active_encode 0.6.0 → 0.7.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 22be45555a3ea9208b56552c17e86a0052865e753cd2cd03fae9a876ebee6964
-  data.tar.gz: 55ba9a105b56e2d660b8bbe237cee26dc10caea24c87dd0b7e81a46b2ad28979
+  metadata.gz: 5aed6c18b1e8d3b2a0ecb4e8dd296660f2af65ca00973edad791b0315964da4c
+  data.tar.gz: 2ce90e1367403bc7d45ffb3d3ba5ada6991ca1f80d0a1b59944abeb63d47d372
 SHA512:
-  metadata.gz: b3ced87039d84e596fc2f011f70b586ef46d2501af00d94e64e91a2241df93c9e93b153814125454eb8af96ae63a2289188067fa4193fe01ee6568d5b9471113
-  data.tar.gz: 5dc21ccd48592fe2460599404f17c886f242f77a57cd3dbc9a8064176da44f9f1566a9cbf2c529491f037f25da252e0dd3fe40e345b15d9310b7c4464add310a
+  metadata.gz: a1edfa784257928dc9d06463434a4132a7ab07995a5679aca9051aa68335dd1ed644cd2c5f9544dc439911f29087d2af1b358c839ef2692a29a0478fae5e5ab2
+  data.tar.gz: f3702f56f989045165a70b29850fc7e480f02c85877912db3115b4a9909e39673090d79fe23127a0a1a30d5b0be96c15f6bba4ea04af751b388531b864786e33

data/.circleci/config.yml

@@ -1,6 +1,6 @@
 version: 2.1
 orbs:
-  samvera: samvera/circleci-orb@0
+  samvera: samvera/circleci-orb@0.3.2
 jobs:
   bundle_and_test:
     parameters:
@@ -12,35 +12,50 @@ jobs:
         default: 5.2.3
       bundler_version:
         type: string
-        default: 2.0.1
+        default: 2.0.2
+      ffmpeg_version:
+        type: string
+        default: 4.1.4
     executor:
       name: 'samvera/ruby'
       ruby_version: << parameters.ruby_version >>
     environment:
       ENGINE_CART_RAILS_OPTIONS: --skip-git --skip-bundle --skip-listen --skip-spring --skip-yarn --skip-keeps --skip-coffee --skip-puma --skip-test
       RAILS_VERSION: << parameters.rails_version >>
+      FFMPEG_PATH: /tmp/ffmpeg
     working_directory: ~/project
     steps:
-      - run: sudo apt-get install libmms0
-      - run: sudo wget -P /tmp/ https://mediaarea.net/download/binary/libzen0/0.4.37/libzen0v5_0.4.37-1_amd64.Debian_9.0.deb
-      - run: sudo wget -P /tmp/ https://mediaarea.net/download/binary/libmediainfo0/19.04/libmediainfo0v5_19.04-1_amd64.Debian_9.0.deb
-      - run: sudo wget -P /tmp/ https://mediaarea.net/download/binary/mediainfo/19.04/mediainfo_19.04-1_amd64.Debian_9.0.deb
-      - run: sudo dpkg -i /tmp/libzen0v5_0.4.37-1_amd64.Debian_9.0.deb /tmp/libmediainfo0v5_19.04-1_amd64.Debian_9.0.deb /tmp/mediainfo_19.04-1_amd64.Debian_9.0.deb
+      - run: 
+          command: |
+            sudo apt-get install libmms0
+            sudo wget -P /tmp/ https://mediaarea.net/download/binary/libzen0/0.4.37/libzen0v5_0.4.37-1_amd64.Debian_9.0.deb
+            sudo wget -P /tmp/ https://mediaarea.net/download/binary/libmediainfo0/19.04/libmediainfo0v5_19.04-1_amd64.Debian_9.0.deb
+            sudo wget -P /tmp/ https://mediaarea.net/download/binary/mediainfo/19.04/mediainfo_19.04-1_amd64.Debian_9.0.deb
+            sudo dpkg -i /tmp/libzen0v5_0.4.37-1_amd64.Debian_9.0.deb /tmp/libmediainfo0v5_19.04-1_amd64.Debian_9.0.deb /tmp/mediainfo_19.04-1_amd64.Debian_9.0.deb
+
+      - restore_cache:
+          keys:
+            - v1-ffmpeg-<< parameters.ffmpeg_version >>
+
+      - run: curl https://www.johnvansickle.com/ffmpeg/old-releases/ffmpeg-<< parameters.ffmpeg_version >>-amd64-static.tar.xz | tar xJ -C /tmp/ --strip-components=1
+
+      - save_cache:
+          key: v1-ffmpeg-<< parameters.ffmpeg_version >>`
+          paths:
+            - /tmp/ffmpeg
 
