make canvadocs_submissions entries unique

closes CNVS-26089

Test plan:
  * in a rails console, call
    DataFixup::CreateCanvadocsSubmissionsRecords.run
    a few times
  * run migrations
  * there should be no duplicate entries in canvadocs_submissions
  * you should not be able to create duplicate entries in
    canvadocs_submissions

Change-Id: I38cd3c990c13434d5daa016f3b4d192bdc877e16
Reviewed-on: https://gerrit.instructure.com/69327
Tested-by: Jenkins
Reviewed-by: Keith T. Garner <kgarner@instructure.com>
QA-Review: Amber Taniuchi <amber@instructure.com>
Product-Review: Cameron Matheson <cameron@instructure.com>

db/migrate/20151221185407_ensure_uniqueness_on_canvadocs_submissions.rb

@@ -0,0 +1,30 @@
+class EnsureUniquenessOnCanvadocsSubmissions < ActiveRecord::Migration
+  tag :postdeploy
+  disable_ddl_transaction!
+
+  def up
+    DataFixup::RemoveDuplicateCanvadocsSubmissions.run
+
+    remove_index :canvadocs_submissions, :canvadoc_id
+    remove_index :canvadocs_submissions, :crocodoc_document_id
+
+    add_index :canvadocs_submissions, [:submission_id, :canvadoc_id],
+      where: "canvadoc_id IS NOT NULL",
+      name: "unique_submissions_and_canvadocs",
+      unique: true, algorithm: :concurrently
+    add_index :canvadocs_submissions, [:submission_id, :crocodoc_document_id],
+      where: "crocodoc_document_id IS NOT NULL",
+      name: "unique_submissions_and_crocodocs",
+      unique: true, algorithm: :concurrently
+  end
+
+  def down
+    remove_index "canvadocs_submissions", name: "unique_submissions_and_canvadocs"
+    remove_index "canvadocs_submissions", name: "unique_submissions_and_crocodocs"
+
+    add_index :canvadocs_submissions, :canvadoc_id,
+      where: "canvadoc_id IS NOT NULL"
+    add_index :canvadocs_submissions, :crocodoc_document_id,
+      where: "crocodoc_document_id IS NOT NULL"
+  end
+end

lib/data_fixup/remove_duplicate_canvadocs_submissions.rb

@@ -0,0 +1,21 @@
+module DataFixup::RemoveDuplicateCanvadocsSubmissions
+  def self.run
+    %w[crocodoc_document_id canvadoc_id].each do |column|
+      duplicates = CanvadocsSubmission.
+        select("#{column}, submission_id").
+        group("#{column}, submission_id").
+        having("count(*) > 1")
+
+      duplicates.find_each do |dup|
+        scope = CanvadocsSubmission.where(
+          column => dup[column],
+          submission_id: dup.submission_id
+        )
+        keeper = scope.first
+        scope.where("id <> ?", keeper.id).delete_all
+      end
+    end
+  end
+
+  class CanvadocsSubmission < ActiveRecord::Base; end
+end