ehbrs_ruby_utils 0.35.0 → 0.36.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 763112ae8325274f9c7a277a78e6984bb2c3e89632214f89d3bc3bced8e1acde
-  data.tar.gz: 61448a174beb1981acaa3615a923a61fc39c9bd1c615839abef486a8c716b2a4
+  metadata.gz: a51803bc2fc98c7439c7e0e8808e00866d7130f898e1f9738c1c9f1483b326aa
+  data.tar.gz: a4146dc9718ed5356187a725e1537895c679301ca63d67619f9f3edaa8374223
 SHA512:
-  metadata.gz: 7a6948ad99b05f1f9742c1a79391ea3a8f7138420efaa6b6f4b966b86d3460f74e01027311170ba0b051a7ac66e27f7c08998344374bded000cb9ba0d93a5a31
-  data.tar.gz: 6515790504ba6e0c848f0c0941de5f3f5fd6f1b9bb396e356af85ea66957f857bfc3865b7b16258d2d5c54fffb55dac5439944884920e4e236766cba3407ff29
+  metadata.gz: 91760450578b621f60892d253f686ee1c5f8662c70a0177bace04b9f3cead3b4148999f0c8b7492f0de473be71e15e52e79d58957fd7fbcc04f59c939105e7ae
+  data.tar.gz: 2b9b4af1d0b6a47e99930adc678487721da804b6978d23700d5c33f2cdb04bf5757401c598c4ddc06212309c2bb3ac68bded943aaf4365a618bc169e4217ddac

data/lib/ehbrs_ruby_utils/fs/to_file_format.rb

@@ -0,0 +1,83 @@
+# frozen_string_literal: true
+
+require 'eac_fs/patches'
+require 'eac_ruby_utils/core_ext'
+
+module EhbrsRubyUtils
+  module Fs
+    class ToFileFormat
+      enable_abstract_methods
+      enable_simple_cache
+      abstract_methods :convert, :convert?, :target_encoding
+
+      common_constructor(:file) do
+        self.file = file.to_pathname
+      end
+
+      class << self
+        def convert_self(file)
+          new(file).run
+        end
+
+        def convert_to_file(source, target)
+          ::FileUtils.cp(source, target)
+          convert_self(target)
+        end
+
+        def convert_to_string(source)
+          ::EacRubyUtils::Fs::Temp.on_file do |target|
+            convert_to_file(source, target)
+            target.open('rb', &:read)
+          end
+        end
+      end
+
+      def run
+        return false unless convert?
+
+        convert
+        true
+      end
+
+      protected
+
+      def convert_to_target_encoding
+        ::EacRubyUtils::Fs::Temp.on_file do |temp|
+          ::EacRubyUtils::Envs.local.command(
+            'iconv', '-c', '-f', source_encoding, '-t', target_encoding, '-o', temp, file
+          ).execute!
+          ::FileUtils.mv(temp, file)
+        end
+        reset_cache
+      end
+
+      def crlf?
+        file_type?('CRLF')
+      end
+
+      def file_info_uncached
+        ::EacFs::FileInfo.new(file)
+      end
+
+      def file_type?(*include)
+        return false unless ::File.file?(file)
+
+        include.any? { |i| file_type.include?(i) }
+      end
+
+      def file_type_uncached
+        ::EacRubyUtils::Envs.local.command('file', '-b', file).execute!.strip
+      end
+
+      def source_encoding
+        r = file.info.charset
+        r = 'iso-8859-15' if r == 'unknown-8bit'
+        r
+      end
+
+      def text?
+        file.info.content_type.type == 'text'
+      end
+    end
+  end
+end

data/lib/ehbrs_ruby_utils/fs/to_utf8_unix.rb

@@ -0,0 +1,60 @@
+# frozen_string_literal: true
+
+require 'ehbrs_ruby_utils/fs/to_file_format'
+
+module EhbrsRubyUtils
+  module Fs
+    class ToUtf8Unix < ::EhbrsRubyUtils::Fs::ToFileFormat
+      UTF8_ENCODINGS = %w[us-ascii utf-8].freeze
+      ISO885915_ENCODINGS = %w[iso-8859-1].freeze
+
+      protected
+
+      def convert
+        check_utf8
+        check_crlf
+      end
+
+      def check_utf8
+        return if utf8?
+
+        convert_to_target_encoding
+        reset_cache
+      end
+
+      def file_attr(option)
+        ::EacRubyUtils::Envs.local.command('file', '--brief', option, file).execute!
+      end
+
+      def mime_encoding_uncached
+        file_attr('--mime-encoding')
+      end
+
+      def check_crlf
+        return unless crlf?
+
+        convert_crlf
+      end
+
+      def convert_crlf
+        ::EacRubyUtils::Envs.local.command('dos2unix', file).execute!
+      end
+
+      def convert?
+        text? && (!utf8? || crlf?)
+      end
+
+      def utf8?
+        UTF8_ENCODINGS.include?(mime_encoding)
+      end
+
+      def iso885915?
+        ISO885915_ENCODINGS.include?(mime_encoding)
+      end
+
+      def target_encoding
+        'utf-8'
+      end
+    end
+  end
+end

