draft.py

#!/usr/bin/python
# -*- python -*-

"""
NAME
	%(program)s - Extract meta-information from an IETF draft.

SYNOPSIS
	%(program)s [OPTIONS] DRAFTLIST_FILE

DESCRIPTION
        Extract information about authors' names and email addresses,
        intended status and number of pages from Internet Drafts.
        The information is emitted in the form of a line containing
        xml-style attributes, prefixed with the name of the draft.

%(options)s

AUTHOR
	Written by Henrik Levkowetz, <henrik@levkowetz.com>

COPYRIGHT
	Copyright 2008 Henrik Levkowetz

	This program is free software; you can redistribute it and/or modify
	it under the terms of the GNU General Public License as published by
	the Free Software Foundation; either version 2 of the License, or (at
	your option) any later version. There is NO WARRANTY; not even the
	implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
	PURPOSE. See the GNU General Public License for more details.

"""

import datetime
import getopt
import os
import os.path
import re
import stat
import sys
import time

version = "0.19"
program = os.path.basename(sys.argv[0])
progdir = os.path.dirname(sys.argv[0])

# ----------------------------------------------------------------------
# Data
# ----------------------------------------------------------------------


opt_debug = False
opt_timestamp = False
opt_trace = False

# The following is an alias list for short forms which starts with a
# different letter than the long form.

longform = {
    "Beth": "Elizabeth",
    "Bill": "William",
    "Bob": "Robert",
    "Dick": "Richard",
    "Fred": "Alfred",
    "Jerry": "Gerald",
    "Liz": "Elizabeth",
    "Lynn": "Carolyn",
    "Ned": "Edward",
    "Ted":"Edward",
}
longform = dict([ (short+" ", longform[short]+" ") for short in longform ])


# ----------------------------------------------------------------------
# Functions
# ----------------------------------------------------------------------
def _debug(string):
    if opt_debug:
        sys.stderr.write("%s\n" % (string))

# ----------------------------------------------------------------------
def _note(string):
    sys.stdout.write("%s: %s\n" % (program, string))
    
# ----------------------------------------------------------------------
def _warn(string):
    sys.stderr.write("%s: Warning: %s\n" % (program, string))
    
# ----------------------------------------------------------------------
def _err(string):
    sys.stderr.write("%s: Error: %s\n" % (program, string))
    sys.exit(1)

# ----------------------------------------------------------------------
def _gettext(file):
    file = open(file)
    text = file.read()
    file.close()

    text = re.sub(".\x08", "", text)    # Get rid of inkribbon backspace-emphasis
    text = text.replace("\r\n", "\n")   # Convert DOS to unix
    text = text.replace("\r", "\n")     # Convert MAC to unix
    text = text.expandtabs()
    text = text.strip()

    return text

def acronym_match(s, l):
    acronym = re.sub("[^A-Z]", "", l)
    #_debug(" s:%s; l:%s => %s; %s" % (s, l, acronym, s==acronym)) 
    return s == acronym

# ----------------------------------------------------------------------

class Draft():

    def __init__(self, text):
        self.rawtext = text

        text = re.sub(".\x08", "", text)    # Get rid of inkribbon backspace-emphasis
        text = text.replace("\r\n", "\n")   # Convert DOS to unix
        text = text.replace("\r", "\n")     # Convert MAC to unix
        text = text.strip()
        self.text = text
        self.errors = {}

        self.rawlines = self.text.split("\n")
        self.lines, self.pages = self._stripheaders()

        # Some things (such as the filename) has to be on the first page.  If
        # we didn't get back a set of pages, only one single page with the
        # whole document, then we need to do an enforced page split in order
        # to limit later searches to the first page.
        if len(self.pages) <= 1:
            self.pages = []
            for pagestart in range(0, len(self.lines), 58):
                self.pages += [ "\n".join(self.lines[pagestart:pagestart+54]) ]

        self.filename, self.revision = self._parse_draftname()
        
        self._authors = None
        self._author_info = None
        self._abstract = None
        self._pagecount = None
        self._status = None
        self._creation_date = None
        self._title = None

