better wrapper script

3 files changed, 47 insertions, 66 deletions

diff --git a/util/de-tok.rb b/util/de-tok.rb
deleted file mode 100755
index 92c563f..0000000
--- a/util/de-tok.rb
+++ /dev/null
@@ -1,36 +0,0 @@
-#!/usr/bin/env ruby
-
-require 'nanomsg'
-require 'open3'
-require 'trollop'
-
-conf = Trollop::options do
-  opt :action, "tokenize (T) or detokenize (D)", :type => :string, :requred => true
-  opt :addr, "socket address", :short => "-S", :type => :string, :required => true
-  opt :scripts, "path to scripts directory", :short => "-p", :type => :string, :required => true
-  opt :lang, "language", :short => "-l", :type => :string, :required => true
-end
-
-sock = NanoMsg::PairSocket.new
-sock.bind conf[:addr]
-sock.send "hello"
-
-if conf[:action] == "D"
-  cmd = "#{conf[:scripts]}/detokenizer.perl -q -b -u -l #{conf[:lang]}"
-elsif conf[:action] == "T"
-  cmd = "#{conf[:scripts]}/tokenizer-no-escape.perl -q -b -a -l #{conf[:lang]}"
-else
-  # ERROR
-end
-while true
-  inp = sock.recv
-  break if !inp||inp=="shutdown"
-  Open3.popen3(cmd) do |pin, pout, perr|
-    pin.write inp
-    pin.close
-    sock.send pout.gets.strip
-  end
-end
-
-sock.send "off"
-
diff --git a/util/truecase.rb b/util/truecase.rb
deleted file mode 100755
index 3e97bd5..0000000
--- a/util/truecase.rb
+++ /dev/null
@@ -1,30 +0,0 @@
-#!/usr/bin/env ruby
-
-require 'nanomsg'
-require 'open3'
-require 'trollop'
-
-conf = Trollop::options do
-  opt :addr, "socket address", :short => "-S", :type => :string, :required => true
-  opt :moses, "path to moses directory", :short => "-m", :type => :string, :required => true
-  opt :model, "model file", :short => "-n", :type => :string, :required => true
-end
-
-sock = NanoMsg::PairSocket.new
-sock.bind conf[:addr]
-sock.send "hello"
-
-cmd = "#{conf[:moses]}/scripts/recaser/truecase.perl -b --model #{conf[:model]}"
-while true
-  inp = sock.recv + " " # FIXME?
-  break if !inp||inp=="shutdown"
-  Open3.popen3(cmd) do |pin, pout, perr|
-    pin.write inp
-    pin.close
-    s = pout.gets.strip
-    sock.send s #pout.gets.strip
-  end
-end
-
-sock.send "off"
-
diff --git a/util/wrapper.rb b/util/wrapper.rb
new file mode 100755
index 0000000..8445f0f
--- /dev/null
+++ b/util/wrapper.rb
@@ -0,0 +1,47 @@
+#!/usr/bin/env ruby
+
+require 'nanomsg'
+require 'open3'
+require 'trollop'
+
+conf = Trollop::options do
+  opt :action,         "tokenize,  detokenize or truecase", :short => "-a", :type => :string, :required => true
+  opt :addr,           "socket address",                    :short => "-S", :type => :string, :required => true
+  opt :ext,            "path to externals",                 :short => "-e", :type => :string, :required => true
+  opt :lang,           "language",                          :short => "-l", :type => :string
+  opt :truecase_model, "model file for truecaser",          :short => "-t", :type => :string
+end
+
+sock = NanoMsg::PairSocket.new
+sock.bind conf[:addr]
+sock.send "hello"
+
+if conf[:action] == "detokenize"
+  cmd = "#{conf[:ext]}/detokenizer.perl -q -b -u -l #{conf[:lang]}"
+  if !conf[:lang]
+    STDERR.write "[detokenizer] No language given, exiting!\n"; exit
+  end
+elsif conf[:action] == "tokenize"
+  cmd = "#{conf[:ext]}/tokenizer-no-escape.perl -q -b -a -l #{conf[:lang]}"
+  if !conf[:lang]
+    STDERR.write "[tokenizer] No language given, exiting!\n"; exit
+  end
+elsif conf[:action] == "truecase"
+  cmd = "#{conf[:ext]}/truecase.perl -b --model #{conf[:truecase_model]}"
+  if !conf[:truecase_model]
+    STDERR.write "[truecaser] No model given for truecaser, exiting!\n"; exit
+  end
+else
+  STDERR.write "[wrapper] Unknown action #{conf[:action]}, exiting!\n"; exit
+end
+pin, pout, perr = Open3.popen3(cmd)
+while true
+  inp = sock.recv.strip+"\n"
+  break if !inp||inp=="shutdown"
+  pin.write inp
+  sock.send pout.gets.strip
+end
+
+pin.close; pout.close; perr.close
+sock.send "off"
+