web/translations/translate.pl

#!/usr/bin/perl
#
# Routino translation replacement Perl script
#
# Part of the Routino routing software.
#
# This file Copyright 2014 Andrew M. Bishop
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU Affero General Public License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.
#

use strict;

# Constants

my @translation_files=(<translation.*.txt>);
my $xml_output_file="../../xml/routino-translations.xml";
my $html_output_dir="../www/routino";

my @html_template_files=(<*.html>);

my %languages=();
my %translations=();


# Read in the translations

foreach my $translation_file (@translation_files)
  {
   $translation_file =~ m%translation.([^.]+).txt%;

   # Add to list of languages

   my $language=$1;

   if(! defined $languages{$language})
     {
      $languages{$language}=1;

      $translations{$language}={};
      $translations{$language}->{codes}={};
      $translations{$language}->{html}=0;
      $translations{$language}->{xml}=0;
     }

   # Process the file

   open(FILE,"<$translation_file");

   while(<FILE>)
     {
      s%\r*\n%%;

      next if(m%^#%);
      next if(m%^$%);

      # Single line HTML entries

      if(m%\@\@%)
        {
         my($code,$text)=split("\t");

         if(defined $translations{$language}->{codes}->{$code})
           {
            print STDERR "Language: $language DUPLICATED codeword '$code'\n";
           }
         else
           {
            $translations{$language}->{html}++;
            $translations{$language}->{codes}->{$code}={};
            $translations{$language}->{codes}->{$code}->{text}=$text;
            $translations{$language}->{codes}->{$code}->{used}=0;
           }
        }

      # Multi-line HTML entries

      if(m%(\$\$[^\$]+\$\$)%)
        {
         my($code,$text)=($1,"");

         while(<FILE>)
           {
            last if(m%\$\$%);

            $text=$text.$_;
           }

         $text =~ s%\r*\n$%%;

         if(defined $translations{$language}->{codes}->{$code})
           {
            print STDERR "Language: $language DUPLICATED codeword '$code'\n";
           }
         else
           {
            $translations{$language}->{html}++;
            $translations{$language}->{codes}->{$code}={};
            $translations{$language}->{codes}->{$code}->{text}=$text;
            $translations{$language}->{codes}->{$code}->{used}=0;
           }
        }

      # Single line XML entries

      if(m%\%\%%)
        {
         my($code,$text)=split("\t");

         if(defined $translations{$language}->{codes}->{$code})
           {
            print STDERR "Language: $language DUPLICATED codeword '$code'\n";
           }
         else
           {
            $translations{$language}->{xml}++;
            $translations{$language}->{codes}->{$code}={};
            $translations{$language}->{codes}->{$code}->{text}=$text;
            $translations{$language}->{codes}->{$code}->{used}=0;
           }
        }
     }

   close(FILE);
  }


# Sort out the languages

my @languages=();

push(@languages,"en");

foreach my $language (sort (keys %languages))
  {
   push(@languages,$language) if($language ne "en");
  }


# Create the HTML files

foreach my $html_template_file (@html_template_files)
  {
   foreach my $language (@languages)
     {
      next if(!$translations{$language}->{html});

      print "Language: $language File: $html_template_file\n";

      my $language_meta=0;
      my $language_meta_string="";

      open(HTML_IN ,"<$html_template_file");
      open(HTML_OUT,">$html_output_dir/$html_template_file.$language");

      while(<HTML_IN>)
        {
         my $line=$_;

         # Language selection - special handling

         if($line =~ m%\*\*LANGUAGES-META\*\*%)
           {
            $language_meta=1-$language_meta;

            if($language_meta==0)
              {
               foreach my $language2 (@languages)
                 {
                  my $LANGUAGE2=$language2;
                  $LANGUAGE2 =~ tr%a-z%A-Z%;

                  $line=$language_meta_string;

                  if($language eq $language2)
                    {
                     $line =~ s%~~CHECKED~~%checked%g;
                    }
                  else
                    {
                     $line =~ s%~~CHECKED~~%%g;
                    }

                  $line =~ s%~~lang~~%$language2%g;
                  $line =~ s%~~LANG~~%$LANGUAGE2%g;

                  if(!$translations{$language2}->{html})
                    {
                     $line =~ s%<a.+</a>%%;
                    }

                  if(!$translations{$language2}->{xml})
                    {
                     $line =~ s%<input .+>%%;
                    }

                  foreach my $code (keys $translations{$language2}->{codes})
                    {
                     if($line =~ s%$code%$translations{$language2}->{codes}->{$code}->{text}%g)
                       {$translations{$language2}->{codes}->{$code}->{used} = 1;}
                    }

                  if($line =~ m%((\@\@|\$\$|\*\*|\~\~)[^\@\$*~]+(\@\@|\$\$|\*\*|\~\~))%)
                    {
                     print STDERR "   Unmatched codeword '$1' in line: $line";
                    }