    # ------------------------------------------------------------------
    def _parse_draftname(self):
        draftname_regex = r"(draft-[a-z0-9-]*)-(\d\d)(\w|\.txt|\n|$)"
        draftname_match = re.search(draftname_regex, self.pages[0])
        if draftname_match:
            return (draftname_match.group(1), draftname_match.group(2) )
        else:
            self.errors["draftname"] = "Could not find the draft name and revision on the first page."
            return ("", "")

    # ----------------------------------------------------------------------
    def _stripheaders(self):
        stripped = []
        pages = []
        page = []
        line = ""
        debug = False
        newpage = False
        sentence = False
        blankcount = 0
        linecount = 0
        # two functions with side effects
        def endpage(pages, page, newpage, line):
            if line:
                page += [ line ]
            return begpage(pages, page, newpage)
        def begpage(pages, page, newpage, line=None):
            if page and len(page) > 5:
                pages += [ "\n".join(page) ]
                page = []
                newpage = True
            if line:
                page += [ line ]
            return pages, page, newpage
        for line in self.rawlines:
            linecount += 1
            line = line.rstrip()
            if re.search("\[?[Pp]age [0-9ivx]+\]?[ \t\f]*$", line, re.I):
                pages, page, newpage = endpage(pages, page, newpage, line)
                continue
            if re.search("\f", line, re.I):
                pages, page, newpage = begpage(pages, page, newpage)
                continue
            if re.search("^ *Internet.Draft.+[12][0-9][0-9][0-9] *$", line, re.I):
                pages, page, newpage = begpage(pages, page, newpage, line)
                continue
    #        if re.search("^ *Internet.Draft  +", line, re.I):
    #            newpage = True
    #            continue
            if re.search("^ *Draft.+[12][0-9][0-9][0-9] *$", line, re.I):
                pages, page, newpage = begpage(pages, page, newpage, line)
                continue
            if re.search("^RFC[ -]?[0-9]+.*(  +)[12][0-9][0-9][0-9]$", line, re.I):
                pages, page, newpage = begpage(pages, page, newpage, line)
                continue
            if re.search("^draft-[-a-z0-9_.]+.*[0-9][0-9][0-9][0-9]$", line, re.I):
                pages, page, newpage = endpage(pages, page, newpage, line)
                continue
            if linecount > 15 and re.search(".{58,}(Jan|Feb|Mar|March|Apr|April|May|Jun|June|Jul|July|Aug|Sep|Oct|Nov|Dec) (19[89][0-9]|20[0-9][0-9]) *$", line, re.I):
                pages, page, newpage = begpage(pages, page, newpage, line)
                continue
            if newpage and re.search("^ *draft-[-a-z0-9_.]+ *$", line, re.I):
                pages, page, newpage = begpage(pages, page, newpage, line)
                continue
            if re.search("^[^ \t]+", line):
                sentence = True
            if re.search("[^ \t]", line):
                if newpage:
                    if sentence:
                        stripped += [""]
                else:
                    if blankcount:
                        stripped += [""]*blankcount
                blankcount = 0
                sentence = False
                newpage = False
            if re.search("[.:]$", line):
                sentence = True
            if re.search("^[ \t]*$", line):
                blankcount += 1
                page += [ line ]
                continue
            page += [ line ]
            stripped += [ line ]
        pages, page, newpage = begpage(pages, page, newpage)
        return stripped, pages

    # ----------------------------------------------------------------------
    def get_pagecount(self):
        if self._pagecount == None:
            self._pagecount = len(re.findall("\[[Pp]age [0-9ixldv]+\]", self.text)) or len(self.lines)/58
        return self._pagecount

    # ----------------------------------------------------------------------
    def get_status(self):
        if self._status == None:
            for line in self.lines[:10]:
                status_match = re.search("^\s*Intended [Ss]tatus:\s*(.*?)   ", line)
                if status_match:
                    self._status = status_match.group(1)
                    break
        return self._status

