/usr/lib/python3/dist-packages/matplotlib/dviread.py is in python3-matplotlib 2.1.1-2ubuntu3.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 | """
A module for reading dvi files output by TeX. Several limitations make
this not (currently) useful as a general-purpose dvi preprocessor, but
it is currently used by the pdf backend for processing usetex text.
Interface::
with Dvi(filename, 72) as dvi:
# iterate over pages:
for page in dvi:
w, h, d = page.width, page.height, page.descent
for x,y,font,glyph,width in page.text:
fontname = font.texname
pointsize = font.size
...
for x,y,height,width in page.boxes:
...
"""
from __future__ import (absolute_import, division, print_function,
unicode_literals)
import six
from six.moves import xrange
from collections import namedtuple
import errno
from functools import partial, wraps
import matplotlib
import matplotlib.cbook as mpl_cbook
from matplotlib.compat import subprocess
from matplotlib import rcParams
import numpy as np
import re
import struct
import sys
import textwrap
import os
if six.PY3:
def ord(x):
return x
# Dvi is a bytecode format documented in
# http://mirrors.ctan.org/systems/knuth/dist/texware/dvitype.web
# http://texdoc.net/texmf-dist/doc/generic/knuth/texware/dvitype.pdf
#
# The file consists of a preamble, some number of pages, a postamble,
# and a finale. Different opcodes are allowed in different contexts,
# so the Dvi object has a parser state:
#
# pre: expecting the preamble
# outer: between pages (followed by a page or the postamble,
# also e.g. font definitions are allowed)
# page: processing a page
# post_post: state after the postamble (our current implementation
# just stops reading)
# finale: the finale (unimplemented in our current implementation)
_dvistate = mpl_cbook.Bunch(pre=0, outer=1, inpage=2, post_post=3, finale=4)
# The marks on a page consist of text and boxes. A page also has dimensions.
Page = namedtuple('Page', 'text boxes height width descent')
Text = namedtuple('Text', 'x y font glyph width')
Box = namedtuple('Box', 'x y height width')
# Opcode argument parsing
#
# Each of the following functions takes a Dvi object and delta,
# which is the difference between the opcode and the minimum opcode
# with the same meaning. Dvi opcodes often encode the number of
# argument bytes in this delta.
def _arg_raw(dvi, delta):
"""Return *delta* without reading anything more from the dvi file"""
return delta
def _arg(bytes, signed, dvi, _):
"""Read *bytes* bytes, returning the bytes interpreted as a
signed integer if *signed* is true, unsigned otherwise."""
return dvi._arg(bytes, signed)
def _arg_slen(dvi, delta):
"""Signed, length *delta*
Read *delta* bytes, returning None if *delta* is zero, and
the bytes interpreted as a signed integer otherwise."""
if delta == 0:
return None
return dvi._arg(delta, True)
def _arg_slen1(dvi, delta):
"""Signed, length *delta*+1
Read *delta*+1 bytes, returning the bytes interpreted as signed."""
return dvi._arg(delta+1, True)
def _arg_ulen1(dvi, delta):
"""Unsigned length *delta*+1
Read *delta*+1 bytes, returning the bytes interpreted as unsigned."""
return dvi._arg(delta+1, False)
def _arg_olen1(dvi, delta):
"""Optionally signed, length *delta*+1
Read *delta*+1 bytes, returning the bytes interpreted as
unsigned integer for 0<=*delta*<3 and signed if *delta*==3."""
return dvi._arg(delta + 1, delta == 3)
_arg_mapping = dict(raw=_arg_raw,
u1=partial(_arg, 1, False),
u4=partial(_arg, 4, False),
s4=partial(_arg, 4, True),
slen=_arg_slen,
olen1=_arg_olen1,
slen1=_arg_slen1,
ulen1=_arg_ulen1)
def _dispatch(table, min, max=None, state=None, args=('raw',)):
"""Decorator for dispatch by opcode. Sets the values in *table*
from *min* to *max* to this method, adds a check that the Dvi state
matches *state* if not None, reads arguments from the file according
to *args*.
