gitlab-elasticsearch-git 0.0.15 → 0.0.16

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 805299d2c9977f838c91ad88a5ae27415345ad8d
-  data.tar.gz: f4c2c5211cb767f10042b1373d4ebbcafe9e0b89
+  metadata.gz: 054efbf7262d71c052ff192800e9bc22920085b4
+  data.tar.gz: c49f66ffa207e47d68527065a7c0cc656a3b658d
 SHA512:
-  metadata.gz: 73ce23ee1c90b94713d5b4dc49878cb6a110de5ebec9a002abc423249532337e0cf544ff0e6bb6b43c118373afd27b34e87645bb4e6a6b1c080e7497287e1b99
-  data.tar.gz: 9e1227a3b202f421667e45e3adfa1bef6db62096f2f00ab8eae3c35bafe60ef12b92bdc533413ccbcb1ead5f5e0d48f1d40b049a96d0ba032c64a05127002a34
+  metadata.gz: fe1ff8b03b42c23ebf79210a2b7e0ee576e5d60934dbafd8b5d08c3371ac74c916d3308d294b7d23f766c9673a38f4563e767baa512439c8f86eaa77aa6eb6ec
+  data.tar.gz: 2e04673c2d31379398109e8cf2534bc9b2880dd3c603334fd12eb1e063cb4d83add0a9ce25ab8baf3561f23a7d4ded9daef672be68c23ee3bd85f73aad091b06

data/.gitlab-ci.yml

@@ -0,0 +1,16 @@
+image: "ruby:2.3"
+
+services:
+  - elasticsearch:latest
+
+variables:
+  ELASTICSEARCH_HOST: "elasticsearch"
+
+before_script:
+  - apt-get update -yqqq; apt-get -o dir::cache::archives="vendor/apt" install -y -qq --force-yes libicu-dev cmake
+  - gem install bundler         # Bundler is not installed with the image
+  - bundle install -j $(nproc)  # Install dependencies
+
+rspec:
+  script:
+  - rspec

data/CHANGELOG

@@ -1,3 +1,6 @@
+0.0.16
+ - Analyzer for CamelCased terms
+
 0.0.15
  - Search through the filenames
 

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    gitlab-elasticsearch-git (0.0.13)
+    gitlab-elasticsearch-git (0.0.15)
       activemodel (~> 4.2)
       activesupport (~> 4.2)
       charlock_holmes (~> 0.7)

data/lib/elasticsearch/git/model.rb

@@ -24,11 +24,6 @@ module Elasticsearch
           index: {
           analysis: {
             analyzer: {
-              human_analyzer: {
-                type: 'custom',
-                tokenizer: 'human_tokenizer',
-                filter: %w(lowercase asciifolding)
-              },
               path_analyzer: {
                 type: 'custom',
                 tokenizer: 'path_tokenizer',
@@ -42,32 +37,30 @@ module Elasticsearch
               code_analyzer: {
                 type: 'custom',
                 tokenizer: 'standard',
-                filter: %w(lowercase asciifolding code_stemmer),
+                filter: %w(code lowercase asciifolding),
                 char_filter: ["code_mapping"]
               }
             },
             tokenizer: {
               sha_tokenizer: {
                 type: "edgeNGram",
-                min_gram: 8,
+                min_gram: 5,
                 max_gram: 40,
                 token_chars: %w(letter digit)
               },
-              human_tokenizer: {
-                type: "nGram",
-                min_gram: 1,
-                max_gram: 20,
-                token_chars: %w(letter digit)
-              },
               path_tokenizer: {
                 type: 'path_hierarchy',
                 reverse: true
               },
             },
             filter: {
-              code_stemmer: {
-                type: "stemmer",
-                name: "minimal_english"
+              code: {
+                type: "pattern_capture",
+                preserve_original: 1,
+                patterns: [
+                  "(\\p{Ll}+|\\p{Lu}\\p{Ll}+|\\p{Lu}+)",
+                  "(\\d+)"
+                ]
               }
             },
             char_filter: {