       - samvera/cached_checkout
 
-      - samvera/bundle_for_gem:
+      - samvera/bundle:
           ruby_version: << parameters.ruby_version >>
           bundler_version: << parameters.bundler_version >>
-          project: active_encode
 
       - samvera/engine_cart_generate:
           cache_key: v2-internal-test-app-{{ checksum "active_encode.gemspec" }}-{{ checksum "spec/test_app_templates/lib/generators/test_app_generator.rb" }}-<< parameters.rails_version >>-<< parameters.ruby_version >>
 
-      - samvera/bundle_for_gem:
+      - samvera/bundle:
           ruby_version: << parameters.ruby_version >>
           bundler_version: << parameters.bundler_version >>
-          project: active_encode
 
       - samvera/rubocop
 
@@ -55,9 +70,6 @@ workflows:
       - bundle_and_test:
           name: "ruby2-5_rails5-2"
           ruby_version: "2.5.5"
-      - bundle_and_test:
-          name: "ruby2-4_rails5-2"
-          ruby_version: "2.4.6"
       - bundle_and_test:
           name: "ruby2-6_rails5-1"
           ruby_version: "2.6.3"
@@ -66,7 +78,3 @@ workflows:
           name: "ruby2-5_rails5-1"
           ruby_version: "2.5.5"
           rails_version: "5.1.7"
-      - bundle_and_test:
-          name: "ruby2-4_rails5-1"
-          ruby_version: "2.4.6"
-          rails_version: "5.1.7"

data/.rubocop_todo.yml

@@ -40,6 +40,7 @@ Metrics/PerceivedComplexity:
 RSpec/AnyInstance:
   Exclude:
     - 'spec/integration/ffmpeg_adapter_spec.rb'
+    - 'spec/integration/pass_through_adapter_spec.rb'
 
 RSpec/ExampleLength:
   Enabled: false
@@ -47,6 +48,7 @@ RSpec/ExampleLength:
 RSpec/InstanceVariable:
   Exclude:
     - 'spec/integration/ffmpeg_adapter_spec.rb'
+    - 'spec/integration/pass_through_adapter_spec.rb'
 
 RSpec/MessageSpies:
   Exclude:
@@ -61,4 +63,5 @@ RSpec/VerifiedDoubles:
 
 Style/GuardClause:
   Exclude:
+    - 'lib/active_encode/engine_adapters/pass_through_adapter.rb'
     - 'lib/file_locator.rb'

data/active_encode.gemspec

@@ -33,4 +33,7 @@ Gem::Specification.new do |spec|
   spec.add_development_dependency "rspec-its"
   spec.add_development_dependency 'rspec_junit_formatter'
   spec.add_development_dependency "rspec-rails"
+
+  # Pin sprockets to < 4 so it works with ruby 2.5+
+  spec.add_dependency 'sprockets', '< 4'
 end

