1 |
/* |
2 |
gen_latex.h - Part of psiconv, a PSION 5 file formats converter |
3 |
Copyright (c) 2001 Jim Ottaway <j.ottaway@lse.ac.uk> |
4 |
|
5 |
This program is free software; you can redistribute it and/or modify |
6 |
it under the terms of the GNU General Public License as published by |
7 |
the Free Software Foundation; either version 2 of the License, or |
8 |
(at your option) any later version. |
9 |
|
10 |
This program is distributed in the hope that it will be useful, |
11 |
but WITHOUT ANY WARRANTY; without even the implied warranty of |
12 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
13 |
GNU General Public License for more details. |
14 |
|
15 |
You should have received a copy of the GNU General Public License |
16 |
along with this program; if not, write to the Free Software |
17 |
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. |
18 |
*/ |
19 |
|
20 |
/* gen_latex.c |
21 |
|
22 |
Sun Feb 11 12:16:29 2001 |
23 |
|
24 |
Jim Ottaway <j.ottaway@lse.ac.uk> |
25 |
|
26 |
Conversion to LaTeX: |
27 |
|
28 |
* Converts to the article class |
29 |
|
30 |
* If the file is a Word file and has outline levels (i.e. is not a |
31 |
Psion 5 file), outline levels 1 to 5 are converted to sectioning |
32 |
commands, otherwise headings are just formatted |
33 |
|
34 |
* If there is a style with the name 'quotation', the paragraph is |
35 |
converted to a quotation environment (other styles/environments |
36 |
could be added) |
37 |
|
38 |
* Also does formatting commands (italics -> \emph, bold -> \textbf, |
39 |
underline -> \underline), and character translation |
40 |
|
41 |
*/ |
42 |
|
43 |
#include "config.h" |
44 |
#include <stdio.h> |
45 |
#include <string.h> |
46 |
#include <stdlib.h> |
47 |
#include "psiconv/data.h" |
48 |
#include "psiconv/list.h" |
49 |
#include "gen.h" |
50 |
#include "psiconv.h" |
51 |
|
52 |
#ifdef DMALLOC |
53 |
#include "dmalloc.h" |
54 |
#endif |
55 |
|
56 |
/* This is incomplete at the moment. |
57 |
|
58 |
Most of the translation/faking of chars is borrowed from the |
59 |
HTML::Latex.pm perl module. */ |
60 |
static const char *char_table[0x100] = |
61 |
{ |
62 |
/* 0x00 */ "" ,"" ,"" ,"" ,"" ,"" ,"\n\n","\\\\" |
63 |
, |
64 |
/* 0x08 */ "\n\n" ," " ,"" ,"" ,"" ,"" ,"" ,"" |
65 |
, |
66 |
/* 0x10 */ " " ,"" ,"" ,"" ,"" ,"" ,"" ,"" |
67 |
, |
68 |
/* 0x18 */ "" ,"" ,"" ,"" ,"" ,"" ,"" ,"" |
69 |
, |
70 |
/* 0x20 */ " " ,"!" ,"\"" ,"\\#" ,"\\$" ,"\\%" ,"\\&","'" |
71 |
, |
72 |
/* 0x28 */ "(" ,")" ,"*" ,"+" ,"," ,"-" ,"." ,"/" |
73 |
, |
74 |
/* 0x30 */ "0" ,"1" ,"2" ,"3" ,"4" ,"5" ,"6" ,"7" |
75 |
, |
76 |
/* 0x38 */ "8" ,"9" ,":" ,";" ,"$<$" ,"=" ,"$>$" ,"?" , |
77 |
/* 0x40 */ "@" ,"A" ,"B" ,"C" ,"D" ,"E" ,"F" ,"G" |
78 |
, |
79 |
/* 0x48 */ "H" ,"I" ,"J" ,"K" ,"L" ,"M" ,"N" ,"O" |
80 |
, |
81 |
/* 0x50 */ "P" ,"Q" ,"R" ,"S" ,"T" ,"U" ,"V" ,"W" |
82 |
, |
83 |
/* 0x58 */ "X" ,"Y" ,"Z" ,"[" ,"$\\backslash$" ,"]" ,"\\^{}" |
84 |
,"\\_" , |
85 |
/* 0x60 */ "`" ,"a" ,"b" ,"c" ,"d" ,"e" ,"f" ,"g" |
86 |
, |
87 |
/* 0x68 */ "h" ,"i" ,"j" ,"k" ,"l" ,"m" ,"n" ,"o" |
88 |
, |
89 |
/* 0x70 */ "p" ,"q" ,"r" ,"s" ,"t" ,"u" ,"v" ,"w" |
90 |
, |
91 |
/* 0x78 */ "x" ,"y" ,"z" ,"\\{" ,"$|$" ,"\\}" ,"~" |
92 |