    # ------------------------------------------------------------------
    def get_creation_date(self):
        if self._creation_date:
            return self._creation_date
        month_names = [ 'jan', 'feb', 'mar', 'apr', 'may', 'jun', 'jul', 'aug', 'sep', 'oct', 'nov', 'dec' ]
        date_regexes = [
            r'^(?P<month>\w+)\s+(?P<day>\d{1,2}),?\s+(?P<year>\d{4})',
            r'^(?P<day>\d{1,2}),?\s+(?P<month>\w+)\s+(?P<year>\d{4})',
            r'^(?P<day>\d{1,2})-(?P<month>\w+)-(?P<year>\d{4})',
            r'^(?P<month>\w+)\s+(?P<year>\d{4})',
            r'\s{3,}(?P<month>\w+)\s+(?P<day>\d{1,2}),?\s+(?P<year>\d{4})',
            r'\s{3,}(?P<day>\d{1,2}),?\s+(?P<month>\w+)\s+(?P<year>\d{4})',
            r'\s{3,}(?P<day>\d{1,2})-(?P<month>\w+)-(?P<year>\d{4})',
            # 'October 2008' - default day to today's.
            r'\s{3,}(?P<month>\w+)\s+(?P<year>\d{4})',
        ]

        for regex in date_regexes:
            match = re.search(regex, self.pages[0], re.MULTILINE)
            if match:
                md = match.groupdict()
                mon = md['month'][0:3].lower()
                day = int( md.get( 'day', 0 ) )
                year = int( md['year'] )
                try:
                    month = month_names.index( mon ) + 1
                    today = datetime.date.today()
                    if day==0:
                        # if the date was given with only month and year, use
                        # today's date if month and year is today's month and
                        # year, otherwise pick the middle of the month.
                        # Don't use today's day for month and year in the past
                        if month==today.month and year==today.year:
                            day = today.day
                        else:
                            day = 15
                    self._creation_date = datetime.date(year, month, day)
                    return self._creation_date
                except ValueError:
                    # mon abbreviation not in _MONTH_NAMES
                    # or month or day out of range
                    pass
        self.errors['creation_date'] = 'Creation Date field is empty or the creation date is not in a proper format.'
        return self._creation_date


    # ------------------------------------------------------------------
    def get_abstract(self):
        if self._abstract:
            return self._abstract
        abstract_re = re.compile('^(\s*)abstract', re.I)
        header_re = re.compile("^(\s*)([0-9]+\.? |Appendix|Status of|Table of|Full Copyright|Copyright|Intellectual Property|Acknowled|Author|Index|Disclaimer).*", re.I)
        begin = False
        abstract = []
        abstract_indent = 0
        look_for_header = False
        for line in self.lines:
            if not begin and abstract_re.match(line):
                begin=True
                abstract_indent = len(abstract_re.match(line).group(0))
                continue
            if begin:
                if not line and not abstract:
                    continue
                if not line:
                    look_for_header=True
                    abstract.append(line)
                    continue
                if look_for_header and header_re.match(line):
                    break
                look_for_header = False
                abstract.append(line)
        abstract = '\n'.join(abstract)
        abstract = self._clean_abstract(abstract)
        self._abstract = self._check_abstract_indent(abstract, abstract_indent)
        return self._abstract


    def _check_abstract_indent(self, abstract, indent):
        indentation_re = re.compile('^(\s)*')
        indent_lines = []
        for line in abstract.split('\n'):
            if line:
                indent = len(indentation_re.match(line).group(0))
                indent_lines.append(indent)
        percents = {}
        total = float(len(indent_lines))
        formated = False
        for indent in set(indent_lines):
            count = indent_lines.count(indent)/total
            percents[indent] = count
            if count > 0.9:
                formated = True
        if not formated:
            return abstract
        new_abstract = []
        for line in abstract.split('\n'):
            if line:
                indent = len(indentation_re.match(line).group(0))
                if percents[indent] < 0.9:
                    break
            new_abstract.append(line)
        return '\n'.join(new_abstract)


    def _clean_abstract(self, text):
        text = re.sub("(?s)(Conventions [Uu]sed in this [Dd]ocument|Requirements [Ll]anguage)?[\n ]*The key words \"MUST\", \"MUST NOT\",.*$", "", text)
        # Get rid of status/copyright boilerplate
        text = re.sub("(?s)\nStatus of [tT]his Memo\n.*$", "", text)
        # wrap long lines without messing up formatting of Ok paragraphs:
        while re.match("([^\n]{72,}?) +", text):
            text = re.sub("([^\n]{72,}?) +([^\n ]*)(\n|$)", "\\1\n\\2 ", text)
        return text


    # ------------------------------------------------------------------
    def get_authors(self):
        """Returns a list of strings with author name and email within angle brackets"""
        if self._authors == None:
            self.extract_authors()
        return self._authors

    def get_author_info(self):
        """Returns a list of tuples, with each tuple containing (given_names, surname, email).  Email will be None if unknown."""
        if self._author_info == None:
            self.extract_authors()
        return self._author_info

    def extract_authors(self):
        """Extract author information from draft text.