data/lib/active_encode/base.rb

@@ -1,6 +1,7 @@
 # frozen_string_literal: true
 require 'active_encode/core'
 require 'active_encode/engine_adapter'
+require 'active_encode/errors'
 require 'active_encode/status'
 require 'active_encode/technical_metadata'
 require 'active_encode/input'
@@ -18,6 +19,4 @@ module ActiveEncode #:nodoc:
     include Callbacks
     include GlobalID
   end
-
-  class NotFound < RuntimeError; end
 end

data/lib/active_encode/engine_adapters.rb

@@ -14,6 +14,7 @@ module ActiveEncode
     autoload :ElasticTranscoderAdapter
     autoload :TestAdapter
     autoload :FfmpegAdapter
+    autoload :PassThroughAdapter
 
     ADAPTER = 'Adapter'
     private_constant :ADAPTER

data/lib/active_encode/engine_adapters/elastic_transcoder_adapter.rb

@@ -121,7 +121,7 @@ module ActiveEncode
         end
 
         def convert_input(job)
-          job.input
+          job.input.key
         end
 
         def copy_to_input_bucket(input_url, bucket)
@@ -161,11 +161,12 @@ module ActiveEncode
         end
 
         def read_preset(id)
-          client.read_preset(id: id).preset
+          @presets ||= {}
+          @presets[id] ||= client.read_preset(id: id).preset
         end
 
         def convert_output(job)
-          pipeline = client.read_pipeline(id: job.pipeline_id).pipeline
+          @pipeline ||= client.read_pipeline(id: job.pipeline_id).pipeline
           job.outputs.collect do |joutput|
             preset = read_preset(joutput.preset_id)
             extension = preset.container == 'ts' ? '.m3u8' : ''
@@ -173,7 +174,7 @@ module ActiveEncode
               managed: false,
               id: joutput.id,
               label: joutput.key.split("/", 2).first,
-              url: "s3://#{pipeline.output_bucket}/#{job.output_key_prefix}#{joutput.key}#{extension}"
+              url: "s3://#{@pipeline.output_bucket}/#{job.output_key_prefix}#{joutput.key}#{extension}"
             }
             tech_md = convert_tech_metadata(joutput, preset).merge(additional_metadata)
 

data/lib/active_encode/engine_adapters/ffmpeg_adapter.rb

@@ -1,13 +1,27 @@
 # frozen_string_literal: true
 require 'fileutils'
 require 'nokogiri'
+require 'shellwords'
 
 module ActiveEncode
   module EngineAdapters
     class FfmpegAdapter
       WORK_DIR = ENV["ENCODE_WORK_DIR"] || "encodes" # Should read from config
+      MEDIAINFO_PATH = ENV["MEDIAINFO_PATH"] || "mediainfo"
+      FFMPEG_PATH = ENV["FFMPEG_PATH"] || "ffmpeg"
 
       def create(input_url, options = {})
+        # Decode file uris for ffmpeg (mediainfo works either way)
+        case input_url
+        when /^file\:\/\/\//
+          input_url = URI.decode(input_url)
+        when /^s3\:\/\//
+          require 'file_locator'
+
+          s3_object = FileLocator::S3File.new(input_url).object
+          input_url = URI.parse(s3_object.presigned_url(:get))
+        end
+
         new_encode = ActiveEncode::Base.new(input_url, options)
         new_encode.id = SecureRandom.uuid
         new_encode.created_at = Time.now.utc
@@ -20,7 +34,7 @@ module ActiveEncode
         FileUtils.mkdir_p working_path("outputs", new_encode.id)
 
         # Extract technical metadata from input file
-        `mediainfo --Output=XML --LogFile=#{working_path("input_metadata", new_encode.id)} #{input_url}`
+        `#{MEDIAINFO_PATH} --Output=XML --LogFile=#{working_path("input_metadata", new_encode.id)} "#{input_url}"`
         new_encode.input = build_input new_encode
 
         if new_encode.input.duration.blank?
@@ -43,14 +57,20 @@ module ActiveEncode
 
         # Run the ffmpeg command and save its pid
         command = ffmpeg_command(input_url, new_encode.id, options)
-        pid = Process.spawn(command)
+        pid = Process.spawn(command, err: working_path('error.log', new_encode.id))
         File.open(working_path("pid", new_encode.id), 'w') { |file| file.write pid }
         new_encode.input.id = pid
 