*table*
the dispatch table to be filled in
*min*
minimum opcode for calling this function
*max*
maximum opcode for calling this function, None if only *min* is allowed
*state*
state of the Dvi object in which these opcodes are allowed
*args*
sequence of argument specifications:
``'raw'``: opcode minus minimum
``'u1'``: read one unsigned byte
``'u4'``: read four bytes, treat as an unsigned number
``'s4'``: read four bytes, treat as a signed number
``'slen'``: read (opcode - minimum) bytes, treat as signed
``'slen1'``: read (opcode - minimum + 1) bytes, treat as signed
``'ulen1'``: read (opcode - minimum + 1) bytes, treat as unsigned
``'olen1'``: read (opcode - minimum + 1) bytes, treat as unsigned
if under four bytes, signed if four bytes
"""
def decorate(method):
get_args = [_arg_mapping[x] for x in args]
@wraps(method)
def wrapper(self, byte):
if state is not None and self.state != state:
raise ValueError("state precondition failed")
return method(self, *[f(self, byte-min) for f in get_args])
if max is None:
table[min] = wrapper
else:
for i in xrange(min, max+1):
assert table[i] is None
table[i] = wrapper
return wrapper
return decorate
class Dvi(object):
"""
A reader for a dvi ("device-independent") file, as produced by TeX.
The current implementation can only iterate through pages in order,
and does not even attempt to verify the postamble.
This class can be used as a context manager to close the underlying
file upon exit. Pages can be read via iteration. Here is an overly
simple way to extract text without trying to detect whitespace::
>>> with matplotlib.dviread.Dvi('input.dvi', 72) as dvi:
>>> for page in dvi:
>>> print ''.join(unichr(t.glyph) for t in page.text)
"""
# dispatch table
_dtable = [None for _ in xrange(256)]
dispatch = partial(_dispatch, _dtable)
def __init__(self, filename, dpi):
"""
Read the data from the file named *filename* and convert
TeX's internal units to units of *dpi* per inch.
*dpi* only sets the units and does not limit the resolution.
Use None to return TeX's internal units.
"""
matplotlib.verbose.report('Dvi: ' + filename, 'debug')
self.file = open(filename, 'rb')
self.dpi = dpi
self.fonts = {}
self.state = _dvistate.pre
self.baseline = self._get_baseline(filename)
def _get_baseline(self, filename):
if rcParams['text.latex.preview']:
base, ext = os.path.splitext(filename)
baseline_filename = base + ".baseline"
if os.path.exists(baseline_filename):
with open(baseline_filename, 'rb') as fd:
l = fd.read().split()
height, depth, width = l
return float(depth)
return None
def __enter__(self):
"""
Context manager enter method, does nothing.
"""
return self
def __exit__(self, etype, evalue, etrace):
"""
Context manager exit method, closes the underlying file if it is open.
"""
self.close()
def __iter__(self):
"""
Iterate through the pages of the file.
Yields
------
Page
Details of all the text and box objects on the page.
The Page tuple contains lists of Text and Box tuples and
the page dimensions, and the Text and Box tuples contain
coordinates transformed into a standard Cartesian
coordinate system at the dpi value given when initializing.
The coordinates are floating point numbers, but otherwise
precision is not lost and coordinate values are not clipped to
integers.
"""
while True:
have_page = self._read()
if have_page:
yield self._output()
else:
break
def close(self):
"""
Close the underlying file if it is open.
"""
if not self.file.closed:
self.file.close()
def _output(self):
"""
Output the text and boxes belonging to the most recent page.
page = dvi._output()
"""
minx, miny, maxx, maxy = np.inf, np.inf, -np.inf, -np.inf
maxy_pure = -np.inf
for elt in self.text + self.boxes:
if isinstance(elt, Box):
x, y, h, w = elt
e = 0 # zero depth
else: # glyph
x, y, font, g, w = elt
h, e = font._height_depth_of(g)
minx = min(minx, x)
miny = min(miny, y - h)
maxx = max(maxx, x + w)
maxy = max(maxy, y + e)
maxy_pure = max(maxy_pure, y)
if self.dpi is None:
# special case for ease of debugging: output raw dvi coordinates
return Page(text=self.text, boxes=self.boxes,
width=maxx-minx, height=maxy_pure-miny,
descent=maxy-maxy_pure)
# convert from TeX's "scaled points" to dpi units
d = self.dpi / (72.27 * 2**16)
if self.baseline is None:
descent = (maxy - maxy_pure) * d
else:
descent = self.baseline
text = [Text((x-minx)*d, (maxy-y)*d - descent, f, g, w*d)
for (x, y, f, g, w) in self.text]
boxes = [Box((x-minx)*d, (maxy-y)*d - descent, h*d, w*d)
for (x, y, h, w) in self.boxes]
return Page(text=text, boxes=boxes, width=(maxx-minx)*d,
height=(maxy_pure-miny)*d, descent=descent)
def _read(self):
"""
Read one page from the file. Return True if successful,
False if there were no more pages.