data/lib/elasticsearch/git/repository.rb

@@ -23,34 +23,51 @@ module Elasticsearch
 
         mapping _timestamp: { enabled: true } do
           indexes :blob do
-            indexes :id,          type: :string, index_options: 'offsets', analyzer: :human_analyzer
-            indexes :rid,         type: :string, index: :not_analyzed
-            indexes :oid,         type: :string, index_options: 'offsets', analyzer: :code_analyzer
-            indexes :commit_sha,  type: :string, index_options: 'offsets', analyzer: :sha_analyzer
-            indexes :path,        type: :string,                           analyzer: :path_analyzer
-            indexes :file_name,   type: :string,                           analyzer: :code_analyzer
-            indexes :content,     type: :string, index_options: 'offsets', analyzer: :code_analyzer
-            indexes :language,    type: :string, index: :not_analyzed
+            indexes :id,          type: :string,
+                                  index_options: 'offsets',
+                                  analyzer: :sha_analyzer
+            indexes :rid,         type: :string,
+                                  index: :not_analyzed
+            indexes :oid,         type: :string,
+                                  index_options: 'offsets',
+                                  analyzer: :sha_analyzer
+            indexes :commit_sha,  type: :string,
+                                  index_options: 'offsets',
+                                  analyzer: :sha_analyzer
+            indexes :path,        type: :string,
+                                  analyzer: :path_analyzer
+            indexes :file_name,   type: :string,
+                                  analyzer: :code_analyzer
+            indexes :content,     type: :string,
+                                  index_options: 'offsets',
+                                  analyzer: :code_analyzer
+            indexes :language,    type: :string,
+                                  index: :not_analyzed
           end
 
           indexes :commit do
-            indexes :id,          type: :string, index_options: 'offsets', analyzer: :human_analyzer
-            indexes :rid,         type: :string, index: :not_analyzed
-            indexes :sha,         type: :string, index_options: 'offsets', analyzer: :sha_analyzer
+            indexes :id,          type: :string,
+                                  index_options: 'offsets',
+                                  analyzer: :sha_analyzer
+            indexes :rid,         type: :string,
+                                  index: :not_analyzed
+            indexes :sha,         type: :string,
+                                  index_options: 'offsets',
+                                  analyzer: :sha_analyzer
 
             indexes :author do
-              indexes :name,      type: :string, index_options: 'offsets', analyzer: :code_analyzer
-              indexes :email,     type: :string, index_options: 'offsets', analyzer: :code_analyzer
+              indexes :name,      type: :string, index_options: 'offsets'
+              indexes :email,     type: :string, index_options: 'offsets'
               indexes :time,      type: :date, format: :basic_date_time_no_millis
             end
 
             indexes :commiter do
-              indexes :name,      type: :string, index_options: 'offsets', analyzer: :code_analyzer
-              indexes :email,     type: :string, index_options: 'offsets', analyzer: :code_analyzer
+              indexes :name,      type: :string, index_options: 'offsets'
+              indexes :email,     type: :string, index_options: 'offsets'
               indexes :time,      type: :date, format: :basic_date_time_no_millis
             end
 
-            indexes :message,     type: :string, index_options: 'offsets', analyzer: :code_analyzer
+            indexes :message,     type: :string, index_options: 'offsets'
           end
         end
 
@@ -331,11 +348,6 @@ module Elasticsearch
           @repository_id
         end
 
-        # For Overwrite
-        def self.repositories_count
-          10
-        end
-
         unless defined?(path_to_repo)
           def path_to_repo
             if @path_to_repo.blank?
@@ -359,8 +371,15 @@ module Elasticsearch
         def client_for_indexing
           @client_for_indexing ||= Elasticsearch::Client.new retry_on_failure: 5
         end
+      end
 
-        def self.search(query, type: :all, page: 1, per: 20, options: {})
+      module ClassMethods
+        # For Overwrite
+        def repositories_count
+          10
+        end
+
+        def search(query, type: :all, page: 1, per: 20, options: {})
           results = { blobs: [], commits: []}
 
           case type.to_sym
@@ -375,9 +394,7 @@ module Elasticsearch
 
           results
         end
-      end
 
-      module ClassMethods
         def search_commit(query, page: 1, per: 20, options: {})
           page ||= 1
 

data/lib/elasticsearch/git/version.rb

@@ -1,5 +1,5 @@
 module Elasticsearch
   module Git
-    VERSION = '0.0.15'
+    VERSION = '0.0.16'
   end
 end

data/spec/main_spec.rb

@@ -3,18 +3,16 @@ require 'spec_helper'
 describe TestRepository do
   before do
     remove_index(TestRepository.index_name)
+    TestRepository.__elasticsearch__.create_index!
   end
 
-  it "creates an index" do
-    expect(index_exist?(TestRepository.index_name)).to be_falsey
-
-    TestRepository.__elasticsearch__.create_index!
+  let(:repo) { TestRepository.new }
 
+  it "creates an index" do
     expect(index_exist?(TestRepository.index_name)).to be_truthy
   end
 
   it "indexes all blobs and searches" do
-    repo = TestRepository.new
     repo.index_blobs
 
     TestRepository.__elasticsearch__.refresh_index!
@@ -23,27 +21,24 @@ describe TestRepository do
   end
 
   it "indexes all commits and searches" do
-    repo = TestRepository.new
     repo.index_commits
 
     TestRepository.__elasticsearch__.refresh_index!
 