                  # Remove un-needed spaces

                  $line =~ s%[\t ]+% %g;
                  $line =~ s%\n %\n%g;
                  $line =~ s%^ %%g;

                  print HTML_OUT $line;
                 }
              }

            next;
           }

         if($language_meta)
           {
            $language_meta_string.=$line;
            next;
           }

         # Replace with translated phrases

         foreach my $code (keys $translations{$language}->{codes})
           {
            if($line =~ s%\Q$code\E%$translations{$language}->{codes}->{$code}->{text}%g)
              {$translations{$language}->{codes}->{$code}->{used} = 1;}
           }

         # Replace what is left with English phrases

         foreach my $code (keys $translations{$languages[0]}->{codes})
           {
            $line =~ s%\Q$code\E%$translations{$languages[0]}->{codes}->{$code}->{text}%g;
           }

         if($line =~ m%((\@\@|\$\$|\*\*|\~\~)[^\@\$*~]+(\@\@|\$\$|\*\*|\~\~))%)
           {
            print STDERR "   Unmatched codeword '$1' in line: $line";
           }

         # Remove un-needed spaces

         $line =~ s%[\t ]+% %g;
         $line =~ s%\n %\n%g;
         $line =~ s%^ %%g;

         print HTML_OUT $line;
        }

      close(HTML_IN);
      close(HTML_OUT);
     }
  }


# Create the XML file

open(XML_OUT,">$xml_output_file");

open(XML_IN ,"<translations-head.xml");

while(<XML_IN>)
  {
   print XML_OUT;
  }

close(XML_IN);

foreach my $language (@languages)
  {
   next if(!$translations{$language}->{xml});

   print "Language: $language File: translations.xml\n";

   open(XML_IN ,"<translations-body.xml");

   while(<XML_IN>)
     {
      my $line=$_;

      $line =~ s%~~lang~~%$language%g;

      # Replace with translated phrases

      foreach my $code (keys $translations{$language}->{codes})
        {
         if($line =~ s%$code%$translations{$language}->{codes}->{$code}->{text}%g)
           {$translations{$language}->{codes}->{$code}->{used} = 1;}
        }

      # Replace what is left with a note about missing translations

      if($line =~ m%\%\%%)
        {
         foreach my $code (keys $translations{$languages[0]}->{codes})
           {
            $line =~ s%$code%$translations{$languages[0]}->{codes}->{$code}->{text}%g;
           }

         $line =~ s%<%<!-- TRANSLATION REQUIRED: %;
         $line =~ s%>% -->%;

         if($line =~ m%((\%\%|\~\~)[^\%~]+(\%\%|\~\~))%)
           {
            print STDERR "   Unmatched codeword '$1' in line: $line";
           }
        }

      print XML_OUT $line;
     }

   close(XML_IN);
  }

open(XML_IN ,"<translations-tail.xml");

while(<XML_IN>)
  {
   print XML_OUT;
  }

close(XML_IN);

close(XML_OUT);


# Check the languages

foreach my $language (@languages)
  {
   foreach my $code (keys $translations{$language}->{codes})
     {
      if(! $translations{$language}->{codes}->{$code}->{used})
        {
         print STDERR "Language: $language UNUSED codeword: $code\n";
        }
     }
  }


# Print the translation coverage

print "\n";

print "Translation Coverage\n";
print "====================\n";

print "\n";
print "            Number      Fraction\n";
print "Language  HTML   XML   HTML   XML\n";
print "--------  ----   ---   ----   ---\n";