"""
while True:
byte = ord(self.file.read(1)[0])
self._dtable[byte](self, byte)
if byte == 140: # end of page
return True
if self.state == _dvistate.post_post: # end of file
self.close()
return False
def _arg(self, nbytes, signed=False):
"""
Read and return an integer argument *nbytes* long.
Signedness is determined by the *signed* keyword.
"""
str = self.file.read(nbytes)
value = ord(str[0])
if signed and value >= 0x80:
value = value - 0x100
for i in range(1, nbytes):
value = 0x100*value + ord(str[i])
return value
@dispatch(min=0, max=127, state=_dvistate.inpage)
def _set_char_immediate(self, char):
self._put_char_real(char)
self.h += self.fonts[self.f]._width_of(char)
@dispatch(min=128, max=131, state=_dvistate.inpage, args=('olen1',))
def _set_char(self, char):
self._put_char_real(char)
self.h += self.fonts[self.f]._width_of(char)
@dispatch(132, state=_dvistate.inpage, args=('s4', 's4'))
def _set_rule(self, a, b):
self._put_rule_real(a, b)
self.h += b
@dispatch(min=133, max=136, state=_dvistate.inpage, args=('olen1',))
def _put_char(self, char):
self._put_char_real(char)
def _put_char_real(self, char):
font = self.fonts[self.f]
if font._vf is None:
self.text.append(Text(self.h, self.v, font, char,
font._width_of(char)))
else:
scale = font._scale
for x, y, f, g, w in font._vf[char].text:
newf = DviFont(scale=_mul2012(scale, f._scale),
tfm=f._tfm, texname=f.texname, vf=f._vf)
self.text.append(Text(self.h + _mul2012(x, scale),
self.v + _mul2012(y, scale),
newf, g, newf._width_of(g)))
self.boxes.extend([Box(self.h + _mul2012(x, scale),
self.v + _mul2012(y, scale),
_mul2012(a, scale), _mul2012(b, scale))
for x, y, a, b in font._vf[char].boxes])
@dispatch(137, state=_dvistate.inpage, args=('s4', 's4'))
def _put_rule(self, a, b):
self._put_rule_real(a, b)
def _put_rule_real(self, a, b):
if a > 0 and b > 0:
self.boxes.append(Box(self.h, self.v, a, b))
@dispatch(138)
def _nop(self, _):
pass
@dispatch(139, state=_dvistate.outer, args=('s4',)*11)
def _bop(self, c0, c1, c2, c3, c4, c5, c6, c7, c8, c9, p):
self.state = _dvistate.inpage
self.h, self.v, self.w, self.x, self.y, self.z = 0, 0, 0, 0, 0, 0
self.stack = []
self.text = [] # list of Text objects
self.boxes = [] # list of Box objects
@dispatch(140, state=_dvistate.inpage)
def _eop(self, _):
self.state = _dvistate.outer
del self.h, self.v, self.w, self.x, self.y, self.z, self.stack
@dispatch(141, state=_dvistate.inpage)
def _push(self, _):
self.stack.append((self.h, self.v, self.w, self.x, self.y, self.z))
@dispatch(142, state=_dvistate.inpage)
def _pop(self, _):
self.h, self.v, self.w, self.x, self.y, self.z = self.stack.pop()
@dispatch(min=143, max=146, state=_dvistate.inpage, args=('slen1',))
def _right(self, b):
self.h += b
@dispatch(min=147, max=151, state=_dvistate.inpage, args=('slen',))
def _right_w(self, new_w):
if new_w is not None:
self.w = new_w
self.h += self.w
@dispatch(min=152, max=156, state=_dvistate.inpage, args=('slen',))
def _right_x(self, new_x):
if new_x is not None:
self.x = new_x
self.h += self.x
@dispatch(min=157, max=160, state=_dvistate.inpage, args=('slen1',))
def _down(self, a):
self.v += a
@dispatch(min=161, max=165, state=_dvistate.inpage, args=('slen',))
def _down_y(self, new_y):
if new_y is not None:
self.y = new_y
self.v += self.y
@dispatch(min=166, max=170, state=_dvistate.inpage, args=('slen',))