-        # Prevent zombie process
-        Process.detach(pid)
-
         new_encode
+      rescue StandardError => e
+        new_encode.state = :failed
+        new_encode.percent_complete = 1
+        new_encode.errors = [e.full_message]
+        write_errors new_encode
+        return new_encode
+      ensure
+        # Prevent zombie process
+        Process.detach(pid) if pid.present?
       end
 
       # Return encode object from file system
@@ -67,28 +87,21 @@ module ActiveEncode
         pid = get_pid(id)
         encode.input.id = pid if pid.present?
 
-        if File.file? working_path("error.log", id)
-          error = File.read working_path("error.log", id)
-          if error.present?
-            encode.state = :failed
-            encode.errors = [error]
-
-            return encode
-          end
-        end
-
-        encode.errors = []
-
         encode.current_operations = []
         encode.created_at, encode.updated_at = get_times encode.id
-
-        if running? pid
+        encode.errors = read_errors(id)
+        if encode.errors.present?
+          encode.state = :failed
+        elsif running? pid
           encode.state = :running
           encode.current_operations = ["transcoding"]
         elsif progress_ended?(encode.id) && encode.percent_complete == 100
           encode.state = :completed
-        elsif encode.percent_complete < 100
+        elsif cancelled? encode.id
           encode.state = :cancelled
+        elsif encode.percent_complete < 100
+          encode.errors << "Encoding has completed but the output duration is shorter than the input"
+          encode.state = :failed
         end
 
         encode.output = build_outputs encode if encode.completed?
@@ -98,10 +111,30 @@ module ActiveEncode
 
       # Cancel ongoing encode using pid file
       def cancel(id)
-        pid = get_pid(id)
-        Process.kill 'SIGTERM', pid.to_i
+        encode = find id
+        if encode.running?
+          pid = get_pid(id)
+
+          IO.popen("ps -ef | grep #{pid}") do |pipe|
+            child_pids = pipe.readlines.map do |line|
+              parts = line.split(/\s+/)
+              parts[1] if parts[2] == pid.to_s && parts[1] != pipe.pid.to_s
+            end.compact
+
+            child_pids.each do |cpid|
+              Process.kill 'SIGTERM', cpid.to_i
+            end
+          end
 
-        find id
+          Process.kill 'SIGTERM', pid.to_i
+          File.write(working_path("cancelled", id), "")
+          encode = find id
+        end
+        encode
+      rescue Errno::ESRCH
+        raise NotRunningError
+      rescue StandardError
+        raise CancelError
       end
 
     private
@@ -122,6 +155,16 @@ module ActiveEncode
         File.write(working_path("error.log", encode.id), encode.errors.join("\n"))
       end
 
+      def read_errors(id)
+        err_path = working_path("error.log", id)
+        error = File.read(err_path) if File.file? err_path
+        if error.present?
+          [error]
+        else
+          []
+        end
+      end
+
       def build_input(encode)
         input = ActiveEncode::Input.new
         metadata = get_tech_metadata(working_path("input_metadata", encode.id))
@@ -140,16 +183,15 @@ module ActiveEncode
         Dir["#{File.absolute_path(working_path('outputs', id))}/*"].each do |file_path|
           output = ActiveEncode::Output.new
           output.url = "file://#{file_path}"
-          original_extension = File.extname(encode.input.url)
-          original_filename = File.basename(encode.input.url).chomp(original_extension)
-          output.label = file_path[/#{Regexp.quote(original_filename)}\-(.*?)#{Regexp.quote(File.extname(file_path))}$/, 1]
+          sanitized_filename = sanitize_base encode.input.url
+          output.label = file_path[/#{Regexp.quote(sanitized_filename)}\-(.*?)#{Regexp.quote(File.extname(file_path))}$/, 1]
           output.id = "#{encode.input.id}-#{output.label}"
           output.created_at = encode.created_at
           output.updated_at = File.mtime file_path
 