data/lib/ehbrs_ruby_utils/fs/to_windows_pt_br.rb

@@ -0,0 +1,59 @@
+# frozen_string_literal: true
+
+require 'eac_ruby_utils/core_ext'
+require 'ehbrs_ruby_utils/fs/to_file_format'
+
+module EhbrsRubyUtils
+  module Fs
+    class ToWindowsPtBr < ::EhbrsRubyUtils::Fs::ToFileFormat
+      TARGET_CHARSETS = %w[ISO-8859].freeze
+      ICONV_TO = 'ISO-8859-1'
+
+      protected
+
+      def convert
+        check_bom
+        check_target_charset
+        check_crlf
+      end
+
+      private
+
+      def check_bom
+        ::EacRubyUtils::Envs.local.command(
+          'sed', '-i', '1s/^\\xEF\\xBB\\xBF//', file
+        ).system!
+        reset_cache
+      end
+
+      def check_target_charset
+        return if target_charset?
+
+        convert_to_target_encoding
+      end
+
+      def check_crlf
+        return if crlf?
+
+        convert_crlf
+      end
+
+      def convert_crlf
+        ::EacRubyUtils::Envs.local.command('unix2dos', file).execute!
+        reset_cache
+      end
+
+      def convert?
+        text? && (!target_charset? || !crlf?)
+      end
+
+      def target_charset?
+        file_type?(*TARGET_CHARSETS)
+      end
+
+      def target_encoding
+        ICONV_TO
+      end
+    end
+  end
+end

data/lib/ehbrs_ruby_utils/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module EhbrsRubyUtils
-  VERSION = '0.35.0'
+  VERSION = '0.36.0'
 end

data/lib/ehbrs_ruby_utils/videos/subtitles/sanitize/content_sanitizer.rb

@@ -0,0 +1,87 @@
+# frozen_string_literal: true
+
+require 'eac_ruby_utils/core_ext'
+require 'ehbrs_ruby_utils/fs/to_file_format'
+require 'ehbrs_ruby_utils/videos/subtitles/sanitize/with_pattern_matcher'
+require 'ehbrs_ruby_utils/videos/subtitles/sanitize/with_term_matcher'
+require 'srt'
+
+module EhbrsRubyUtils
+  module Videos
+    module Subtitles
+      class Sanitize < ::EhbrsRubyUtils::Fs::ToFileFormat
+        class ContentSanitizer
+          class << self
+            def build_pattern(slim, elim)
+              /#{::Regexp.quote(slim)}[^#{::Regexp.quote(elim)}]*#{::Regexp.quote(elim)}/
+            end
+          end
+
+          REMOVE_PATTERNS = [%w[< >], %w[( )], ['[', ']']].map do |args|
+            build_pattern(*args)
+          end.freeze
+          REMOVE_TERMS = %w[subtitle osdb legenda @ united4ever unitedteam pt-subs capejuna maniacs
+                            |]
+                           .map(&:downcase)
+
+          common_constructor :input
+
+          def line_processors
+            REMOVE_PATTERNS.map do |pattern|
+              ::EhbrsRubyUtils::Videos::Subtitles::Sanitize::WithPatternMatcher.new(pattern)
+            end
+          end
+
+          def output
+            output_lines.join("\n")
+          end
+
+          def output_lines
+            r = []
+            last_output_line = nil
+            ::SRT::File.parse_string(input).lines.each do |input_line|
+              output_line(input_line, last_output_line).if_present do |v|
+                r << v
+                last_output_line = v
+              end
+            end
+            r
+          end
+
+          def output_line(input_line, last_output_line)
+            text = output_line_text(input_line.text)
+            return nil if text.blank?
+
+            r = input_line.dup
+            r.sequence = last_output_line.if_present(1) { |v| v.sequence + 1 }
+            r.text = text
+            r
+          end
+
+          def output_line_text(text)
+            text_processors.each do |term|
+              text = term.process(text)
+              return nil if text.blank?
+            end
+
+            text.map { |line| process_line(line) }.compact_blank
+          end
+
+          def process_line(line)
+            remove_tags(line)
+          end
+
+          def remove_tags(line)
+            line_processors.inject(line) { |a, e| e.process(a) }.strip
+          end
+
+          def text_processors
+            REMOVE_TERMS.map do |term|
+              ::EhbrsRubyUtils::Videos::Subtitles::Sanitize::WithTermMatcher.new(term)
+            end
+          end
+        end
+      end
+    end
+  end
+end