foreach my $language (@languages)
  {
   printf("%-6s     %3d   %3d  %4.0f%% %4.0f%%\n",$language,
                                                  $translations{$language}->{html},
                                                  $translations{$language}->{xml},
                                                  100.0*$translations{$language}->{html}/$translations{$languages[0]}->{html},
                                                  100.0*$translations{$language}->{xml} /$translations{$languages[0]}->{xml});
  }
1	#!/usr/bin/perl
2	#
3	# Routino translation replacement Perl script
4	#
5	# Part of the Routino routing software.
6	#
7	# This file Copyright 2014 Andrew M. Bishop
8	#
9	# This program is free software: you can redistribute it and/or modify
10	# it under the terms of the GNU Affero General Public License as published by
11	# the Free Software Foundation, either version 3 of the License, or
12	# (at your option) any later version.
13	#
14	# This program is distributed in the hope that it will be useful,
15	# but WITHOUT ANY WARRANTY; without even the implied warranty of
16	# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17	# GNU Affero General Public License for more details.
18	#
19	# You should have received a copy of the GNU Affero General Public License
20	# along with this program. If not, see <http://www.gnu.org/licenses/>.
21	#
22
23	use strict;
24
25	# Constants
26
27	my @translation_files=(<translation.*.txt>);
28	my $xml_output_file="../../xml/routino-translations.xml";
29	my $html_output_dir="../www/routino";
30
31	my @html_template_files=(<*.html>);
32
33	my %languages=();
34	my %translations=();
35
36
37	# Read in the translations
38
39	foreach my $translation_file (@translation_files)
40	{
41	$translation_file =~ m%translation.([^.]+).txt%;
42
43	# Add to list of languages
44
45	my $language=$1;
46
47	if(! defined $languages{$language})
48	{
49	$languages{$language}=1;
50
51	$translations{$language}={};
52	$translations{$language}->{codes}={};
53	$translations{$language}->{html}=0;
54	$translations{$language}->{xml}=0;
55	}
56
57	# Process the file
58
59	open(FILE,"<$translation_file");
60
61	while(<FILE>)
62	{
63	s%\r*\n%%;
64
65	next if(m%^#%);
66	next if(m%^$%);
67
68	# Single line HTML entries
69
70	if(m%\@\@%)
71	{
72	my($code,$text)=split("\t");
73
74	if(defined $translations{$language}->{codes}->{$code})
75	{
76	print STDERR "Language: $language DUPLICATED codeword '$code'\n";
77	}
78	else
79	{
80	$translations{$language}->{html}++;
81	$translations{$language}->{codes}->{$code}={};
82	$translations{$language}->{codes}->{$code}->{text}=$text;
83	$translations{$language}->{codes}->{$code}->{used}=0;
84	}
85	}
86
87	# Multi-line HTML entries
88
89	if(m%(\$\$[^\$]+\$\$)%)
90	{
91	my($code,$text)=($1,"");
92
93	while(<FILE>)
94	{
95	last if(m%\$\$%);
96
97	$text=$text.$_;
98	}
99
100	$text =~ s%\r*\n$%%;
101
102	if(defined $translations{$language}->{codes}->{$code})
103	{
104	print STDERR "Language: $language DUPLICATED codeword '$code'\n";
105	}
106	else
107	{
108	$translations{$language}->{html}++;
109	$translations{$language}->{codes}->{$code}={};
110	$translations{$language}->{codes}->{$code}->{text}=$text;
111	$translations{$language}->{codes}->{$code}->{used}=0;
112	}
113	}
114
115	# Single line XML entries
116
117	if(m%\%\%%)
118	{
119	my($code,$text)=split("\t");
120
121	if(defined $translations{$language}->{codes}->{$code})
122	{
123	print STDERR "Language: $language DUPLICATED codeword '$code'\n";
124	}
125	else
126	{
127	$translations{$language}->{xml}++;
128	$translations{$language}->{codes}->{$code}={};
129	$translations{$language}->{codes}->{$code}->{text}=$text;
130	$translations{$language}->{codes}->{$code}->{used}=0;
131	}
132	}
133	}
134
135	close(FILE);
136	}
137
138
139	# Sort out the languages
140
141	my @languages=();
142
143	push(@languages,"en");
144
145	foreach my $language (sort (keys %languages))
146	{
147	push(@languages,$language) if($language ne "en");
148	}
149
150
151	# Create the HTML files
152
153	foreach my $html_template_file (@html_template_files)
154	{
155	foreach my $language (@languages)
156	{
157	next if(!$translations{$language}->{html});
158
159	print "Language: $language File: $html_template_file\n";
160
161	my $language_meta=0;
162	my $language_meta_string="";
163
164	open(HTML_IN ,"<$html_template_file");
165	open(HTML_OUT,">$html_output_dir/$html_template_file.$language");
166
167	while(<HTML_IN>)
168	{
169	my $line=$_;
170
171	# Language selection - special handling