           # Extract technical metadata from output file
           metadata_path = working_path("output_metadata-#{output.label}", id)
-          `mediainfo --Output=XML --LogFile=#{metadata_path} #{output.url}` unless File.file? metadata_path
+          `#{MEDIAINFO_PATH} --Output=XML --LogFile=#{metadata_path} #{output.url}` unless File.file? metadata_path
           output.assign_tech_metadata(get_tech_metadata(metadata_path))
 
           outputs << output
@@ -160,11 +202,19 @@ module ActiveEncode
 
       def ffmpeg_command(input_url, id, opts)
         output_opt = opts[:outputs].collect do |output|
-          file_name = "outputs/#{File.basename(input_url, File.extname(input_url))}-#{output[:label]}.#{output[:extension]}"
+          sanitized_filename = sanitize_base input_url
+          file_name = "outputs/#{sanitized_filename}-#{output[:label]}.#{output[:extension]}"
           " #{output[:ffmpeg_opt]} #{working_path(file_name, id)}"
         end.join(" ")
+        "#{FFMPEG_PATH} -y -loglevel error -progress #{working_path('progress', id)} -i \"#{input_url}\" #{output_opt}"
+      end
 
-        "ffmpeg -y -loglevel error -progress #{working_path('progress', id)} -i #{input_url} #{output_opt} > #{working_path('error.log', id)} 2>&1"
+      def sanitize_base(input_url)
+        if input_url.is_a? URI::HTTP
+          File.basename(input_url.path, File.extname(input_url.path))
+        else
+          File.basename(input_url, File.extname(input_url)).gsub(/[^0-9A-Za-z.\-]/, '_')
+        end
       end
 
       def get_pid(id)
@@ -188,10 +238,16 @@ module ActiveEncode
           1
         else
           progress_in_milliseconds = progress_value("out_time_ms=", data).to_i / 1000.0
-          (progress_in_milliseconds / encode.input.duration * 100).round
+          output = (progress_in_milliseconds / encode.input.duration * 100).ceil
+          return 100 if output > 100
+          output
         end
       end
 
+      def cancelled?(id)
+        File.exist? working_path("cancelled", id)
+      end
+
       def read_progress(id)
         File.read working_path("progress", id) if File.file? working_path("progress", id)
       end