def _down_z(self, new_z):
if new_z is not None:
self.z = new_z
self.v += self.z
@dispatch(min=171, max=234, state=_dvistate.inpage)
def _fnt_num_immediate(self, k):
self.f = k
@dispatch(min=235, max=238, state=_dvistate.inpage, args=('olen1',))
def _fnt_num(self, new_f):
self.f = new_f
@dispatch(min=239, max=242, args=('ulen1',))
def _xxx(self, datalen):
special = self.file.read(datalen)
if six.PY3:
chr_ = chr
else:
def chr_(x):
return x
matplotlib.verbose.report(
'Dvi._xxx: encountered special: %s'
% ''.join([(32 <= ord(ch) < 127) and chr_(ch)
or '<%02x>' % ord(ch)
for ch in special]),
'debug')
@dispatch(min=243, max=246, args=('olen1', 'u4', 'u4', 'u4', 'u1', 'u1'))
def _fnt_def(self, k, c, s, d, a, l):
self._fnt_def_real(k, c, s, d, a, l)
def _fnt_def_real(self, k, c, s, d, a, l):
n = self.file.read(a + l)
fontname = n[-l:].decode('ascii')
tfm = _tfmfile(fontname)
if tfm is None:
if six.PY2:
error_class = OSError
else:
error_class = FileNotFoundError
raise error_class("missing font metrics file: %s" % fontname)
if c != 0 and tfm.checksum != 0 and c != tfm.checksum:
raise ValueError('tfm checksum mismatch: %s' % n)
vf = _vffile(fontname)
self.fonts[k] = DviFont(scale=s, tfm=tfm, texname=n, vf=vf)
@dispatch(247, state=_dvistate.pre, args=('u1', 'u4', 'u4', 'u4', 'u1'))
def _pre(self, i, num, den, mag, k):
comment = self.file.read(k)
if i != 2:
raise ValueError("Unknown dvi format %d" % i)
if num != 25400000 or den != 7227 * 2**16:
raise ValueError("nonstandard units in dvi file")
# meaning: TeX always uses those exact values, so it
# should be enough for us to support those
# (There are 72.27 pt to an inch so 7227 pt =
# 7227 * 2**16 sp to 100 in. The numerator is multiplied
# by 10^5 to get units of 10**-7 meters.)
if mag != 1000:
raise ValueError("nonstandard magnification in dvi file")
# meaning: LaTeX seems to frown on setting \mag, so
# I think we can assume this is constant
self.state = _dvistate.outer
@dispatch(248, state=_dvistate.outer)
def _post(self, _):
self.state = _dvistate.post_post
# TODO: actually read the postamble and finale?
# currently post_post just triggers closing the file
@dispatch(249)
def _post_post(self, _):
raise NotImplementedError
@dispatch(min=250, max=255)
def _malformed(self, offset):
raise ValueError("unknown command: byte %d", 250 + offset)
class DviFont(object):
"""
Encapsulation of a font that a DVI file can refer to.
This class holds a font's texname and size, supports comparison,
and knows the widths of glyphs in the same units as the AFM file.
There are also internal attributes (for use by dviread.py) that
are *not* used for comparison.
The size is in Adobe points (converted from TeX points).
Parameters
----------
scale : float
Factor by which the font is scaled from its natural size.
tfm : Tfm
TeX font metrics for this font
texname : bytes
Name of the font as used internally by TeX and friends, as an
ASCII bytestring. This is usually very different from any external
font names, and :class:`dviread.PsfontsMap` can be used to find
the external name of the font.
vf : Vf
A TeX "virtual font" file, or None if this font is not virtual.
Attributes
----------
texname : bytes
size : float
Size of the font in Adobe points, converted from the slightly
smaller TeX points.
widths : list
Widths of glyphs in glyph-space units, typically 1/1000ths of
the point size.
