GitHub Importer - Find users based on their email address

The returned email by the GitHub API is the user's publicly visible
email address (or null if the user has not specified a public email
address in their profile)

5 files changed, 89 insertions, 36 deletions

diff --git a/lib/gitlab/github_import/base_formatter.rb b/lib/gitlab/github_import/base_formatter.rb
index 95dba9a327b..8c80791e7c9 100644
--- a/lib/gitlab/github_import/base_formatter.rb
+++ b/lib/gitlab/github_import/base_formatter.rb
@@ -1,11 +1,12 @@
 module Gitlab
   module GithubImport
     class BaseFormatter
-      attr_reader :formatter, :project, :raw_data
+      attr_reader :client, :formatter, :project, :raw_data
 
-      def initialize(project, raw_data)
+      def initialize(project, raw_data, client = nil)
         @project = project
         @raw_data = raw_data
+        @client = client
         @formatter = Gitlab::ImportFormatter.new
       end
 
@@ -18,19 +19,6 @@ module Gitlab
       def url
         raw_data.url || ''
       end
-
-      private
-
-      def gitlab_user_id(github_id)
-        User.joins(:identities).
-          find_by("identities.extern_uid = ? AND identities.provider = 'github'", github_id.to_s).
-          try(:id)
-      end
-
-      def gitlab_author_id
-        return @gitlab_author_id if defined?(@gitlab_author_id)
-        @gitlab_author_id = gitlab_user_id(raw_data.user.id)
-      end
     end
   end
 end
diff --git a/lib/gitlab/github_import/comment_formatter.rb b/lib/gitlab/github_import/comment_formatter.rb
index 2bddcde2b7c..e21922070c1 100644
--- a/lib/gitlab/github_import/comment_formatter.rb
+++ b/lib/gitlab/github_import/comment_formatter.rb
@@ -1,6 +1,8 @@
 module Gitlab
   module GithubImport
     class CommentFormatter < BaseFormatter
+      attr_writer :author_id
+
       def attributes
         {
           project: project,
@@ -17,11 +19,11 @@ module Gitlab
       private
 
       def author
-        raw_data.user.login
+        @author ||= UserFormatter.new(client, raw_data.user)
       end
 
       def author_id
-        gitlab_author_id || project.creator_id
+        author.gitlab_id || project.creator_id
       end
 
       def body
@@ -52,10 +54,10 @@ module Gitlab
       end
 
       def note
-        if gitlab_author_id
+        if author.gitlab_id
           body
         else
-          formatter.author_line(author) + body
+          formatter.author_line(author.login) + body
         end
       end
 
diff --git a/lib/gitlab/github_import/importer.rb b/lib/gitlab/github_import/importer.rb
index 9a4ffd28438..d95ff4fd104 100644
--- a/lib/gitlab/github_import/importer.rb
+++ b/lib/gitlab/github_import/importer.rb
@@ -110,7 +110,7 @@ module Gitlab
       def import_issues
         fetch_resources(:issues, repo, state: :all, sort: :created, direction: :asc, per_page: 100) do |issues|
           issues.each do |raw|
-            gh_issue = IssueFormatter.new(project, raw)
+            gh_issue = IssueFormatter.new(project, raw, client)
 
             begin
               issuable =
@@ -131,7 +131,8 @@ module Gitlab
       def import_pull_requests
         fetch_resources(:pull_requests, repo, state: :all, sort: :created, direction: :asc, per_page: 100) do |pull_requests|
           pull_requests.each do |raw|
-            gh_pull_request = PullRequestFormatter.new(project, raw)
+            gh_pull_request = PullRequestFormatter.new(project, raw, client)
+
             next unless gh_pull_request.valid?
 