172
173	if($line =~ m%\\LANGUAGES-META\\%)
174	{
175	$language_meta=1-$language_meta;
176
177	if($language_meta==0)
178	{
179	foreach my $language2 (@languages)
180	{
181	my $LANGUAGE2=$language2;
182	$LANGUAGE2 =~ tr%a-z%A-Z%;
183
184	$line=$language_meta_string;
185
186	if($language eq $language2)
187	{
188	$line =~ s%~~CHECKED~~%checked%g;
189	}
190	else
191	{
192	$line =~ s%~~CHECKED~~%%g;
193	}
194
195	$line =~ s%~~lang~~%$language2%g;
196	$line =~ s%~~LANG~~%$LANGUAGE2%g;
197
198	if(!$translations{$language2}->{html})
199	{
200	$line =~ s%<a.+</a>%%;
201	}
202
203	if(!$translations{$language2}->{xml})
204	{
205	$line =~ s%<input .+>%%;
206	}
207
208	foreach my $code (keys $translations{$language2}->{codes})
209	{
210	if($line =~ s%$code%$translations{$language2}->{codes}->{$code}->{text}%g)
211	{$translations{$language2}->{codes}->{$code}->{used} = 1;}
212	}
213
214	if($line =~ m%((\@\@\|\$\$\|\\\|\~\~)[^\@\$~]+(\@\@\|\$\$\|\\*\|\~\~))%)
215	{
216	print STDERR " Unmatched codeword '$1' in line: $line";
217	}
218
219	# Remove un-needed spaces
220
221	$line =~ s%[\t ]+% %g;
222	$line =~ s%\n %\n%g;
223	$line =~ s%^ %%g;
224
225	print HTML_OUT $line;
226	}
227	}
228
229	next;
230	}
231
232	if($language_meta)
233	{
234	$language_meta_string.=$line;
235	next;
236	}
237
238	# Replace with translated phrases
239
240	foreach my $code (keys $translations{$language}->{codes})
241	{
242	if($line =~ s%\Q$code\E%$translations{$language}->{codes}->{$code}->{text}%g)
243	{$translations{$language}->{codes}->{$code}->{used} = 1;}
244	}
245
246	# Replace what is left with English phrases
247
248	foreach my $code (keys $translations{$languages[0]}->{codes})
249	{
250	$line =~ s%\Q$code\E%$translations{$languages[0]}->{codes}->{$code}->{text}%g;
251	}
252
253	if($line =~ m%((\@\@\|\$\$\|\\\|\~\~)[^\@\$~]+(\@\@\|\$\$\|\\*\|\~\~))%)
254	{
255	print STDERR " Unmatched codeword '$1' in line: $line";
256	}
257
258	# Remove un-needed spaces
259
260	$line =~ s%[\t ]+% %g;
261	$line =~ s%\n %\n%g;
262	$line =~ s%^ %%g;
263
264	print HTML_OUT $line;
265	}
266
267	close(HTML_IN);
268	close(HTML_OUT);
269	}
270	}
271
272
273	# Create the XML file
274
275	open(XML_OUT,">$xml_output_file");
276
277	open(XML_IN ,"<translations-head.xml");
278
279	while(<XML_IN>)
280	{
281	print XML_OUT;
282	}
283
284	close(XML_IN);
285
286	foreach my $language (@languages)
287	{
288	next if(!$translations{$language}->{xml});
289
290	print "Language: $language File: translations.xml\n";
291
292	open(XML_IN ,"<translations-body.xml");
293
294	while(<XML_IN>)
295	{
296	my $line=$_;
297
298	$line =~ s%~~lang~~%$language%g;
299
300	# Replace with translated phrases
301
302	foreach my $code (keys $translations{$language}->{codes})
303	{
304	if($line =~ s%$code%$translations{$language}->{codes}->{$code}->{text}%g)
305	{$translations{$language}->{codes}->{$code}->{used} = 1;}
306	}
307
308	# Replace what is left with a note about missing translations
309
310	if($line =~ m%\%\%%)
311	{
312	foreach my $code (keys $translations{$languages[0]}->{codes})
313	{
314	$line =~ s%$code%$translations{$languages[0]}->{codes}->{$code}->{text}%g;
315	}
316
317	$line =~ s%<%<!-- TRANSLATION REQUIRED: %;
318	$line =~ s%>% -->%;
319
320	if($line =~ m%((\%\%\|\~\~)[^\%~]+(\%\%\|\~\~))%)
321	{
322	print STDERR " Unmatched codeword '$1' in line: $line";
323	}
324	}
325
326	print XML_OUT $line;
327	}
328
329	close(XML_IN);
330	}
331
332	open(XML_IN ,"<translations-tail.xml");
333
334	while(<XML_IN>)
335	{
336	print XML_OUT;
337	}
338
339	close(XML_IN);
340
341	close(XML_OUT);
342
343
344	# Check the languages
345
346	foreach my $language (@languages)
347	{
348	foreach my $code (keys $translations{$language}->{codes})
349	{
350	if(! $translations{$language}->{codes}->{$code}->{used})
351	{
352	print STDERR "Language: $language UNUSED codeword: $code\n";
353	}
354	}
355	}
356
357
358	# Print the translation coverage
359
360	print "\n";
361
362	print "Translation Coverage\n";
363	print "====================\n";
364
365	print "\n";
366	print " Number Fraction\n";
367	print "Language HTML XML HTML XML\n";
368	print "-------- ---- --- ---- ---\n";
369
370	foreach my $language (@languages)
371	{
372	printf("%-6s %3d %3d %4.0f%% %4.0f%%\n",$language,
373	$translations{$language}->{html},
374	$translations{$language}->{xml},
375	100.0*$translations{$language}->{html}/$translations{$languages[0]}->{html},
376	100.0*$translations{$language}->{xml} /$translations{$languages[0]}->{xml});
377	}