        """
        aux = {
            "honor" : r"(?:Dr\.?|Prof(?:\.?|essor)|Sir|Lady|Dame|Sri)",
            "prefix": r"([Dd]e|Hadi|van|van de|van der|Ver|von|[Ee]l)",
            "suffix": r"(jr.?|Jr.?|II|2nd|III|3rd|IV|4th)",
            "first" : r"([A-Z][-A-Za-z]*)((\.?[- ]{1,2}[A-Za-z]+)*)",
            "last"  : r"([-A-Za-z']{2,})",
            }
        authformats = [
            r" {6}(%(first)s[ \.]{1,3}((%(prefix)s )?%(last)s)( %(suffix)s)?)([, ]?(.+\.?|\(.+\.?|\)))?$" % aux,
            r" {6}(((%(prefix)s )?%(last)s)( %(suffix)s)?, %(first)s)?$" % aux,
            r" {6}(%(last)s)$" % aux,
            ]
        multiauthformats = [
            (
                r" {6}(%(first)s[ \.]{1,3}((%(prefix)s )?%(last)s)( %(suffix)s)?)(, ?%(first)s[ \.]{1,3}((%(prefix)s )?%(last)s)( %(suffix)s)?)+$" % aux,
                r"(%(first)s[ \.]{1,3}((%(prefix)s )?%(last)s)( %(suffix)s)?)" % aux
            ),
        ]
        editorformats = [
            r"(?:, | )([Ee]d\.?|\([Ee]d\.?\)|[Ee]ditor)$",
            ]

        ignore = [
            "Standards Track", "Current Practice", "Internet Draft", "Working Group",
            "No Affiliation", 
            ]
        # group       12                   34            5            6
        authors = []
        author_info = []
        companies = []

        # Collect first-page author information first
        have_blankline = False
        have_draftline = False
        prev_blankline = False
        for line in self.lines[:15]:
            #_debug( "**" + line)
            leading_space = len(re.findall("^ *", line)[0])
            line_len = len(line.rstrip())
            trailing_space = line_len <= 72 and 72 - line_len or 0
            # Truncate long lines at the first space past column 80:
            trunc_space = line.find(" ", 80)
            if line_len > 80 and  trunc_space > -1:
                line = line[:trunc_space]
            if line_len > 60:
                # Look for centered title, break if found:
                if (leading_space > 5 and abs(leading_space - trailing_space) < 5):
                    _debug("Breaking for centered line")
                    break
                for editorformat in editorformats:
                    if re.search(editorformat, line):
                        line = re.sub(editorformat, "", line)
                        break
                for lineformat, authformat in multiauthformats:
                    match = re.search(lineformat, line)
                    if match:
                        #_debug("Multiauth format: '%s'" % lineformat)
                        author_list = re.findall(authformat, line)
                        authors += [ a[0] for a in author_list ]
                        #_debug("\nLine:   " + line)
                        #_debug("Format: " + authformat)
                        for author in author_list:
                            _debug("Author: '%s'" % author[0])
                else:
                    for authformat in authformats:                        
                        match = re.search(authformat, line)
                        if match:
                            #_debug("Auth format: '%s'" % authformat)
                            author = match.group(1)
                            authors += [ author ]
                            #_debug("\nLine:   " + line)
                            #_debug("Format: " + authformat)
                            _debug("Author: '%s'" % author)
            if line.strip() == "":
                if prev_blankline and authors:
                    break
                have_blankline = True
                prev_blankline = True
            else:
                prev_blankline = False
            if "draft-" in line:
                have_draftline = True
            if have_blankline and have_draftline:
                break