"""
__slots__ = ('texname', 'size', 'widths', '_scale', '_vf', '_tfm')
def __init__(self, scale, tfm, texname, vf):
if not isinstance(texname, bytes):
raise ValueError("texname must be a bytestring, got %s"
% type(texname))
self._scale, self._tfm, self.texname, self._vf = \
scale, tfm, texname, vf
self.size = scale * (72.0 / (72.27 * 2**16))
try:
nchars = max(tfm.width) + 1
except ValueError:
nchars = 0
self.widths = [(1000*tfm.width.get(char, 0)) >> 20
for char in xrange(nchars)]
def __eq__(self, other):
return self.__class__ == other.__class__ and \
self.texname == other.texname and self.size == other.size
def __ne__(self, other):
return not self.__eq__(other)
def _width_of(self, char):
"""
Width of char in dvi units. For internal use by dviread.py.
"""
width = self._tfm.width.get(char, None)
if width is not None:
return _mul2012(width, self._scale)
matplotlib.verbose.report(
'No width for char %d in font %s' % (char, self.texname),
'debug')
return 0
def _height_depth_of(self, char):
"""
Height and depth of char in dvi units. For internal use by dviread.py.
"""
result = []
for metric, name in ((self._tfm.height, "height"),
(self._tfm.depth, "depth")):
value = metric.get(char, None)
if value is None:
matplotlib.verbose.report(
'No %s for char %d in font %s' % (
name, char, self.texname),
'debug')
result.append(0)
else:
result.append(_mul2012(value, self._scale))
return result
class Vf(Dvi):
"""
A virtual font (\\*.vf file) containing subroutines for dvi files.
Usage::
vf = Vf(filename)
glyph = vf[code]
glyph.text, glyph.boxes, glyph.width
Parameters
----------
filename : string or bytestring
Notes
-----
The virtual font format is a derivative of dvi:
http://mirrors.ctan.org/info/knuth/virtual-fonts
This class reuses some of the machinery of `Dvi`
but replaces the `_read` loop and dispatch mechanism.
"""
def __init__(self, filename):
Dvi.__init__(self, filename, 0)
try:
self._first_font = None
self._chars = {}
self._read()
finally:
self.close()
def __getitem__(self, code):
return self._chars[code]
def _read(self):
"""
Read one page from the file. Return True if successful,
False if there were no more pages.
"""
packet_len, packet_char, packet_width = None, None, None
while True:
byte = ord(self.file.read(1)[0])
# If we are in a packet, execute the dvi instructions
if self.state == _dvistate.inpage:
byte_at = self.file.tell()-1
if byte_at == packet_ends:
self._finalize_packet(packet_char, packet_width)
packet_len, packet_char, packet_width = None, None, None
# fall through to out-of-packet code
elif byte_at > packet_ends:
raise ValueError("Packet length mismatch in vf file")
else:
if byte in (139, 140) or byte >= 243:
raise ValueError(
"Inappropriate opcode %d in vf file" % byte)
Dvi._dtable[byte](self, byte)
continue
# We are outside a packet
if byte < 242: # a short packet (length given by byte)
packet_len = byte
packet_char, packet_width = self._arg(1), self._arg(3)
packet_ends = self._init_packet(byte)
self.state = _dvistate.inpage
elif byte == 242: # a long packet
packet_len, packet_char, packet_width = \
[self._arg(x) for x in (4, 4, 4)]
self._init_packet(packet_len)
elif 243 <= byte <= 246:
k = self._arg(byte - 242, byte == 246)
c, s, d, a, l = [self._arg(x) for x in (4, 4, 4, 1, 1)]
self._fnt_def_real(k, c, s, d, a, l)
if self._first_font is None:
self._first_font = k
elif byte == 247: # preamble
i, k = self._arg(1), self._arg(1)
x = self.file.read(k)
cs, ds = self._arg(4), self._arg(4)
self._pre(i, x, cs, ds)
elif byte == 248: # postamble (just some number of 248s)
break
else:
raise ValueError("unknown vf opcode %d" % byte)
def _init_packet(self, pl):
if self.state != _dvistate.outer:
raise ValueError("Misplaced packet in vf file")
self.h, self.v, self.w, self.x, self.y, self.z = 0, 0, 0, 0, 0, 0
self.stack, self.text, self.boxes = [], [], []
self.f = self._first_font
return self.file.tell() + pl
def _finalize_packet(self, packet_char, packet_width):
self._chars[packet_char] = Page(
text=self.text, boxes=self.boxes, width=packet_width,
height=None, descent=None)
self.state = _dvistate.outer
def _pre(self, i, x, cs, ds):
if self.state != _dvistate.pre:
raise ValueError("pre command in middle of vf file")
if i != 202:
raise ValueError("Unknown vf format %d" % i)
if len(x):
matplotlib.verbose.report('vf file comment: ' + x, 'debug')
self.state = _dvistate.outer
# cs = checksum, ds = design size
def _fix2comp(num):
"""
Convert from two's complement to negative.