data/lib/active_encode/engine_adapters/pass_through_adapter.rb

@@ -0,0 +1,239 @@
+# frozen_string_literal: true
+require 'fileutils'
+require 'nokogiri'
+require 'shellwords'
+require 'file_locator'
+
+# PassThroughAdapter accepts an input file url and a number of derivative urls in the options
+# E.g. `create(input, outputs: [{ label: 'low',  url: 'file:///derivatives/low.mp4' }, { label: 'high', url: 'file:///derivatives/high.mp4' }])`
+# This adapter mirrors the ffmpeg adapter but differs in a few ways:
+#    1. It starts by copying the derivative files to the work directory
+#    2. It runs Mediainfo on the input and output files and skips ffmpeg
+#    3. All work is done in the create method so it's status is always completed or failed
+module ActiveEncode
+  module EngineAdapters
+    class PassThroughAdapter
+      WORK_DIR = ENV["ENCODE_WORK_DIR"] || "encodes" # Should read from config
+      MEDIAINFO_PATH = ENV["MEDIAINFO_PATH"] || "mediainfo"
+
+      def create(input_url, options = {})
+        # Decode file uris for ffmpeg (mediainfo works either way)
+        input_url = URI.decode(input_url) if input_url.starts_with? "file:///"
+
+        new_encode = ActiveEncode::Base.new(input_url, options)
+        new_encode.id = SecureRandom.uuid
+        new_encode.current_operations = []
+        new_encode.output = []
+
+        # Create a working directory that holds all output files related to the encode
+        FileUtils.mkdir_p working_path("", new_encode.id)
+        FileUtils.mkdir_p working_path("outputs", new_encode.id)
+
+        # Extract technical metadata from input file
+        `#{MEDIAINFO_PATH} --Output=XML --LogFile=#{working_path("input_metadata", new_encode.id)} #{input_url.shellescape}`
+        new_encode.input = build_input new_encode
+        new_encode.input.id = new_encode.id
+        new_encode.created_at, new_encode.updated_at = get_times new_encode.id
+
+        if new_encode.input.duration.blank?
+          new_encode.state = :failed
+          new_encode.percent_complete = 1
+
+          new_encode.errors = if new_encode.input.file_size.blank?
+                                ["#{input_url} does not exist or is not accessible"]
+                              else
+                                ["Error inspecting input: #{input_url}"]
+                              end
+
+          write_errors new_encode
+          return new_encode
+        end
+
+        # For saving filename to label map used to find the label when building outputs
+        filename_label_hash = {}
+
+        # Copy derivatives to work directory
+        options[:outputs].each do |opt|
+          url = opt[:url]
+          output_path = working_path("outputs/#{sanitize_base opt[:url]}#{File.extname opt[:url]}", new_encode.id)
+          FileUtils.cp FileLocator.new(url).location, output_path
+          filename_label_hash[output_path] = opt[:label]
+        end
+
+        # Write filename-to-label map so we can retrieve them on build_output
+        File.write working_path("filename_label.yml", new_encode.id), filename_label_hash.to_yaml
+
+        new_encode.percent_complete = 1
+        new_encode.state = :running
+        new_encode.errors = []
+
+        new_encode
+      rescue StandardError => e
+        new_encode.state = :failed
+        new_encode.percent_complete = 1
+        new_encode.errors = [e.full_message]
+        write_errors new_encode
+        return new_encode
+      end
+
+      # Return encode object from file system
+      def find(id, opts = {})
+        encode_class = opts[:cast]
+        encode_class ||= ActiveEncode::Base
+        encode = encode_class.new(nil, opts)
+        encode.id = id
+        encode.created_at, encode.updated_at = get_times encode.id
+        encode.input = build_input encode
+        encode.input.id = encode.id
+        encode.output = []
+        encode.current_operations = []
+
+        encode.errors = read_errors(id)
+        if encode.errors.present?
+          encode.state = :failed
+          encode.percent_complete = 1
+        elsif cancelled?(id)
+          encode.state = :cancelled
+          encode.percent_complete = 1
+        elsif completed?(id)
+          encode.state = :completed
+          encode.percent_complete = 100
+        else
+          encode.output = build_outputs encode
+          encode.state = :completed
+          encode.percent_complete = 100
+        end
+
+        encode
+      rescue StandardError => e
+        encode.state = :failed
+        encode.percent_complete = 1
+        encode.errors = [e.full_message]
+        write_errors encode
