#! /usr/local/bin/python
# Convert GNU texinfo files into HTML, one file per node.
# Based on Texinfo 2.14.
# Usage: texi2html [-d] [-d] [-c] inputfile outputdirectory
# The input file must be a complete texinfo file, e.g. emacs.texi.
# This creates many files (one per info node) in the output directory,
# overwriting existing files of the same name. All files created have
# ".html" as their extension.
# XXX To do:
# - handle @comment*** correctly
# - handle @xref {some words} correctly
# - handle @ftable correctly (items aren't indexed?)
# - handle @itemx properly
# - handle @exdent properly
# - add links directly to the proper line from indices
# - check against the definitive list of @-cmds; we still miss (among others):
# - @defindex (hard)
# - @c(omment) in the middle of a line (rarely used)
# - @this* (not really needed, only used in headers anyway)
# - @today{} (ever used outside title page?)
# More consistent handling of chapters/sections/etc.
# Lots of documentation
# Many more options:
# -top designate top node
# -links customize which types of links are included
# -split split at chapters or sections instead of nodes
# -name Allow different types of filename handling. Non unix systems
# will have problems with long node names
# ...
# Support the most recent texinfo version and take a good look at HTML 3.0
# More debugging output (customizable) and more fexible error handling
# How about icons ?
import os
import regex
import regsub
import string
MAGIC = '\\input texinfo'
cmprog = regex.compile('^@\([a-z]+\)\([ \t]\|$\)') # Command (line-oriented)
blprog = regex.compile('^[ \t]*$') # Blank line
kwprog = regex.compile('@[a-z]+') # Keyword (embedded, usually with {} args)
spprog = regex.compile('[\n@{}&<>]') # Special characters in running text
miprog = regex.compile( \
'^\* \([^:]*\):\(:\|[ \t]*\([^\t,\n.]+\)\([^ \t\n]*\)\)[ \t\n]*')
# menu item (Yuck!)
class Node:
__doc__ = """
Some of the parser's functionality is separated into this class.
A Node accumulates its contents, takes care of links to other Nodes
and saves itself when it is finished and all links are resolved. """
def __init__ (self, dir, name, topname, title, next, prev, up):
self.dirname = dir
self.name = name
if topname:
self.topname = topname
else:
self.topname = name
self.title = title
self.next = next
self.prev = prev
self.up = up
self.lines = []
self.type = 0
self.cont = ''
def write (self, *lines):
map(self.lines.append, lines)
def flush (self):
fp = open (self.dirname + '/' + makefile(self.name), 'w')
fp.write (self.prologue)
fp.write (self.text)
fp.write (self.epilogue)
fp.close ()
def link(self, label, nodename):
if nodename:
if string.lower(nodename) == '(dir)':
addr = '../dir.html'
else:
addr = makefile(nodename)
self.write(label, ': ', nodename, ' \n')
def finalize(self):
length = len (self.lines)
self.text = string.joinfields (self.lines, '')
self.lines = []
self.write ('
\n')
if self.cont != self.next:
self.link('Cont', self.cont)
self.link('Next', self.next)
self.link('Prev', self.prev)
self.link('Up', self.up)
if self.name <> self.topname:
self.link('Top', self.topname)
self.write ('
\n')
links = string.joinfields(self.lines, '')
self.lines = []
self.prologue = ('\n'
'\n'
'\n'
' ' + self.title + '\n'
'\n' + \
links)
if length > 20:
self.epilogue = '\n%s\n' % links
else:
self.epilogue = '
\n'
class TexinfoParser:
# Initialize an instance
def __init__(self):
self.unknown = {} # statistics about unknown @-commands
self.filenames = {} # Check for identical filenames
self.debugging = 0 # larger values produce more output
self.nodefp = None # open file we're writing to
self.nodelineno = 0 # Linenumber relative to node
self.links = None # Links from current node
self.savetext = None # If not None, save text head instead
self.savestack = [] # If not None, save text head instead
self.dirname = 'tmp' # directory where files are created
self.includedir = '.' # directory to search @include files
self.nodename = '' # name of current node
self.topname = '' # name of top node (first node seen)
self.title = '' # title of this whole Texinfo tree
self.resetindex() # Reset all indices
self.contents = [] # Reset table of contents
self.numbering = [] # Reset section numbering counters
self.nofill = 0 # Normal operation: fill paragraphs
self.values={'html': 1} # Names that should be parsed in ifset
self.stackinfo={} # Keep track of state in the stack
# XXX The following should be reset per node?!
self.footnotes = [] # Reset list of footnotes
self.itemarg = None # Reset command used by @item
self.itemnumber = None # Reset number for @item in @enumerate
self.itemindex = None # Reset item index name
self.node = None
self.nodestack = []
self.cont = 0
self.includedepth = 0
# Set (output) directory name
def setdirname(self, dirname):
self.dirname = dirname
# Set include directory name
def setincludedir(self, includedir):
self.includedir = includedir
# Parse the contents of an entire file
def parse(self, fp):
line = fp.readline()
lineno = 1
while line and (line[0] == '%' or blprog.match(line) >= 0):
line = fp.readline()
lineno = lineno + 1
if line[:len(MAGIC)] <> MAGIC:
raise SyntaxError, 'file does not begin with '+`MAGIC`
self.parserest(fp, lineno)
# Parse the contents of a file, not expecting a MAGIC header
def parserest(self, fp, initial_lineno):
lineno = initial_lineno
self.done = 0
self.skip = 0
self.stack = []
accu = []
while not self.done:
line = fp.readline()
self.nodelineno = self.nodelineno + 1
if not line:
if accu:
if not self.skip: self.process(accu)
accu = []
if initial_lineno > 0:
print '*** EOF before @bye'
break
lineno = lineno + 1
if cmprog.match(line) >= 0:
a, b = cmprog.regs[1]
cmd = line[a:b]
if cmd in ('noindent', 'refill'):
accu.append(line)
else:
if accu:
if not self.skip:
self.process(accu)
accu = []
self.command(line)
elif blprog.match(line) >= 0 and \
'format' not in self.stack and \
'example' not in self.stack:
if accu:
if not self.skip:
self.process(accu)
self.write('