data/lib/ehbrs_ruby_utils/videos/subtitles/sanitize/with_pattern_matcher.rb

@@ -0,0 +1,20 @@
+# frozen_string_literal: true
+
+require 'eac_ruby_utils/core_ext'
+require 'ehbrs_ruby_utils/fs/to_file_format'
+
+module EhbrsRubyUtils
+  module Videos
+    module Subtitles
+      class Sanitize < ::EhbrsRubyUtils::Fs::ToFileFormat
+        class WithPatternMatcher
+          common_constructor :pattern
+
+          def process(line)
+            line.gsub(pattern, '')
+          end
+        end
+      end
+    end
+  end
+end

data/lib/ehbrs_ruby_utils/videos/subtitles/sanitize/with_term_matcher.rb

@@ -0,0 +1,20 @@
+# frozen_string_literal: true
+
+require 'eac_ruby_utils/core_ext'
+require 'ehbrs_ruby_utils/fs/to_file_format'
+
+module EhbrsRubyUtils
+  module Videos
+    module Subtitles
+      class Sanitize < ::EhbrsRubyUtils::Fs::ToFileFormat
+        class WithTermMatcher
+          common_constructor :term
+
+          def process(lines)
+            lines.map(&:downcase).any? { |line| line.include?(term) } ? nil : lines
+          end
+        end
+      end
+    end
+  end
+end

data/lib/ehbrs_ruby_utils/videos/subtitles/sanitize.rb

@@ -0,0 +1,42 @@
+# frozen_string_literal: true
+
+require 'avm/file_formats/utf8_assert'
+require 'ehbrs_ruby_utils/fs/to_file_format'
+require 'ehbrs_ruby_utils/fs/to_windows_pt_br'
+
+module EhbrsRubyUtils
+  module Videos
+    module Subtitles
+      class Sanitize < ::EhbrsRubyUtils::Fs::ToFileFormat
+        require_sub __FILE__
+
+        def run
+          sanitize_content
+          convert_to_windows_ptbr
+        end
+
+        def subtitle?
+          text? && file.extname == '.srt'
+        end
+
+        private
+
+        def convert_to_windows_ptbr
+          ::EhbrsRubyUtils::Fs::ToWindowsPtBr.convert_self(file)
+        end
+
+        def sanitize_content
+          ::Avm::FileFormats::Utf8Assert.assert_files([file]) do
+            sanitize_content_on_utf8
+          end
+        end
+
+        def sanitize_content_on_utf8
+          input = file.read
+          output = ::EhbrsRubyUtils::Videos::Subtitles::Sanitize::ContentSanitizer.new(input).output
+          file.write(output) if input != output
+        end
+      end
+    end
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: ehbrs_ruby_utils
 version: !ruby/object:Gem::Version
-  version: 0.35.0
+  version: 0.36.0
 platform: ruby
 authors:
 - Eduardo H. Bogoni
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2023-10-28 00:00:00.000000000 Z
+date: 2023-10-30 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: aranha
@@ -154,6 +154,26 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '0.5'
+- !ruby/object:Gem::Dependency
+  name: srt
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.1'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 0.1.5
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '0.1'
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 0.1.5
 - !ruby/object:Gem::Dependency
   name: taglib-ruby
   requirement: !ruby/object:Gem::Requirement
@@ -289,6 +309,9 @@ files:
 - lib/ehbrs_ruby_utils/fs/selected.rb
 - lib/ehbrs_ruby_utils/fs/selected/build.rb
 - lib/ehbrs_ruby_utils/fs/selected/build_file.rb
+- lib/ehbrs_ruby_utils/fs/to_file_format.rb
+- lib/ehbrs_ruby_utils/fs/to_utf8_unix.rb
+- lib/ehbrs_ruby_utils/fs/to_windows_pt_br.rb
 - lib/ehbrs_ruby_utils/gjt1.rb
 - lib/ehbrs_ruby_utils/gjt1/manager.rb
 - lib/ehbrs_ruby_utils/mudslide.rb
@@ -341,6 +364,10 @@ files:
 - lib/ehbrs_ruby_utils/videos/series/rename/results_builder.rb
 - lib/ehbrs_ruby_utils/videos/series/rename/season_group.rb
 - lib/ehbrs_ruby_utils/videos/stream.rb
+- lib/ehbrs_ruby_utils/videos/subtitles/sanitize.rb
+- lib/ehbrs_ruby_utils/videos/subtitles/sanitize/content_sanitizer.rb
+- lib/ehbrs_ruby_utils/videos/subtitles/sanitize/with_pattern_matcher.rb
+- lib/ehbrs_ruby_utils/videos/subtitles/sanitize/with_term_matcher.rb
 - lib/ehbrs_ruby_utils/web_utils.rb
 - lib/ehbrs_ruby_utils/web_utils/instance.rb
 - lib/ehbrs_ruby_utils/web_utils/instance/finances.rb