"""
assert 0 <= num < 2**32
if num & 2**31:
return num - 2**32
else:
return num
def _mul2012(num1, num2):
"""
Multiply two numbers in 20.12 fixed point format.
"""
# Separated into a function because >> has surprising precedence
return (num1*num2) >> 20
class Tfm(object):
"""
A TeX Font Metric file.
This implementation covers only the bare minimum needed by the Dvi class.
Parameters
----------
filename : string or bytestring
Attributes
----------
checksum : int
Used for verifying against the dvi file.
design_size : int
Design size of the font (unknown units)
width, height, depth : dict
Dimensions of each character, need to be scaled by the factor
specified in the dvi file. These are dicts because indexing may
not start from 0.
"""
__slots__ = ('checksum', 'design_size', 'width', 'height', 'depth')
def __init__(self, filename):
matplotlib.verbose.report('opening tfm file ' + filename, 'debug')
with open(filename, 'rb') as file:
header1 = file.read(24)
lh, bc, ec, nw, nh, nd = \
struct.unpack(str('!6H'), header1[2:14])
matplotlib.verbose.report(
'lh=%d, bc=%d, ec=%d, nw=%d, nh=%d, nd=%d' % (
lh, bc, ec, nw, nh, nd), 'debug')
header2 = file.read(4*lh)
self.checksum, self.design_size = \
struct.unpack(str('!2I'), header2[:8])
# there is also encoding information etc.
char_info = file.read(4*(ec-bc+1))
widths = file.read(4*nw)
heights = file.read(4*nh)
depths = file.read(4*nd)
self.width, self.height, self.depth = {}, {}, {}
widths, heights, depths = \
[struct.unpack(str('!%dI') % (len(x)/4), x)
for x in (widths, heights, depths)]
for idx, char in enumerate(xrange(bc, ec+1)):
byte0 = ord(char_info[4*idx])
byte1 = ord(char_info[4*idx+1])
self.width[char] = _fix2comp(widths[byte0])
self.height[char] = _fix2comp(heights[byte1 >> 4])
self.depth[char] = _fix2comp(depths[byte1 & 0xf])
PsFont = namedtuple('Font', 'texname psname effects encoding filename')
class PsfontsMap(object):
"""
A psfonts.map formatted file, mapping TeX fonts to PS fonts.
Usage::
>>> map = PsfontsMap(find_tex_file('pdftex.map'))
>>> entry = map[b'ptmbo8r']
>>> entry.texname
b'ptmbo8r'
>>> entry.psname
b'Times-Bold'
>>> entry.encoding
'/usr/local/texlive/2008/texmf-dist/fonts/enc/dvips/base/8r.enc'
>>> entry.effects
{'slant': 0.16700000000000001}
>>> entry.filename
Parameters
----------
filename : string or bytestring
Notes
-----
For historical reasons, TeX knows many Type-1 fonts by different
names than the outside world. (For one thing, the names have to
fit in eight characters.) Also, TeX's native fonts are not Type-1
but Metafont, which is nontrivial to convert to PostScript except
as a bitmap. While high-quality conversions to Type-1 format exist
and are shipped with modern TeX distributions, we need to know
which Type-1 fonts are the counterparts of which native fonts. For
these reasons a mapping is needed from internal font names to font
file names.
A texmf tree typically includes mapping files called e.g.
:file:`psfonts.map`, :file:`pdftex.map`, or :file:`dvipdfm.map`.
The file :file:`psfonts.map` is used by :program:`dvips`,
:file:`pdftex.map` by :program:`pdfTeX`, and :file:`dvipdfm.map`
by :program:`dvipdfm`. :file:`psfonts.map` might avoid embedding
the 35 PostScript fonts (i.e., have no filename for them, as in
the Times-Bold example above), while the pdf-related files perhaps
only avoid the "Base 14" pdf fonts. But the user may have
configured these files differently.
