Merge branch '341825-integrate-rb-ipynbdiff-into-gitlab' into 'master'

Resolve "Integrate rb-ipynbdiff into Gitlab"

See merge request gitlab-org/gitlab!71477

Gemfile

@@ -537,3 +537,5 @@ gem 'webauthn', '~> 2.3'
 gem 'ipaddress', '~> 0.8.3'
 
 gem 'parslet', '~> 1.8'
+
+gem 'ipynbdiff', '0.3.6'

Gemfile.lock

@@ -640,6 +640,9 @@ GEM
     invisible_captcha (1.1.0)
       rails (>= 4.2)
     ipaddress (0.8.3)
+    ipynbdiff (0.3.6)
+      diffy (= 3.3.0)
+      json (= 2.5.1)
     jaeger-client (1.1.0)
       opentracing (~> 0.3)
       thrift
@@ -1501,6 +1504,7 @@ DEPENDENCIES
   icalendar
   invisible_captcha (~> 1.1.0)
   ipaddress (~> 0.8.3)
+  ipynbdiff (= 0.3.6)
   jira-ruby (~> 2.1.4)
   js_regex (~> 3.7)
   json (~> 2.5.1)

app/presenters/blob_presenter.rb

 # frozen_string_literal: true
+require 'ipynbdiff'
 
 class BlobPresenter < Gitlab::View::Presenter::Delegated
   include ApplicationHelper
@@ -20,6 +21,17 @@ class BlobPresenter < Gitlab::View::Presenter::Delegated
     )
   end
 
+  def highlight_transformed(plain: nil)
+    load_all_blob_data
+
+    Gitlab::Highlight.highlight(
+      blob.path,
+      transformed_blob_data,
+      language: language,
+      plain: plain
+    )
+  end
+
   def plain_data
     return if blob.binary?
 
@@ -107,4 +119,8 @@ class BlobPresenter < Gitlab::View::Presenter::Delegated
   def language
     blob.language_from_gitattributes
   end
+
+  def transformed_blob_data
+    @transformed_blob ||= ( blob.path.ends_with?('.ipynb') && IpynbDiff.transform(blob.data, options: { include_metadata: false, cell_decorator: :percent }) ) || blob.data
+  end
 end

config/feature_flags/development/jupyter_clean_diffs.yml

+---
+name: jupyter_clean_diffs
+introduced_by_url: https://gitlab.com/gitlab-org/gitlab/-/merge_requests/71477
+rollout_issue_url: https://gitlab.com/gitlab-org/gitlab/-/issues/343433
+milestone: '14.5'
+type: development
+group: group::incubation
+default_enabled: false

doc/user/project/repository/jupyter_notebooks/index.md

@@ -20,6 +20,15 @@ rendered to HTML when viewed:
 Interactive features, including JavaScript plots, don't work when viewed in
 GitLab.
 
+## Cleaner diffs 
+
+> [Introduced](https://gitlab.com/gitlab-org/gitlab/-/epics/6589) in GitLab 14.5
+
+When commits include changes to Jupyter Notebook files, GitLab strips out the 
+noise and displays a cleaner version of the diff.
+
+![Jupyter Notebook Clean Diff](img/jupyter_notebook_diff.png)
+
 ## Jupyter Git integration
 
 Jupyter can be configured as an OAuth application with repository access, acting

lib/gitlab/diff/file.rb

@@ -43,6 +43,8 @@ module Gitlab
         # Ensure items are collected in the the batch
         new_blob_lazy
         old_blob_lazy
+
+        preprocess_before_diff(diff) if Feature.enabled?(:jupyter_clean_diffs, @project)
       end
 
       def position(position_marker, position_type: :text)
@@ -448,6 +450,19 @@ module Gitlab
         find_renderable_viewer_class(classes)
       end
 
+      def preprocess_before_diff(diff)
+        return unless diff.new_path.ends_with? '.ipynb'
+
+        from = old_blob_lazy&.data
+        to = new_blob_lazy&.data
+
+        new_diff = IpynbDiff.diff(from, to,
+                                  diff_opts: { context: 5, include_diff_info: true },
+                                  transform_options: { cell_decorator: :percent } )
+
+        diff.diff = new_diff.scan(/.*\n/)[2..-1].join('') if new_diff
+      end
+
       def alternate_viewer_class
         return unless viewer.instance_of?(DiffViewer::Renamed)
 

lib/gitlab/diff/highlight.rb

@@ -152,6 +152,9 @@ module Gitlab
         return [] unless blob
 
         blob.load_all_data!
+
+        return blob.present.highlight_transformed.lines if Feature.enabled?(:jupyter_clean_diffs, @project)
+
         blob.present.highlight.lines
       end
 

spec/support/helpers/test_env.rb

@@ -53,7 +53,7 @@ module TestEnv
     'wip'                                => 'b9238ee',
     'csv'                                => '3dd0896',
     'v1.1.0'                             => 'b83d6e3',
-    'add-ipython-files'                  => 'f6b7a70',
+    'add-ipython-files'                  => 'c10c411',
     'add-pdf-file'                       => 'e774ebd',
     'squash-large-files'                 => '54cec52',
     'add-pdf-text-binary'                => '79faa7b',