私が取り組んでいる Web サイトは再設計中であり、そのため、ユーザー画像のサイズを変更する必要があります。Web サイトは現在、carrierwave gem を使用してすべての画像およびビデオ処理を処理しており、各画像には、次の内容に基づいた一意のファイル名を持つ元のファイルがあります。
def filename
if original_filename
if model && model.read_attribute(:user_image).present?
model.read_attribute(:user_image)
else
@name ||= "#{secure_token}.#{file.extension}" if original_filename.present?
end
end
end
および secure_token によって生成される
def secure_token
var = :"@#{mounted_as}_secure_token"
model.instance_variable_get(var) or model.instance_variable_set(var, SecureRandom.uuid)
end
これを行うために作成されるタスクは次のとおりです。
##
# CarrierWave Amazon S3 File Reprocessor Rake Task
#
# Written (specifically) for:
# - CarrierWave
# - Ruby on Rails 3
# - Amazon S3
#
# Works with:
# - Any server of which you have write-permissions in the Rails.root/tmp directory
# - Works with Heroku
#
# Not tested with, but might work with:
# - Ruby on Rails 2
#
# Might work with, after a couple of tweaks:
# - File System Storage
# - Cloud Files Storage
# - GridFS
#
# Examples:
#
# Reprocess all versions of User#avatar
# rake carrierwave:reprocess class=User mounted_uploader=avatar
#
# Reprocess the versions: thumb, small, medium for User#avatar
# rake carrierwave:reprocess class=User mounted_uploader=avatar versions='thumb, small, medium'
#
# Reprocess for an underlying association, for things like Embedded MongoDB Documents
# which are models you cannot access directly, but rather through a regular Document
#
# Embeds One (picture) Association
# rake carrierwave:reprocess class=User association=picture mounted_uploader=image versions='thumb, small, medium'
#
# Embeds Many (pictures) Association
# rake carrierwave:reprocess class=User association=pictures mounted_uploader=image versions='thumb, small, medium'
#
# WARNING
# There is an issue with "Rake", that you cannot name your mounted_uploader "file".
# If you do this, then you will not be able to reprocess images through this rake task
# class User
# include Mongoid::Document
# mount_uploader :file, PictureUploader
# end
#
# This will NOT work with reprocessing through Rake because the mounted_uploader uses the "file" attribute.
namespace :carrierwave do
##
# Only tested with Amazon S3 Storage
# Needs some minor modifications if you want to use this for File System Store, Cloud Files and GridFS probably.
# This should work without Ruby on Rails as well, just set a different TMP_PATH.
desc "Reprocesses Carrier Wave file versions of a given model."
task :reprocess => :environment do
##
# Load in the OPEN URI library to be able
# to pull down and store the original file in a temp directory
require 'open-uri'
##
# Default constants
TMP_PATH = "#{Rails.root}/tmp/carrierwave"
##
# Set environment constants
CLASS = ENV['class'].capitalize
ASSOCIATION = ENV['association'] || nil
MOUNTED_UPLOADER = ENV['mounted_uploader'].to_sym
VERSIONS = ENV['versions'].nil? ? Array.new : ENV['versions'].split(',').map {|version| version.strip.to_sym}
##
# Find the Model
MODEL = Kernel.const_get(CLASS)
##
# Create the temp directory
%x(mkdir -p "#{TMP_PATH}")
##
# Find all records for the provided Model
records = MODEL.all
##
# Output to console
puts "\nCarrier Wave Version Reprocessing!"
puts "======================================="
puts "Model: #{CLASS}"
puts "Mounted Uploader: #{MOUNTED_UPLOADER}"
puts "Association: #{ASSOCIATION}" if ASSOCIATION
puts "Versions: #{VERSIONS.empty? ? "all" : VERSIONS.join(', ')}\n\n"
##
# Run through all records
records.each do |record|
##
# Set the mounted uploader object
# If it has a one-to-one association (singular) then that object
# will be returned and wrapped in an array so we can "iterate" through it below.
#
# If it has a one-to-many association then it will return the array of associated objects
#
# If no association is specified, it assumes the amounted uploader is attached to the specified CLASS
if ASSOCIATION
if ASSOCIATION.singular?
objects = [record.send(ASSOCIATION)]
else
objects = record.send(ASSOCIATION)
end
else
objects = [record]
end
##
# Iterates through the objects
objects.each do |object|
##
# Returns the mounted uploader object
mounted_object = object.send(MOUNTED_UPLOADER)
##
# Retrieve Filename
filename = mounted_object.path.split('/').last
##
# Output to console
puts "Reprocessing: #{filename}"
##
# Read out the original file from the remote location
# and write it out to the temp directory (TMP_PATH)
# This file will be used as the base file to reprocess
# the versions. Once all versions have been processed,
# this temp file will be directly removed.
open(mounted_object.url) do |original_object|
File.open(File.join(TMP_PATH, filename), 'w') do |temp_file|
temp_file.write(original_object.read)
end
end
##
# By default it will add all available versions to the versions variable
# which means that all available versions will be reprocessed.
# If the "versions" argument has been provided, then only the specified
# version(s) will be set to the versions variable, and thus, only these
# will be reprocessed.
versions = mounted_object.versions.map {|version| version[0]}
versions = VERSIONS unless VERSIONS.empty?
##
# Reprocesses the versions
versions.each do |version|
mounted_object.send(version).cache!(File.open(File.join(TMP_PATH, filename)))
mounted_object.send(version).store!
end
##
# Removes the temp file
%x(rm "#{TMP_PATH}/#{filename}")
end
end
end
end
問題は、新しい画像を作成している間、ファイルが画像アップローダーで設定された命名に従うのではなく、新しいファイル名で保存されるため、Web サイトがそれらを見つけられないことです。以下は、画像の保存方法の例です。
それがどうあるべきか：
元のファイル: fdk392ks93_39ei.png
サムネイル版：thumb_fdk392ks93_39ei.png
それはどうですか：
元のファイル: fdk392ks93_39ei.png
サムネイル版：thumb_fajeilkadifej_jakdjfi.png
どんな助けでも大歓迎です。
他の情報：
モデル: ユーザー
Uploader: user_image (これは、フォルダー/ファイル名を格納する列名でもあります)