"""
__slots__ = ('_font', '_filename')
def __init__(self, filename):
self._font = {}
self._filename = filename
if six.PY3 and isinstance(filename, bytes):
encoding = sys.getfilesystemencoding() or 'utf-8'
self._filename = filename.decode(encoding, errors='replace')
with open(filename, 'rb') as file:
self._parse(file)
def __getitem__(self, texname):
assert isinstance(texname, bytes)
try:
result = self._font[texname]
except KeyError:
fmt = ('A PostScript file for the font whose TeX name is "{0}" '
'could not be found in the file "{1}". The dviread module '
'can only handle fonts that have an associated PostScript '
'font file. '
'This problem can often be solved by installing '
'a suitable PostScript font package in your (TeX) '
'package manager.')
msg = fmt.format(texname.decode('ascii'), self._filename)
msg = textwrap.fill(msg, break_on_hyphens=False,
break_long_words=False)
matplotlib.verbose.report(msg, 'helpful')
raise
fn, enc = result.filename, result.encoding
if fn is not None and not fn.startswith(b'/'):
fn = find_tex_file(fn)
if enc is not None and not enc.startswith(b'/'):
enc = find_tex_file(result.encoding)
return result._replace(filename=fn, encoding=enc)
def _parse(self, file):
"""
Parse the font mapping file.
The format is, AFAIK: texname fontname [effects and filenames]
Effects are PostScript snippets like ".177 SlantFont",
filenames begin with one or two less-than signs. A filename
ending in enc is an encoding file, other filenames are font
files. This can be overridden with a left bracket: <[foobar
indicates an encoding file named foobar.
There is some difference between <foo.pfb and <<bar.pfb in
subsetting, but I have no example of << in my TeX installation.
"""
# If the map file specifies multiple encodings for a font, we
# follow pdfTeX in choosing the last one specified. Such
# entries are probably mistakes but they have occurred.
# http://tex.stackexchange.com/questions/10826/
# http://article.gmane.org/gmane.comp.tex.pdftex/4914
empty_re = re.compile(br'%|\s*$')
word_re = re.compile(
br'''(?x) (?:
"<\[ (?P<enc1> [^"]+ )" | # quoted encoding marked by [
"< (?P<enc2> [^"]+.enc)" | # quoted encoding, ends in .enc
"<<? (?P<file1> [^"]+ )" | # quoted font file name
" (?P<eff1> [^"]+ )" | # quoted effects or font name
<\[ (?P<enc3> \S+ ) | # encoding marked by [
< (?P<enc4> \S+ .enc) | # encoding, ends in .enc
<<? (?P<file2> \S+ ) | # font file name
(?P<eff2> \S+ ) # effects or font name
)''')
effects_re = re.compile(
br'''(?x) (?P<slant> -?[0-9]*(?:\.[0-9]+)) \s* SlantFont
| (?P<extend>-?[0-9]*(?:\.[0-9]+)) \s* ExtendFont''')
lines = (line.strip()
for line in file
if not empty_re.match(line))
for line in lines:
effects, encoding, filename = b'', None, None
words = word_re.finditer(line)
# The named groups are mutually exclusive and are
# referenced below at an estimated order of probability of
# occurrence based on looking at my copy of pdftex.map.
# The font names are probably unquoted:
w = next(words)
texname = w.group('eff2') or w.group('eff1')
w = next(words)
psname = w.group('eff2') or w.group('eff1')
for w in words:
# Any effects are almost always quoted:
eff = w.group('eff1') or w.group('eff2')
if eff:
effects = eff
continue
# Encoding files usually have the .enc suffix
# and almost never need quoting:
enc = (w.group('enc4') or w.group('enc3') or
w.group('enc2') or w.group('enc1'))
if enc:
if encoding is not None:
matplotlib.verbose.report(
'Multiple encodings for %s = %s'
% (texname, psname),
'debug')
encoding = enc
continue
# File names are probably unquoted:
filename = w.group('file2') or w.group('file1')
effects_dict = {}
for match in effects_re.finditer(effects):
slant = match.group('slant')
if slant:
effects_dict['slant'] = float(slant)
else:
effects_dict['extend'] = float(match.group('extend'))
self._font[texname] = PsFont(
texname=texname, psname=psname, effects=effects_dict,
encoding=encoding, filename=filename)
class Encoding(object):
"""
Parses a \\*.enc file referenced from a psfonts.map style file.