+        return encode
+      end
+
+      # Cancel ongoing encode using pid file
+      def cancel(id)
+        # Check for errors and if not then create cancelled file else raise CancelError?
+        if running?(id)
+          File.write(working_path("cancelled", id), "")
+          find id
+        else
+          raise CancelError
+        end
+      end
+
+    private
+
+      def running?(id)
+        !cancelled?(id) || !failed?(id) || !completed?(id)
+      end
+
+      def cancelled?(id)
+        File.exist? working_path("cancelled", id)
+      end
+
+      def failed?(id)
+        read_errors(id).present?
+      end
+
+      def completed?(id)
+        File.exist? working_path("completed", id)
+      end
+
+      def get_times(id)
+        updated_at = if File.file? working_path("completed", id)
+                       File.mtime(working_path("completed", id))
+                     elsif File.file? working_path("cancelled", id)
+                       File.mtime(working_path("cancelled", id))
+                     elsif File.file? working_path("error.log", id)
+                       File.mtime(working_path("error.log", id))
+                     else
+                       File.mtime(working_path("input_metadata", id))
+                     end
+
+        [File.mtime(working_path("input_metadata", id)), updated_at]
+      end
+
+      def write_errors(encode)
+        File.write(working_path("error.log", encode.id), encode.errors.join("\n"))
+      end
+
+      def read_errors(id)
+        err_path = working_path("error.log", id)
+        error = File.read(err_path) if File.file? err_path
+        if error.present?
+          [error]
+        else
+          []
+        end
+      end
+
+      def build_input(encode)
+        input = ActiveEncode::Input.new
+        metadata = get_tech_metadata(working_path("input_metadata", encode.id))
+        input.url = metadata[:url]
+        input.assign_tech_metadata(metadata)
+        created_at = File.mtime(working_path("input_metadata", encode.id))
+        input.created_at = created_at
+        input.updated_at = created_at
+
+        input
+      end
+
+      def build_outputs(encode)
+        id = encode.id
+        outputs = []
+        filename_label_hash = YAML.safe_load(File.read(working_path("filename_label.yml", id))) if File.exist?(working_path("filename_label.yml", id))
+        Dir["#{File.absolute_path(working_path('outputs', id))}/*"].each do |file_path|
+          output = ActiveEncode::Output.new
+          output.url = "file://#{file_path}"
+          output.label = filename_label_hash[file_path] if filename_label_hash
+          output.id = "#{encode.input.id}-#{output.label}"
+          output.created_at = encode.created_at
+          output.updated_at = File.mtime file_path
+
+          # Extract technical metadata from output file
+          metadata_path = working_path("output_metadata-#{output.label}", id)
+          `#{MEDIAINFO_PATH} --Output=XML --LogFile=#{metadata_path} #{output.url}` unless File.file? metadata_path
+          output.assign_tech_metadata(get_tech_metadata(metadata_path))
+
+          outputs << output
+        end
+        File.write(working_path("completed", id), "")
+
+        outputs
+      end
+
+      def sanitize_base(input_url)
+        File.basename(input_url, File.extname(input_url)).gsub(/[^0-9A-Za-z.\-]/, '_')
+      end
+
+      def working_path(path, id)
+        File.join(WORK_DIR, id, path)
+      end
+
+      def get_tech_metadata(file_path)
+        doc = Nokogiri::XML File.read(file_path)
+        doc.remove_namespaces!
+        duration = get_xpath_text(doc, '//Duration/text()', :to_f)
+        duration *= 1000 unless duration.nil? # Convert to milliseconds
+        { url: get_xpath_text(doc, '//media/@ref', :to_s),
+          width: get_xpath_text(doc, '//Width/text()', :to_f),
+          height: get_xpath_text(doc, '//Height/text()', :to_f),
+          frame_rate: get_xpath_text(doc, '//FrameRate/text()', :to_f),
+          duration: duration,
+          file_size: get_xpath_text(doc, '//FileSize/text()', :to_i),
+          audio_codec: get_xpath_text(doc, '//track[@type="Audio"]/CodecID/text()', :to_s),
+          audio_bitrate: get_xpath_text(doc, '//track[@type="Audio"]/BitRate/text()', :to_i),
+          video_codec: get_xpath_text(doc, '//track[@type="Video"]/CodecID/text()', :to_s),
+          video_bitrate: get_xpath_text(doc, '//track[@type="Video"]/BitRate/text()', :to_i) }
+      end
+
+      def get_xpath_text(doc, xpath, cast_method)
+        doc.xpath(xpath).first&.text&.send(cast_method)
+      end
+    end
+  end
+end