,"" , |
93 |
/* 0x80 */ "" ,"","","$f$","","{\\ldots}", "$\\dagger$","$\\ddagger$", |
94 |
/* 0x88 */ "\\^{}","","","","\\OE","" ,"" ,"" , |
95 |
/* 0x90 */ "","`","'","``","''", "$\\bullet$","--","---", |
96 |
/* 0x98 */ "~","{\\textsc{tm}}","","","\\oe","","","\\\"Y", |
97 |
/* 0xa0 */ "","!`","c","{\\pounds}", |
98 |
"","{Y\\hspace*{-1.4ex}--}","$|$","{\\S}", |
99 |
/* 0xa8 */ |
100 |
"\"","{\\copyright}","$^{\\underline{a}}$","","$\\neg$","$-$","","$^-$", |
101 |
/* 0xb0 */ "$^{\\circ}$","$\\pm$","$^2$","$^3$", |
102 |
"$^\\prime$","$\\mu$","{\\P}","$\\cdot$", |
103 |
/* 0xb8 */ ",","$^1$","$^{\\underline{\\circ}}$","", |
104 |
"$\\frac{1}{4}$","$\\frac{1}{2}$","$\\frac{3}{4}$","?`", |
105 |
/* 0xc0 */ "\\`A","\\'A","\\^A","\\~A", "\\\"A","\\AA","\\AE","\\c{C}", |
106 |
/* 0xc8 */ "\\`E","\\'E","\\^E","\\\"E", "\\`I","\\'I","\\^I","\\\"I", |
107 |
/* 0xd0 */ "{D\\hspace*{-1.7ex}-\\hspace{.9ex}}","\\~N","\\`O","\\'O", |
108 |
"\\^O","\\~O","\\\"O","$\\times$", |
109 |
/* 0xd8 */ "{\\O}","\\`U","\\'U","\\^U", "\\\"U","\\'Y","","", |
110 |
/* 0xe0 */ "\\`a","\\'a","\\^a","\\~a", |
111 |
"\\\"a","\\r{a}","{\\ae}","\\c{c}", |
112 |
/* 0xe8 */ "\\`e","\\'e","\\^e","\\\"e", |
113 |
"\\`{\\i}","\\'{\\i}","\\^{\\i}","\\\"{\\i}", |
114 |
/* 0xf0 */ "\\v{o}","\\~n","\\`o","\\'o", "\\^o","\\~o","\\\"o","$\\div$", |
115 |
/* 0xf8 */ "{\\o}","\\`u","\\'u","\\^u", "\\\"u","\\'y","","\\\"y" |
116 |
}; |
117 |
|
118 |
static psiconv_character_layout gen_base_char(const psiconv_font font, |
119 |
const psiconv_color color, |
120 |
const psiconv_color |
121 |
back_color); |
122 |
static void diff_char(FILE *of, const psiconv_character_layout old, |
123 |
const psiconv_character_layout new, int *flags); |
124 |
static void gen_para(FILE *of, const psiconv_paragraph para, |
125 |
const psiconv_character_layout base_char, |
126 |
psiconv_word_f wf); |
127 |
|
128 |
static void psiconv_gen_latex_word(FILE *of,psiconv_word_f wf); |
129 |
static void psiconv_gen_latex_texted(FILE *of,psiconv_texted_f tf); |
130 |
|
131 |
psiconv_character_layout gen_base_char(const psiconv_font font, |
132 |
const psiconv_color color, |
133 |
const psiconv_color back_color) |
134 |
{ |
135 |
struct psiconv_character_layout_s base_char_struct = |
136 |
{ |
137 |
NULL, /* color */ |
138 |
NULL, /* back_color */ |
139 |
13.0, /* font_size */ |
140 |
psiconv_bool_false, /* italic */ |
141 |
psiconv_bool_false, /* bold */ |
142 |
psiconv_normalscript, /* super_sub */ |
143 |
psiconv_bool_false, /* underline */ |
144 |
psiconv_bool_false, /* strikethrough */ |
145 |
NULL, /* font */ |
146 |
}; |
147 |
base_char_struct.color = color; |
148 |
base_char_struct.back_color = back_color; |
149 |
base_char_struct.font = font; |
150 |
return psiconv_clone_character_layout(&base_char_struct); |
151 |
} |
152 |
|
153 |
/* flags & 1: 1 if in a section |
154 |
flags & 2: 1 if at end-of-paragraph |
155 |
*/ |
156 |
void diff_char(FILE *of, const psiconv_character_layout old, |
157 |
const psiconv_character_layout new, |
158 |
int *flags) |
159 |
{ |
160 |
if ((*flags & 3) == 3) { /* end of section command argument */ |
161 |
putc('}',of); |
162 |
return; |
163 |
} |
164 |
if (old->italic != new->italic) { |
165 |
if (old->italic) |
166 |
putc('}',of); |
167 |
else |
168 |
fputs("\\emph{",of); |
169 |
} |
170 |
if (old->bold != new->bold) { |
171 |
if (old->bold) |
172 |
putc('}',of); |
173 |
else |
174 |
fputs("\\textbf{",of); |
175 |
} |
176 |
if (old->underline != new->underline) { |
177 |
if (old->underline) |
178 |
putc('}',of); |
179 |
else |
180 |
fputs("\\underline{",of); |
181 |
} |
182 |
} |
183 |
|
184 |
|
185 |
const static char *sections[] = { |
186 |
"section", |
187 |
"subsection", |
188 |
"subsubsection", |
189 |
"paragraph", |
190 |
"subparagraph" |
191 |
}; |
192 |
|
193 |
struct environment { |
194 |
char *style_name; |
195 |
char *environment_name; |
196 |
}; |
197 |
|
198 |
const static struct environment environments[] = { |
199 |
{"quotation", "quotation"}, |
200 |
{"", NULL} |
201 |
}; |
202 |
|
203 |
char *find_env(char *style) { |
204 |
int n = 0; |
205 |
while (environments[n].environment_name != NULL) { |
206 |
if (strcmp(style, environments[n].style_name) == 0) { |
207 |
return environments[n].environment_name; |
208 |
} |
209 |
n++; |
210 |
} |
211 |
return NULL; |
212 |
} |
213 |
|
214 |
psiconv_bool_t bullet_switch_on = psiconv_bool_false; |
215 |
|
216 |
void gen_para(FILE *of, const psiconv_paragraph para, |
217 |
const psiconv_character_layout base_char, |
218 |
psiconv_word_f wf) |
219 |
{ |
220 |
int i,j,loc; |
221 |
psiconv_character_layout cur_char; |
222 |
psiconv_in_line_layout inl; |
223 |
int flags = 0; |
224 |
psiconv_word_style sty; |
225 |
char *env = NULL; |
226 |
|
227 |
if (para->base_paragraph->bullet->on) { |
228 |
if (! bullet_switch_on) { |
229 |
fputs("\\begin{itemize}\n\n", of); |
230 |
bullet_switch_on = psiconv_bool_true; |
231 |
} |
232 |
fputs("\\item ",of); |
233 |
} else { |
234 |
if (bullet_switch_on) { |
235 |
fputs("\\end{itemize}\n\n", of); |
236 |
bullet_switch_on = psiconv_bool_false; |
237 |
} |
238 |
} |
239 |
|
240 |
cur_char = base_char; |
241 |
|
242 |
if (wf) { |
243 |
sty = psiconv_get_style(wf->styles_sec, para->base_style); |
244 |
if (sty->name && (env = find_env(sty->name))) { |
245 |
fputs("\\begin{",of); |
246 |
fputs(env,of); |
247 |
fputs("}\n",of); |
248 |
} else { |
249 |
if (sty->outline_level && |
250 |
(sty->outline_level > 0) && (sty->outline_level < 6)) { |
251 |
putc('\\', of); |
252 |
fputs(sections[(sty->outline_level - 1)], of); |
253 |
putc('{', of); |
254 |
cur_char = para->base_character; /* ignore initial formatting */ |
255 |
flags |= 1; |
256 |
} |
257 |
} |
258 |
} |
259 |
|
260 |
if (psiconv_list_length(para->in_lines) == 0) { |
261 |
diff_char(of,cur_char,para->base_character,&flags); |
262 |
cur_char = para->base_character; |
263 |
} |
264 |
loc = 0; |
265 |
|
266 |
for (i = 0; i < psiconv_list_length(para->in_lines); i++) { |
267 |
inl = psiconv_list_get(para->in_lines,i); |
268 |
diff_char(of,cur_char,inl->layout,&flags); |
269 |
cur_char = inl->layout; |
270 |
for (j = loc; j < inl->length + loc; j ++) { |
271 |
fputs(char_table[(unsigned char) (para->text[j])],of); |
272 |
} |
273 |
loc = j; |
274 |
} |
275 |
|
276 |
if (loc < strlen(para->text)) { |
277 |
diff_char(of,cur_char,para->base_character,&flags); |
278 |
cur_char = para->base_character; |
279 |
for (j = loc; j < strlen(para->text); j ++) { |
280 |
fputs(char_table[(unsigned char) (para->text[j])],of); |
281 |
} |
282 |
} |
283 |
|
284 |
flags |= 2; |
285 |
diff_char(of,cur_char,base_char,&flags); |
286 |
|
287 |
if (env) { |
288 |
fputs("\n\\end{",of); |
289 |
fputs(env,of); |
290 |
putc('}',of); |
291 |
} |
292 |
|
293 |
fputs("\n\n", of); |
294 |
} |
295 |
|
296 |
int psiconv_gen_latex(const char * filename,const psiconv_file file, |
297 |
const char *dest) |
298 |
{ |
299 |
FILE *of = fopen(filename,"w"); |
300 |
if (! of) |
301 |
return -1; |
302 |
|
303 |
if (file->type == psiconv_word_file) { |
304 |
psiconv_gen_latex_word(of,(psiconv_word_f) file->file); |
305 |
} else if (file->type == psiconv_texted_file) { |
306 |
psiconv_gen_latex_texted(of,(psiconv_texted_f) file->file); |
307 |
} else { |
308 |
fclose(of); |
309 |
return -1; |
310 |
} |
311 |
return fclose(of); |
312 |
} |
313 |
|
314 |
/* This isn't tested !!! */ |
315 |
void psiconv_gen_latex_texted(FILE *of,psiconv_texted_f tf) |
316 |
{ |
317 |
psiconv_character_layout base_char; |
318 |
psiconv_paragraph para; |
319 |
int i; |
320 |
|
321 |
/* We have nothing better */ |
322 |
base_char = psiconv_basic_character_layout(); |
323 |
|
324 |
fputs("\\documentclass{article}\n\n\\begin{document}\n\n", of); |
325 |
for (i = 0; i < psiconv_list_length(tf->texted_sec->paragraphs); i++) { |
326 |
para = psiconv_list_get(tf->texted_sec->paragraphs,i); |
327 |
gen_para(of,para,base_char, NULL); |
328 |
} |
329 |
fputs("\\end{document}",of); |
330 |
psiconv_free_character_layout(base_char); |
331 |
} |
332 |
|
333 |
void psiconv_gen_latex_word(FILE *of,psiconv_word_f wf) |
334 |
{ |
335 |
int i; |
336 |
psiconv_paragraph para; |
337 |
psiconv_color white,black; |
338 |
psiconv_character_layout base_char; |
339 |
|
340 |
white = malloc(sizeof(*white)); |
341 |
black = malloc(sizeof(*black)); |
342 |
white->red = 0x00; |
343 |
white->green = 0x00; |
344 |
white->blue = 0x00; |
345 |
black->red = 0xff; |
346 |
black->green = 0xff; |
347 |
black->blue = 0xff; |
348 |
|
349 |
/* To keep from generating a font desc for each line */ |
350 |
base_char = gen_base_char(wf->styles_sec->normal->character->font, |
351 |
black,white); |
352 |
|
353 |
psiconv_free_color(black); |
354 |
psiconv_free_color(white); |
355 |
|
356 |
fputs("\\documentclass{article}\n\n\\begin{document}\n\n", of); |
357 |
|
358 |
for (i = 0; i < psiconv_list_length(wf->paragraphs); i++) { |
359 |
para = psiconv_list_get(wf->paragraphs,i); |
360 |
gen_para(of,para,base_char,wf); |
361 |
} |
362 |
fputs("\\end{document}\n",of); |
363 |
for (i = 0; i < |
364 |
psiconv_list_length(wf->page_sec->header->text->paragraphs); i++) { |
365 |
para = psiconv_list_get(wf->page_sec->header->text->paragraphs,i); |
366 |
gen_para(of,para,base_char,wf); |
367 |
} |
368 |
|
369 |
psiconv_free_character_layout(base_char); |
370 |
} |
371 |
|
372 |
static struct psiconv_fileformat_s ff = |
373 |
{ |
374 |
"LaTeX", |
375 |
"LaTeX conversion to article class", |
376 |
&psiconv_gen_latex |
377 |
}; |
378 |
|
379 |
void init_latex(void) |
380 |
{ |
381 |
psiconv_list_add(fileformat_list,&ff); |
382 |
} |
383 |
|