The format this class understands is a very limited subset of
PostScript.
Usage (subject to change)::
for name in Encoding(filename):
whatever(name)
Parameters
----------
filename : string or bytestring
Attributes
----------
encoding : list
List of character names
"""
__slots__ = ('encoding',)
def __init__(self, filename):
with open(filename, 'rb') as file:
matplotlib.verbose.report('Parsing TeX encoding ' + filename,
'debug-annoying')
self.encoding = self._parse(file)
matplotlib.verbose.report('Result: ' + repr(self.encoding),
'debug-annoying')
def __iter__(self):
for name in self.encoding:
yield name
def _parse(self, file):
result = []
lines = (line.split(b'%', 1)[0].strip() for line in file)
data = b''.join(lines)
beginning = data.find(b'[')
if beginning < 0:
raise ValueError("Cannot locate beginning of encoding in {}"
.format(file))
data = data[beginning:]
end = data.find(b']')
if end < 0:
raise ValueError("Cannot locate end of encoding in {}"
.format(file))
data = data[:end]
return re.findall(br'/([^][{}<>\s]+)', data)
def find_tex_file(filename, format=None):
"""
Find a file in the texmf tree.
Calls :program:`kpsewhich` which is an interface to the kpathsea
library [1]_. Most existing TeX distributions on Unix-like systems use
kpathsea. It is also available as part of MikTeX, a popular
distribution on Windows.
Parameters
----------
filename : string or bytestring
format : string or bytestring
Used as the value of the `--format` option to :program:`kpsewhich`.
Could be e.g. 'tfm' or 'vf' to limit the search to that type of files.
References
----------
.. [1] `Kpathsea documentation <http://www.tug.org/kpathsea/>`_
The library that :program:`kpsewhich` is part of.
"""
if six.PY3:
# we expect these to always be ascii encoded, but use utf-8
# out of caution
if isinstance(filename, bytes):
filename = filename.decode('utf-8', errors='replace')
if isinstance(format, bytes):
format = format.decode('utf-8', errors='replace')
cmd = [str('kpsewhich')]
if format is not None:
cmd += ['--format=' + format]
cmd += [filename]
matplotlib.verbose.report('find_tex_file(%s): %s'
% (filename, cmd), 'debug')
# stderr is unused, but reading it avoids a subprocess optimization
# that breaks EINTR handling in some Python versions:
# http://bugs.python.org/issue12493
# https://github.com/matplotlib/matplotlib/issues/633
pipe = subprocess.Popen(cmd, stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
result = pipe.communicate()[0].rstrip()
matplotlib.verbose.report('find_tex_file result: %s' % result,
'debug')
return result.decode('ascii')
# With multiple text objects per figure (e.g., tick labels) we may end
# up reading the same tfm and vf files many times, so we implement a
# simple cache. TODO: is this worth making persistent?
_tfmcache = {}
_vfcache = {}
def _fontfile(texname, class_, suffix, cache):
try:
return cache[texname]
except KeyError:
pass
filename = find_tex_file(texname + suffix)
if filename:
result = class_(filename)
else:
result = None
cache[texname] = result
return result
def _tfmfile(texname):
return _fontfile(texname, Tfm, '.tfm', _tfmcache)
def _vffile(texname):
return _fontfile(texname, Vf, '.vf', _vfcache)
if __name__ == '__main__':
import sys
matplotlib.verbose.set_level('debug-annoying')
fname = sys.argv[1]
try:
dpi = float(sys.argv[2])
except IndexError:
dpi = None
with Dvi(fname, dpi) as dvi:
fontmap = PsfontsMap(find_tex_file('pdftex.map'))
for page in dvi:
print('=== new page ===')
fPrev = None
for x, y, f, c, w in page.text:
if f != fPrev:
print('font', f.texname, 'scaled', f._scale/pow(2.0, 20))
fPrev = f
print(x, y, c, 32 <= c < 128 and chr(c) or '.', w)
for x, y, w, h in page.boxes:
print(x, y, 'BOX', w, h)
|