][:bool][:filter] = {
               terms: {
-                "commit.rid" => [options[:repository_id]].flatten
+                'commit.rid' => [options[:repository_id]].flatten
               }
             }
           end
@@ -451,8 +450,8 @@ module Elasticsearch
 
           query_hash = {
             query: {
-              filtered: {
-                query: {
+              bool: {
+                must: {
                   multi_match: {
                     query: query,
                     operator: :and,
@@ -481,20 +480,20 @@ module Elasticsearch
             from: per * (page - 1)
           }
 
+          query_hash[:query][:bool][:filter] = []
+
           if options[:repository_id]
-            query_hash[:query][:filtered][:filter] ||= { and: [] }
-            query_hash[:query][:filtered][:filter][:and] << {
+            query_hash[:query][:bool][:filter] << {
               terms: {
-                "blob.rid" => [options[:repository_id]].flatten
+                'blob.rid' => [options[:repository_id]].flatten
               }
             }
           end
 
           if options[:language]
-            query_hash[:query][:filtered][:filter] ||= { and: [] }
-            query_hash[:query][:filtered][:filter][:and] << {
+            query_hash[:query][:bool][:filter] << {
               terms: {
-                "blob.language" => [options[:language]].flatten
+                'blob.language' => [options[:language]].flatten
               }
             }
           end
@@ -540,12 +539,12 @@ module Elasticsearch
             fields: ['blob.path'],
             query: {
                 fuzzy: {
-                    "repository.blob.path" => { value: query }
+                    'repository.blob.path' => { value: query }
                 },
             },
             filter: {
                 term: {
-                    "repository.blob.rid" => [options[:repository_id]].flatten
+                    'repository.blob.rid' => [options[:repository_id]].flatten
                 }
             },
             size: per,

data/lib/elasticsearch/git/version.rb

@@ -1,5 +1,5 @@
 module Elasticsearch
   module Git
-    VERSION = "0.0.13"
+    VERSION = "0.0.14"
   end
 end

data/spec/main_spec.rb

@@ -27,7 +27,7 @@ describe TestRepository do
     repo.index_commits
 
     TestRepository.__elasticsearch__.refresh_index!
-    
+
     expect(repo.search('test', type: :commit)[:commits][:total_count]).to eq(2)
   end
 
@@ -37,12 +37,12 @@ describe TestRepository do
     repo.index_blobs
 
     TestRepository.__elasticsearch__.refresh_index!
-    
+
     expect(repo.search('test')[:commits][:total_count]).to eq(2)
     expect(repo.search('def')[:blobs][:total_count]).to eq(4)
   end
 
-  it "it indexes specified commits" do
+  it "indexes specified commits" do
     repo = TestRepository.new
     repo.index_commits(
       from_rev: '40f4a7a617393735a95a0bb67b08385bc1e7c66d',
@@ -50,11 +50,11 @@ describe TestRepository do
     )
 
     TestRepository.__elasticsearch__.refresh_index!
-    
+
     expect(repo.search('empty', type: :commit)[:commits][:total_count]).to eq(1)
   end
 
-  it "it indexes specified blobs" do
+  it "indexes specified blobs" do
     repo = TestRepository.new
     repo.index_blobs(
       from_rev: '40f4a7a617393735a95a0bb67b08385bc1e7c66d',
@@ -62,7 +62,31 @@ describe TestRepository do
     )
 
     TestRepository.__elasticsearch__.refresh_index!
-    
+
     expect(repo.search('Permission is hereby granted', type: :blob)[:blobs][:total_count]).to eq(1)
   end
+
+  it "applies repository_id filter for blobs" do
+    repo = TestRepository.new
+    repo.index_blobs
+
+    TestRepository.new("repo_second").index_blobs
+
+    TestRepository.__elasticsearch__.refresh_index!
+
+    expect(TestRepository.__elasticsearch__.search('def').results.count).to eq(8)
+    expect(repo.search('def')[:blobs][:total_count]).to eq(4)
+  end
+
+  it "applies repository_id filter for commits" do
+    repo = TestRepository.new
+    repo.index_commits
+
+    TestRepository.new("repo_second").index_commits
+
+    TestRepository.__elasticsearch__.refresh_index!
+
+    expect(TestRepository.__elasticsearch__.search('test').results.count).to eq(4)
+    expect(repo.search('test')[:commits][:total_count]).to eq(2)
+  end
 end

data/spec/support/test_repository.rb

@@ -1,8 +1,12 @@
 class TestRepository
   include Elasticsearch::Git::Repository
 
+  def initialize(id = 'test_1')
+    @repo_id = id
+  end
+
   def repository_id
-    'test_1'
+    @repo_id
   end
 
   def path_to_repo

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gitlab-elasticsearch-git
 version: !ruby/object:Gem::Version
-  version: 0.0.13
+  version: 0.0.14
 platform: ruby
 authors:
 - Andrey Kumanyaev
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-04-15 00:00:00.000000000 Z
+date: 2016-05-11 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: elasticsearch-model