             begin
@@ -209,14 +210,16 @@ module Gitlab
         ActiveRecord::Base.no_touching do
           comments.each do |raw|
             begin
-              comment         = CommentFormatter.new(project, raw)
+              comment = CommentFormatter.new(project, raw, client)
+
               # GH does not return info about comment's parent, so we guess it by checking its URL!
               *_, parent, iid = URI(raw.html_url).path.split('/')
-              if parent == 'issues'
-                issuable = Issue.find_by(project_id: project.id, iid: iid)
-              else
-                issuable = MergeRequest.find_by(target_project_id: project.id, iid: iid)
-              end
+
+              issuable = if parent == 'issues'
+                           Issue.find_by(project_id: project.id, iid: iid)
+                         else
+                           MergeRequest.find_by(target_project_id: project.id, iid: iid)
+                         end
 
               next unless issuable
 
diff --git a/lib/gitlab/github_import/issuable_formatter.rb b/lib/gitlab/github_import/issuable_formatter.rb
index 256f360efc7..29fb0f9d333 100644
--- a/lib/gitlab/github_import/issuable_formatter.rb
+++ b/lib/gitlab/github_import/issuable_formatter.rb
@@ -1,6 +1,8 @@
 module Gitlab
   module GithubImport
     class IssuableFormatter < BaseFormatter
+      attr_writer :assignee_id, :author_id
+
       def project_association
         raise NotImplementedError
       end
@@ -23,18 +25,24 @@ module Gitlab
         raw_data.assignee.present?
       end
 
-      def assignee_id
+      def author
+        @author ||= UserFormatter.new(client, raw_data.user)
+      end
+
+      def author_id
+        @author_id ||= author.gitlab_id || project.creator_id
+      end
+
+      def assignee
         if assigned?
-          gitlab_user_id(raw_data.assignee.id)
+          @assignee ||= UserFormatter.new(client, raw_data.assignee)
         end
       end
 
-      def author
-        raw_data.user.login
-      end
+      def assignee_id
+        return @assignee_id if defined?(@assignee_id)
 
-      def author_id
-        gitlab_author_id || project.creator_id
+        @assignee_id = assignee.try(:gitlab_id)
       end
 
       def body
@@ -42,10 +50,10 @@ module Gitlab
       end
 
       def description
-        if gitlab_author_id
+        if author.gitlab_id
           body
         else
-          formatter.author_line(author) + body
+          formatter.author_line(author.login) + body
         end
       end
 
diff --git a/lib/gitlab/github_import/user_formatter.rb b/lib/gitlab/github_import/user_formatter.rb
new file mode 100644
index 00000000000..7550326ffbe
--- /dev/null
+++ b/lib/gitlab/github_import/user_formatter.rb
@@ -0,0 +1,52 @@
+module Gitlab
+  module GithubImport
+    class UserFormatter
+      attr_reader :client, :raw
+
+      delegate :id, :login, to: :raw, allow_nil: true
+
+      def initialize(client, raw)
+        @client = client
+        @raw = raw
+      end
+
+      def gitlab_id
+        return @gitlab_id if defined?(@gitlab_id)
+
+        @gitlab_id = find_by_external_uid || find_by_email
+      end
+
+      private
+
+      def email
+        @email ||= client.user(raw.login).try(:email)
+      end
+
+      def find_by_email
+        return nil unless email
+        users  = ::User.arel_table
+        emails = ::Email.arel_table
+
+        left_join_emails = users.join(emails, Arel::Nodes::OuterJoin).on(
+          users[:id].eq(emails[:user_id])
+        ).join_sources
+
+        User.select(:id)
+            .joins(left_join_emails)
+            .where(users[:email].eq(email).or(emails[:email].eq(email)))
+            .first.try(:id)
+      end
+
+      def find_by_external_uid
+        return nil unless id
+
+        identities = ::Identity.arel_table
+
+        User.select(:id)
+            .joins(:identities).where(identities[:provider].eq(:github)
+            .and(identities[:extern_uid].eq(id))).
+            first.try(:id)
+      end
+    end
+  end
+end