3453062d-fc36-4443-940d-ef7f6c0aac17

Committed 09 Sep 2025 06:51PM UTC coverage: 95.463% (-0.05%) from 95.512%

Build # 3453062d-fc36-4443-940d-ef7f6c0aac17

Build Type

push

circleci

Committed by

web-flow

Commit Message

Allow diacritics in file names (#2158)

Run Details

8 of 10 new or added lines in 3 files covered. (80.0%)

3535 of 3703 relevant lines covered (95.46%)

396.93 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

97.67

/app/models/upload_snapshot.rb

# frozen_string_literal: true
class UploadSnapshot < ApplicationRecord
  belongs_to :work
  attr_writer :upload

  alias_attribute :existing_files, :files

  def snapshot_deletions(work_changes, s3_filenames)
    s3_filenames_sorted = s3_filenames.sort
    existing_files.each do |file|
      filename = file["filename"]
      # Use Ruby's Binary Search functionality instead of a plain Ruby Array `.include?`
      # to detect missing values in the array because the binary search performs
      # much faster when the list of files is large. Notice that the binary search
      # requires that the list of files is sorted.
      # See https://ruby-doc.org/3.3.6/bsearch_rdoc.html
      if s3_filenames_sorted.bsearch { |s3_filename| filename <=> s3_filename }.nil?
        work_changes << { action: "removed", filename:, checksum: file["checksum"] }
      end
    end
  end

  def snapshot_modifications(work_changes, s3_files)
    # check for modifications
    s3_files.each do |s3_file|
      match = existing_files_sorted.bsearch { |file| s3_file.filename <=> file["filename"] }
      if match.nil?
        work_changes << { action: "added", filename: s3_file.filename, checksum: s3_file.checksum }
      elsif UploadSnapshot.checksum_compare(match["checksum"], s3_file.checksum) == false
        work_changes << { action: "replaced", filename: s3_file.filename, checksum: s3_file.checksum }
      end
    end
  end

  def upload
    @upload ||= uploads.find { |s3_file| filenames.include?(s3_file.filename) }
  end

  def uri
    URI.parse(url)
  end

  def filenames
    files.map { |file| file["filename"] }
  end

  def store_files(s3_files)
    self.files = s3_files.map { |file| { "filename" => file.filename, "checksum" => file.checksum } }
  end

  def self.find_by_filename(work_id:, filename:)
    find_by("work_id = ? AND files @> ?", work_id, JSON.dump([{ filename: }]))
  end

  class << self
    # Compares two checksums. Accounts for the case in which one of them is
    # a plain MD5 value and the other has been encoded with base64.
    # See also
    #   https://ruby-doc.org/core-2.7.0/Array.html#method-i-pack
    #   https://ruby-doc.org/core-2.7.0/String.html#method-i-unpack
    def checksum_compare(checksum1, checksum2)
      if checksum1 == checksum2
        true
      elsif checksum1.nil? || checksum2.nil?
        false
      elsif checksum1.length < checksum2.length
        # Decode the first one and then compare
        checksum1.unpack("m0").first.unpack("H*").first == checksum2
      else
        # Decode the second one and then compare
        checksum1 == checksum2.unpack("m0").first.unpack("H*").first
      end
    rescue ArgumentError
      # One of the values was not properly encoded
      false
    rescue NoMethodError
      # One of the values was not properly encoded
      false
    end
  end

  private

    def existing_files_sorted
      @existing_files_sorted ||= files.sort_by { |file| file["filename"] }
    end

    def uploads
      work.uploads
    end
end

1	# frozen_string_literal: true
2	class UploadSnapshot < ApplicationRecord	2✔
3	belongs_to :work	2✔
4	attr_writer :upload	2✔
5
6	alias_attribute :existing_files, :files	2✔
7
8	def snapshot_deletions(work_changes, s3_filenames)	2✔
9	s3_filenames_sorted = s3_filenames.sort	80✔
10	existing_files.each do \|file\|	80✔
11	filename = file["filename"]	78✔
12	# Use Ruby's Binary Search functionality instead of a plain Ruby Array `.include?`
13	# to detect missing values in the array because the binary search performs
14	# much faster when the list of files is large. Notice that the binary search
15	# requires that the list of files is sorted.
16	# See https://ruby-doc.org/3.3.6/bsearch_rdoc.html
17	if s3_filenames_sorted.bsearch { \|s3_filename\| filename <=> s3_filename }.nil?	184✔
18	work_changes << { action: "removed", filename:, checksum: file["checksum"] }	46✔
19	end
20	end
21	end
22
23	def snapshot_modifications(work_changes, s3_files)	2✔
24	# check for modifications
25	s3_files.each do \|s3_file\|	80✔
26	match = existing_files_sorted.bsearch { \|file\| s3_file.filename <=> file["filename"] }	232✔
27	if match.nil?	124✔
28	work_changes << { action: "added", filename: s3_file.filename, checksum: s3_file.checksum }	88✔
29	elsif UploadSnapshot.checksum_compare(match["checksum"], s3_file.checksum) == false	36✔
30	work_changes << { action: "replaced", filename: s3_file.filename, checksum: s3_file.checksum }	6✔
31	end
32	end
33	end
34
35	def upload	2✔
36	@upload \|\|= uploads.find { \|s3_file\| filenames.include?(s3_file.filename) }	12✔
37	end
38
39	def uri	2✔
40	URI.parse(url)	2✔
41	end
42
43	def filenames	2✔
44	files.map { \|file\| file["filename"] }	12✔
45	end
46
47	def store_files(s3_files)	2✔
48	self.files = s3_files.map { \|file\| { "filename" => file.filename, "checksum" => file.checksum } }	182✔
49	end
50
51	def self.find_by_filename(work_id:, filename:)	2✔
52	find_by("work_id = ? AND files @> ?", work_id, JSON.dump([{ filename: }]))	6✔
53	end
54
55	class << self	2✔
56	# Compares two checksums. Accounts for the case in which one of them is
57	# a plain MD5 value and the other has been encoded with base64.
58	# See also
59	# https://ruby-doc.org/core-2.7.0/Array.html#method-i-pack
60	# https://ruby-doc.org/core-2.7.0/String.html#method-i-unpack
61	def checksum_compare(checksum1, checksum2)	2✔
62	if checksum1 == checksum2	48✔
63	true	32✔
64	elsif checksum1.nil? \|\| checksum2.nil?	16✔
65	false	4✔
66	elsif checksum1.length < checksum2.length	12✔
67	# Decode the first one and then compare
68	checksum1.unpack("m0").first.unpack("H*").first == checksum2	6✔
69	else
70	# Decode the second one and then compare
71	checksum1 == checksum2.unpack("m0").first.unpack("H*").first	6✔
72	end
73	rescue ArgumentError
74	# One of the values was not properly encoded
75	false	4✔
76	rescue NoMethodError
77	# One of the values was not properly encoded
NEW 78	false	×
79	end
80	end
81
82	private	2✔
83
84	def existing_files_sorted	2✔
85	@existing_files_sorted \|\|= files.sort_by { \|file\| file["filename"] }	202✔
86	end
87
88	def uploads	2✔
89	work.uploads	4✔
90	end
91	end

pulibrary / pdc_describe / 3453062d-fc36-4443-940d-ef7f6c0aac17

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous