lib/lyx2lyx/LyX.py

   1 # This file is part of lyx2lyx
   2 # -*- coding: iso-8859-1 -*-
   3 # Copyright (C) 2002-2004 Dekel Tsur <dekel@lyx.org>, José Matos <jamatos@lyx.org>
   4 #
   5 # This program is free software; you can redistribute it and/or
   6 # modify it under the terms of the GNU General Public License
   7 # as published by the Free Software Foundation; either version 2
   8 # of the License, or (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., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA.
  18
  19 from parser_tools import get_value, check_token, find_token,\
  20      find_tokens, find_end_of, find_end_of_inset
  21 import os.path
  22 import gzip
  23 import sys
  24 import re
  25 import string
  26 import time
  27
  28 version_lyx2lyx = "1.4.0cvs"
  29 default_debug_level = 2
  30
  31 # Regular expressions used
  32 format_re = re.compile(r"(\d)[\.,]?(\d\d)")
  33 fileformat = re.compile(r"\\lyxformat\s*(\S*)")
  34 original_version = re.compile(r"\#LyX (\S*)")
  35
  36 ##
  37 # file format information:
  38 #  file, supported formats, stable release versions
  39 format_relation = [("0_10",  [210], ["0.10.7","0.10"]),
  40                    ("0_12",  [215], ["0.12","0.12.1","0.12"]),
  41                    ("1_0_0", [215], ["1.0.0","1.0"]),
  42                    ("1_0_1", [215], ["1.0.1","1.0.2","1.0.3","1.0.4", "1.1.2","1.1"]),
  43                    ("1_1_4", [215], ["1.1.4","1.1"]),
  44                    ("1_1_5", [216], ["1.1.5","1.1.5fix1","1.1.5fix2","1.1"]),
  45                    ("1_1_6", [217], ["1.1.6","1.1.6fix1","1.1.6fix2","1.1"]),
  46                    ("1_1_6fix3", [218], ["1.1.6fix3","1.1.6fix4","1.1"]),
  47                    ("1_2", [220], ["1.2.0","1.2.1","1.2.3","1.2.4","1.2"]),
  48                    ("1_3", [221], ["1.3.0","1.3.1","1.3.2","1.3.3","1.3.4","1.3.5","1.3.6","1.3"]),
  49                    ("1_4", range(222,246), ["1.4.0cvs","1.4"])]
  50
  51
  52 def formats_list():
  53     " Returns a list with supported file formats."
  54     formats = []
  55     for version in format_relation:
  56         for format in version[1]:
  57             if format not in formats:
  58                 formats.append(format)
  59     return formats
  60
  61
  62 def get_end_format():
  63     " Returns the more recent file format available."
  64     return format_relation[-1][1][-1]
  65
  66
  67 def get_backend(textclass):
  68     " For _textclass_ returns its backend."
  69     if textclass == "linuxdoc" or textclass == "manpage":
  70         return "linuxdoc"
  71     if textclass[:7] == "docbook":
  72         return "docbook"
  73     return "latex"
  74
  75
  76 def trim_eol(line):
  77     " Remove end of line char(s)."
  78     if line[-2:-1] == '\r':
  79         return line[:-2]
  80     else:
  81         return line[:-1]
  82
  83
  84 ##
  85 # Class
  86 #
  87 class LyX_Base:
  88     """This class carries all the information of the LyX file."""
  89     def __init__(self, end_format = 0, input = "", output = "", error = "", debug = default_debug_level, try_hard = 0):
  90         """Arguments:
  91         end_format: final format that the file should be converted. (integer)
  92         input: the name of the input source, if empty resort to standard input.
  93         output: the name of the output file, if empty use the standard output.
  94         error: the name of the error file, if empty use the standard error.
  95         debug: debug level, O means no debug, as its value increases be more verbose.
  96         """
  97         self.choose_io(input, output)
  98
  99         if error:
 100             self.err = open(error, "w")
 101         else:
 102             self.err = sys.stderr
 103
 104         self.debug = debug
 105         self.try_hard = try_hard
 106
 107         if end_format:
 108             self.end_format = self.lyxformat(end_format)
 109         else:
 110             self.end_format = get_end_format()
 111
 112         self.backend = "latex"
 113         self.textclass = "article"
 114         self.header = []
 115         self.preamble = []
 116         self.body = []
 117         self.status = 0
 118
 119
 120     def warning(self, message, debug_level= default_debug_level):
 121         " Emits warning to self.error, if the debug_level is less than the self.debug."
 122         if debug_level <= self.debug:
 123             self.err.write("Warning: " + message + "\n")
 124
 125
 126     def error(self, message):
 127         " Emits a warning and exits if not in try_hard mode."
 128         self.warning(message)
 129         if not self.try_hard:
 130             self.warning("Quiting.")
 131             sys.exit(1)
 132
 133         self.status = 2
 134
 135
 136     def read(self):
 137         """Reads a file into the self.header and self.body parts, from self.input."""
 138
 139         while 1:
 140             line = self.input.readline()
 141             if not line:
 142                 self.error("Invalid LyX file.")
 143
 144             line = trim_eol(line)
 145             if check_token(line, '\\begin_preamble'):
 146                 while 1:
 147                     line = self.input.readline()
 148                     if not line:
 149                         self.error("Invalid LyX file.")
 150
 151                     line = trim_eol(line)
 152                     if check_token(line, '\\end_preamble'):
 153                         break
 154
 155                     if string.split(line)[:0] in ("\\layout", "\\begin_layout", "\\begin_body"):
 156                         self.warning("Malformed LyX file: Missing '\\end_preamble'.")
 157                         self.warning("Adding it now and hoping for the best.")
 158
 159                     self.preamble.append(line)
 160
 161             if check_token(line, '\\end_preamble'):
 162                 continue
 163
 164             line = string.strip(line)
 165             if not line:
 166                 continue
 167
 168             if string.split(line)[0] in ("\\layout", "\\begin_layout", "\\begin_body"):
 169                 self.body.append(line)
 170                 break
 171
 172             self.header.append(line)
 173
 174         while 1:
 175             line = self.input.readline()
 176             if not line:
 177                 break
 178             self.body.append(trim_eol(line))
 179
 180         self.textclass = get_value(self.header, "\\textclass", 0)
 181         self.backend = get_backend(self.textclass)
 182         self.format  = self.read_format()
 183         self.language = get_value(self.header, "\\language", 0)
 184         if self.language == "":
 185             self.language = "english"
 186         self.initial_version = self.read_version()
 187
 188
 189     def write(self):
 190         " Writes the LyX file to self.output."
 191         self.set_version()
 192         self.set_format()
 193
 194         if self.preamble:
 195             i = find_token(self.header, '\\textclass', 0) + 1
 196             preamble = ['\\begin_preamble'] + self.preamble + ['\\end_preamble']
 197             if i == 0:
 198                 self.error("Malformed LyX file: Missing '\\textclass'.")
 199             else:
 200                 header = self.header[:i] + preamble + self.header[i:]
 201         else:
 202             header = self.header
 203
 204         for line in header + [''] + self.body:
 205             self.output.write(line+"\n")
 206
 207
 208     def choose_io(self, input, output):
 209         """Choose input and output streams, dealing transparently with
 210         compressed files."""
 211
 212         if output:
 213             self.output = open(output, "wb")
 214         else:
 215             self.output = sys.stdout
 216
 217         if input and input != '-':
 218             self.dir = os.path.dirname(os.path.abspath(input))
 219             try:
 220                 gzip.open(input).readline()
 221                 self.input = gzip.open(input)
 222                 self.output = gzip.GzipFile(mode="wb", fileobj=self.output)
 223             except:
 224                 self.input = open(input)
 225         else:
 226             self.input = sys.stdin
 227
 228
 229     def lyxformat(self, format):
 230         " Returns the file format representation, an integer."
 231         result = format_re.match(format)
 232         if result:
 233             format = int(result.group(1) + result.group(2))
 234         else:
 235             self.error(str(format) + ": " + "Invalid LyX file.")
 236
 237         if format in formats_list():
 238             return format
 239
 240         self.error(str(format) + ": " + "Format not supported.")
 241         return None
 242
 243
 244     def read_version(self):
 245         """ Searchs for clues of the LyX version used to write the file, returns the
 246         most likely value, or None otherwise."""
 247         for line in self.header:
 248             if line[0] != "#":
 249                 return None
 250
 251             result = original_version.match(line)
 252             if result:
 253                 return result.group(1)
 254         return None
 255
 256
 257     def set_version(self):
 258         " Set the header with the version used."
 259         self.header[0] = "#LyX %s created this file. For more info see http://www.lyx.org/" % version_lyx2lyx
 260         if self.header[1][0] == '#':
 261             del self.header[1]
 262
 263
 264     def read_format(self):
 265         " Read from the header the fileformat of the present LyX file."
 266         for line in self.header:
 267             result = fileformat.match(line)
 268             if result:
 269                 return self.lyxformat(result.group(1))
 270         else:
 271             self.error("Invalid LyX File.")
 272         return None
 273
 274
 275     def set_format(self):
 276         " Set the file format of the file, in the header."
 277         if self.format <= 217:
 278             format = str(float(self.format)/100)
 279         else:
 280             format = str(self.format)
 281         i = find_token(self.header, "\\lyxformat", 0)
 282         self.header[i] = "\\lyxformat %s" % format
 283
 284
 285     def set_parameter(self, param, value):
 286         " Set the value of the header parameter."
 287         i = find_token(self.header, '\\' + param, 0)
 288         if i == -1:
 289             self.warning('Parameter not found in the header: %s' % param, 3)
 290             return
 291         self.header[i] = '\\%s %s' % (param, str(value))
 292
 293
 294     def convert(self):
 295         "Convert from current (self.format) to self.end_format."
 296         mode, convertion_chain = self.chain()
 297         self.warning("convertion chain: " + str(convertion_chain), 3)
 298
 299         for step in convertion_chain:
 300             steps = getattr(__import__("lyx_" + step), mode)
 301
 302             self.warning("Convertion step: %s - %s" % (step, mode), default_debug_level + 1)
 303             if not steps:
 304                     self.error("The convertion to an older format (%s) is not implemented." % self.format)
 305
 306             multi_conv = len(steps) != 1
 307             for version, table in steps:
 308                 if multi_conv and \
 309                    (self.format >= version and mode == "convert") or\
 310                    (self.format <= version and mode == "revert"):
 311                     continue
 312
 313                 for conv in table:
 314                     init_t = time.time()
 315                     try:
 316                         conv(self)
 317                     except:
 318                         self.warning("An error ocurred in %s, %s" % (version, str(conv)),
 319                                      default_debug_level)
 320                         if not self.try_hard:
 321                             raise
 322                         self.status = 2
 323                     else:
 324                         self.warning("%lf: Elapsed time on %s"  % (time.time() - init_t, str(conv)),
 325                                      default_debug_level + 1)
 326
 327                 self.format = version
 328                 if self.end_format == self.format:
 329                     return
 330
 331
 332     def chain(self):
 333         """ This is where all the decisions related with the convertion are taken.
 334         It returns a list of modules needed to convert the LyX file from
 335         self.format to self.end_format"""
 336
 337         self.start =  self.format
 338         format = self.format
 339         correct_version = 0
 340
 341         for rel in format_relation:
 342             if self.initial_version in rel[2]:
 343                 if format in rel[1]:
 344                     initial_step = rel[0]
 345                     correct_version = 1
 346                     break
 347
 348         if not correct_version:
 349             if format <= 215:
 350                 self.warning("Version does not match file format, discarding it.")
 351             for rel in format_relation:
 352                 if format in rel[1]:
 353                     initial_step = rel[0]
 354                     break
 355             else:
 356                 # This should not happen, really.
 357                 self.error("Format not supported.")
 358
 359         # Find the final step
 360         for rel in format_relation:
 361             if self.end_format in rel[1]:
 362                 final_step = rel[0]
 363                 break
 364         else:
 365             self.error("Format not supported.")
 366
 367         # Convertion mode, back or forth
 368         steps = []
 369         if (initial_step, self.start) < (final_step, self.end_format):
 370             mode = "convert"
 371             first_step = 1
 372             for step in format_relation:
 373                 if  initial_step <= step[0] <= final_step:
 374                     if first_step and len(step[1]) == 1:
 375                         first_step = 0
 376                         continue
 377                     steps.append(step[0])
 378         else:
 379             mode = "revert"
 380             relation_format = format_relation[:]
 381             relation_format.reverse()
 382             last_step = None
 383
 384             for step in relation_format:
 385                 if  final_step <= step[0] <= initial_step:
 386                     steps.append(step[0])
 387                     last_step = step
 388
 389             if last_step[1][-1] == self.end_format:
 390                 steps.pop()
 391
 392         return mode, steps
 393
 394
 395     def get_toc(self, depth = 4):
 396         " Returns the TOC of this LyX document."
 397         paragraphs_filter = {'Title' : 0,'Chapter' : 1, 'Section' : 2, 'Subsection' : 3, 'Subsubsection': 4}
 398         allowed_insets = ['Quotes']
 399         allowed_parameters = '\\paragraph_spacing', '\\noindent', '\\align', '\\labelwidthstring', "\\start_of_appendix", "\\leftindent"
 400
 401         sections = []
 402         for section in paragraphs_filter.keys():
 403             sections.append('\\begin_layout %s' % section)
 404
 405         toc_par = []
 406         i = 0
 407         while 1:
 408             i = find_tokens(self.body, sections, i)
 409             if i == -1:
 410                 break
 411
 412             j = find_end_of(self.body,  i + 1, '\\begin_layout', '\\end_layout')
 413             if j == -1:
 414                 self.warning('Incomplete file.', 0)
 415                 break
 416
 417             section = string.split(self.body[i])[1]
 418             if section[-1] == '*':
 419                 section = section[:-1]
 420
 421             par = []
 422
 423             k = i + 1
 424             # skip paragraph parameters
 425             while not string.strip(self.body[k]) or string.split(self.body[k])[0] in allowed_parameters:
 426                 k = k +1
 427
 428             while k < j:
 429                 if check_token(self.body[k], '\\begin_inset'):
 430                     inset = string.split(self.body[k])[1]
 431                     end = find_end_of_inset(self.body, k)
 432                     if end == -1 or end > j:
 433                         self.warning('Malformed file.', 0)
 434
 435                     if inset in allowed_insets:
 436                         par.extend(self.body[k: end+1])
 437                     k = end + 1
 438                 else:
 439                     par.append(self.body[k])
 440                     k = k + 1
 441
 442             # trim empty lines in the end.
 443             while string.strip(par[-1]) == '' and par:
 444                 par.pop()
 445
 446             toc_par.append(Paragraph(section, par))
 447
 448             i = j + 1
 449
 450         return toc_par
 451
 452
 453 class File(LyX_Base):
 454     " This class reads existing LyX files."
 455     def __init__(self, end_format = 0, input = "", output = "", error = "", debug = default_debug_level, try_hard = 0):
 456         LyX_Base.__init__(self, end_format, input, output, error, debug, try_hard)
 457         self.read()
 458
 459
 460 class NewFile(LyX_Base):
 461     " This class is to create new LyX files."
 462     def set_header(self, **params):
 463         # set default values
 464         self.header.extend([
 465             "#LyX xxxx created this file. For more info see http://www.lyx.org/",
 466             "\\lyxformat xxx",
 467             "\\begin_document",
 468             "\\begin_header",
 469             "\\textclass article",
 470             "\\language english",
 471             "\\inputencoding auto",
 472             "\\fontscheme default",
 473             "\\graphics default",
 474             "\\paperfontsize default",
 475             "\\papersize default",
 476             "\\use_geometry false",
 477             "\\use_amsmath 1",
 478             "\\cite_engine basic",
 479             "\\use_bibtopic false",
 480             "\\paperorientation portrait",
 481             "\\secnumdepth 3",
 482             "\\tocdepth 3",
 483             "\\paragraph_separation indent",
 484             "\\defskip medskip",
 485             "\\quotes_language english",
 486             "\\papercolumns 1",
 487             "\\papersides 1",
 488             "\\paperpagestyle default",
 489             "\\tracking_changes false",
 490             "\\end_header"])
 491
 492         self.format = get_end_format()
 493         for param in params:
 494             self.set_parameter(param, params[param])
 495
 496
 497     def set_body(self, paragraphs):
 498         self.body.extend(['\\begin_body',''])
 499
 500         for par in paragraphs:
 501             self.body.extend(par.asLines())
 502
 503         self.body.extend(['','\\end_body', '\\end_document'])
 504
 505
 506 class Paragraph:
 507     # unfinished implementation, it is missing the Text and Insets representation.
 508     " This class represents the LyX paragraphs."
 509     def __init__(self, name, body=[], settings = [], child = []):
 510         """ Parameters:
 511         name: paragraph name.
 512         body: list of lines of body text.
 513         child: list of paragraphs that descend from this paragraph.
 514         """
 515         self.name = name
 516         self.body = body
 517         self.settings = settings
 518         self.child = child
 519
 520     def asLines(self):
 521         " Converts the paragraph to a list of strings, representing it in the LyX file."
 522         result = ['','\\begin_layout %s' % self.name]
 523         result.extend(self.settings)
 524         result.append('')
 525         result.extend(self.body)
 526         result.append('\\end_layout')
 527
 528         if not self.child:
 529             return result
 530
 531         result.append('\\begin_deeper')
 532         for node in self.child:
 533             result.extend(node.asLines())
 534         result.append('\\end_deeper')
 535
 536         return result
 537
 538
 539 class Inset:
 540     " This class represents the LyX insets."
 541     pass
 542
 543
 544 class Text:
 545     " This class represents simple chuncks of text."
 546     pass