[PATCH] player: support guessing encodings for comments

[Eric Wong <bofh@bogomips.org>]

Eric Wong wrote:
> Ugh, this is taking a while.  I have a mix of UTF-8 and
> ISO-8859-1 and probably some totally bogus filenames available to me :x

Maybe the following patch is alright, a few other things I want
to work on around mlib before I release.

---8<---
Subject: [PATCH] player: support guessing encodings for comments

This can be helpful for end users and is close to what other
players use.  We can fallback to Encoding.default_external by
default (typically UTF-8) and then again using `charlock_holmes'
if installed.

Note: path names remain binary, because that's how proper
filesystems operate.
---
 lib/dtas.rb            |  2 ++
 lib/dtas/encoding.rb   | 58 ++++++++++++++++++++++++++++++++++++++++++++++++++
 lib/dtas/source/sox.rb |  4 +++-
 test/test_encoding.rb  | 20 +++++++++++++++++
 4 files changed, 83 insertions(+), 1 deletion(-)
 create mode 100644 lib/dtas/encoding.rb
 create mode 100644 test/test_encoding.rb

diff --git a/lib/dtas.rb b/lib/dtas.rb
index ac416d7..3c2cdb4 100644
--- a/lib/dtas.rb
+++ b/lib/dtas.rb
@@ -42,3 +42,5 @@ def self.dedupe_str(str)
 
 require_relative 'dtas/compat_onenine'
 require_relative 'dtas/spawn_fix'
+require_relative 'dtas/encoding'
+DTAS.extend(DTAS::Encoding)
diff --git a/lib/dtas/encoding.rb b/lib/dtas/encoding.rb
new file mode 100644
index 0000000..71c877f
--- /dev/null
+++ b/lib/dtas/encoding.rb
@@ -0,0 +1,58 @@
+# Copyright (C) 2018 all contributors <dtas-all@nongnu.org>
+# License: GPL-3.0+ <https://www.gnu.org/licenses/gpl-3.0.txt>
+# frozen_string_literal: true
+
+# This module gets included in DTAS
+module DTAS::Encoding # :nodoc:
+  def self.extended(mod)
+    mod.instance_eval { @charlock_holmes = nil}
+  end
+
+private
+
+  def try_enc_harder(str, enc, old) # :nodoc:
+    case @charlock_holmes
+    when nil
+      begin
+        require 'charlock_holmes'
+        @charlock_holmes = CharlockHolmes::EncodingDetector.new
+      rescue LoadError
+        warn "`charlock_holmes` gem not available for encoding detection"
+        @charlock_holmes = false
+      end
+    when false
+      enc_fallback(str, enc, old)
+    else
+      res = @charlock_holmes.detect(str)
+      if det = res[:ruby_encoding]
+        str.force_encoding(det)
+        warn "charlock_holmes detected #{str.inspect} as #{det}..."
+        str.valid_encoding? or enc_fallback(str, det, old)
+      else
+        enc_fallback(str, enc, old)
+      end
+    end
+    str
+  end
+
+  def enc_fallback(str, enc, old) # :nodoc:
+    str.force_encoding(old)
+    warn "could not detect encoding for #{str.inspect} (not #{enc})"
+  end
+
+public
+
+  def try_enc(str, enc, harder = true) # :nodoc:
+    old = str.encoding
+    return str if old == enc
+    str.force_encoding(enc)
+    unless str.valid_encoding?
+      if harder
+        try_enc_harder(str, enc, old)
+      else
+        enc_fallback(str, enc, old)
+      end
+    end
+    str
+  end
+end
diff --git a/lib/dtas/source/sox.rb b/lib/dtas/source/sox.rb
index f702b41..03487fe 100644
--- a/lib/dtas/source/sox.rb
+++ b/lib/dtas/source/sox.rb
@@ -50,17 +50,19 @@ def mcache_lookup(infile)
       out =~ /^Sample Rate\s*:\s*(\d+)/n and dst['rate'] = $1.to_i
       out =~ /^Precision\s*:\s*(\d+)-bit/n and dst['bits'] = $1.to_i
 
+      enc = Encoding.default_external
       if out =~ /\nComments\s*:[ \t]*\n?(.*)\z/mn
         comments = dst['comments'] = {}
         key = nil
         $1.split(/\n/n).each do |line|
           if line.sub!(/^([^=]+)=/ni, '')
-            key = DTAS.dedupe_str($1.upcase)
+            key = DTAS.dedupe_str(DTAS.try_enc($1.upcase, enc))
           end
           (comments[key] ||= ''.b) << "#{line}\n" unless line.empty?
         end
         comments.each do |k,v|
           v.chomp!
+          DTAS.try_enc(v, enc)
           comments[k] = DTAS.dedupe_str(v)
         end
       end
diff --git a/test/test_encoding.rb b/test/test_encoding.rb
new file mode 100644
index 0000000..d9af968
--- /dev/null
+++ b/test/test_encoding.rb
@@ -0,0 +1,20 @@
+# Copyright (C) 2018 all contributors <dtas-all@nongnu.org>
+# License: GPL-3.0+ <https://www.gnu.org/licenses/gpl-3.0.txt>
+# frozen_string_literal: true
+require './test/helper'
+require 'dtas'
+require 'yaml'
+
+class TestEncoding < Testcase
+  def test_encoding
+    data = <<EOD # <20180111114546.77906b35@cumparsita.ch>
+---
+comments:
+  ARTIST: !binary |-
+    RW5yaXF1ZSBSb2Ryw61ndWV6
+EOD
+    hash = YAML.load(data)
+    artist = DTAS.try_enc(hash['comments']['ARTIST'], Encoding::UTF_8)
+    assert_equal 'Enrique Rodríguez', artist
+  end
+end
-- 
EW