-    expect(repo.search('test', type: :commit)[:commits][:total_count]).to eq(2)
+    expect(repo.search('test', type: :commit)[:commits][:total_count]).to eq(3)
   end
 
   it "searches through all types" do
-    repo = TestRepository.new
     repo.index_commits
     repo.index_blobs
 
     TestRepository.__elasticsearch__.refresh_index!
 
-    expect(repo.search('test')[:commits][:total_count]).to eq(2)
+    expect(repo.search('test')[:commits][:total_count]).to eq(3)
     expect(repo.search('def')[:blobs][:total_count]).to eq(4)
   end
 
   it "searches through filename" do
-    repo = TestRepository.new
     repo.index_blobs
 
     TestRepository.__elasticsearch__.refresh_index!
@@ -55,8 +50,20 @@ describe TestRepository do
     expect(found_version_file).to be_truthy
   end
 
+  it "searches through camel cased words" do
+    TestRepository.__elasticsearch__.create_index!(force: true)
+    repo.index_blobs
+
+    TestRepository.__elasticsearch__.refresh_index!
+
+    found_version_file = repo.search('Hip')[:blobs][:results].any? do |result|
+      result["_source"]["blob"]["file_name"] == "camelCase.rb"
+    end
+
+    expect(found_version_file).to be_truthy
+  end
+
   it "indexes specified commits" do
-    repo = TestRepository.new
     repo.index_commits(
       from_rev: '40f4a7a617393735a95a0bb67b08385bc1e7c66d',
       to_rev: '732401c65e924df81435deb12891ef570167d2e2'
@@ -68,7 +75,6 @@ describe TestRepository do
   end
 
   it "indexes specified blobs" do
-    repo = TestRepository.new
     repo.index_blobs(
       from_rev: '40f4a7a617393735a95a0bb67b08385bc1e7c66d',
       to_rev: '732401c65e924df81435deb12891ef570167d2e2'
@@ -80,7 +86,6 @@ describe TestRepository do
   end
 
   it "applies repository_id filter for blobs" do
-    repo = TestRepository.new
     repo.index_blobs
 
     TestRepository.new("repo_second").index_blobs
@@ -92,14 +97,13 @@ describe TestRepository do
   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)
+    expect(TestRepository.__elasticsearch__.search('test').results.count).to eq(6)
+    expect(repo.search('test')[:commits][:total_count]).to eq(3)
   end
-end
+end

data/spec/spec_helper.rb

@@ -13,4 +13,4 @@ RSpec.configure do |config|
   config.include ElasticHelper
   config.include SeedHelper
   config.before(:all) { ensure_seeds }
-end
+end

data/spec/support/elastic_helper.rb

@@ -9,6 +9,7 @@ module ElasticHelper
   end
 
   def elastic_url
-    'http://localhost:9200/'
+    host = ENV['ELASTICSEARCH_HOST'] || 'localhost'
+    "http://#{host}:9200/"
   end
-end
+end

data/spec/support/seed_helper.rb

@@ -1,7 +1,7 @@
 require 'fileutils'
 
 module SeedHelper
-  GITLAB_URL = "git@gitlab.com:gitlab-org/gitlab-elasticsearch-git-test.git"
+  GITLAB_URL = "https://gitlab.com/gitlab-org/gitlab-elasticsearch-git-test.git"
 
   def ensure_seeds
     if File.exists?(SUPPORT_PATH)
@@ -22,4 +22,4 @@ module SeedHelper
   def git_env
     {'GIT_TEMPLATE_DIR' => ''}
   end
-end
+end

data/spec/support/test_repository.rb

@@ -1,10 +1,20 @@
 class TestRepository
   include Elasticsearch::Git::Repository
 
+  CLIENT = Elasticsearch::Client.new(
+    host: (ENV['ELASTICSEARCH_HOST'] || 'localhost')
+  )
+
+  self.__elasticsearch__.client = CLIENT
+
   def initialize(id = 'test_1')
     @repo_id = id
   end
 
+  def client_for_indexing
+    CLIENT
+  end
+
   def repository_id
     @repo_id
   end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gitlab-elasticsearch-git
 version: !ruby/object:Gem::Version
-  version: 0.0.15
+  version: 0.0.16
 platform: ruby
 authors:
 - Andrey Kumanyaev
@@ -10,7 +10,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-05-20 00:00:00.000000000 Z
+date: 2016-08-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: elasticsearch-model
@@ -119,6 +119,7 @@ extensions: []
 extra_rdoc_files: []
 files:
 - ".gitignore"
+- ".gitlab-ci.yml"
 - ".rspec"
 - CHANGELOG
 - Gemfile