        found_pos = []
        for i in range(len(authors)):
            _debug("1: authors[%s]: %s" % (i, authors[i]))
            author = authors[i]
            if author == None:
                continue
            suffix_match = re.search(" %(suffix)s$" % aux, author)
            if suffix_match:
                suffix = suffix_match.group(1)
                author = author[:-len(suffix)].strip()
            else:
                suffix = None
            if "," in author:
                last, first = author.split(",",1)
                author = "%s %s" % (first.strip(), last.strip())
            if not " " in author:
                if "." in author:
                    first, last = author.rsplit(".", 1)
                    first += "."
                else:
                    author = "[A-Z].+ " + author
                    first, last = author.rsplit(" ", 1)
            else:
                first, last = author.rsplit(" ", 1)
            prefix_match = re.search(" %(prefix)s$" % aux, first)
            if prefix_match:
                prefix = prefix_match.group(1)
                first = first[:-len(prefix)].strip()
                last = prefix+" "+last
            _debug("First, Last: '%s' '%s'" % (first, last))
            for firstname, surname, casefixname in [ (first,last,last), (last,first,first), (first,last,last.upper()), (last,first,first.upper()), ]:
                author = "%s %s" % (firstname, casefixname)
                _debug("\nAuthors: "+str(authors))
                _debug("Author: "+author)
                # Pattern for full author information search, based on first page author name:
                authpat = author
                # Permit expansion of first name
                authpat = re.sub("\. ", ".* ", authpat)
                authpat = re.sub("\.$", ".*", authpat)
                # Permit insertsion of middle name or initial
                authpat = re.sub(" ", "\S*( +[^ ]+)* +", authpat)
                # Permit expansion of double-name initials
                if not "[A-Z]" in authpat:
                    authpat = re.sub("-", ".*?-", authpat)
                # Some chinese names are shown with double-letter(latin) abbreviated given names, rather than
                # a single-letter(latin) abbreviation:
                authpat = re.sub("^([A-Z])[A-Z]+\.\*", r"\1[-\w]+", authpat) 
                authpat = "(?:^| and )(?:%s ?)?(%s)( *\(.*\)|,( [A-Z][-A-Za-z0-9]*)?| %s| [A-Z][a-z]+)?" % (aux["honor"], authpat, aux["suffix"])
                _debug("Authpat: " + authpat)
                start = 0
                col = None
                # Find start of author info for this author (if any).
                # Scan from the end of the file, looking for a match to  authpath
                # Scan towards the front from the end of the file, looking for a match to authpath
                for j in range(len(self.lines)-1, 15, -1):
                    line = self.lines[j]
                    forms = [ line ] + [ line.replace(short, longform[short]) for short in longform if short in line ]
                    for form in forms:
                        try:
                            if re.search(authpat, form.strip()) and not j in found_pos:
                                start = j
                                found_pos += [ start ]
                                _debug( " ==> start %s, normalized '%s'" % (start, form.strip()))
                                # The author info could be formatted in multiple columns...
                                columns = re.split("(    +| and )", form)
                                # _debug( "Columns:" + columns; sys.stdout.flush())
                                # Find which column:
                                #_debug( "Col range:" + range(len(columns)); sys.stdout.flush())

                                cols = [ c for c in range(len(columns)) if re.search(authpat+r"( and |, |$)", columns[c].strip()) ]
                                if cols:
                                    col = cols[0]
                                    if not (start, col) in found_pos:
                                        found_pos += [ (start, col) ]
                                        _debug( "Col:   %d" % col)
                                        beg = len("".join(columns[:col]))
                                        _debug( "Beg:   %d '%s'" % (beg, "".join(columns[:col])))
                                        _debug( "Len:   %d" % len(columns))
                                        if col == len(columns) or col == len(columns)-1:
                                            end = None
                                            _debug( "End1:  %s" % end)
                                        else:
                                            end = beg + len("".join(columns[col:col+2]))
                                            _debug( "End2:  %d '%s'" % (end, "".join(columns[col:col+2])))
                                        _debug( "Cut:   '%s'" % form[beg:end])
                                        author_match = re.search(authpat, columns[col].strip()).group(1)
                                        _debug( "AuthMatch: '%s'" % (author_match,))
                                        if author_match in companies:
                                            authors[i] = None
                                        else:
                                            if casefixname in author_match:
                                                fullname = author_match.replace(casefixname, surname)
                                            else:
                                                fullname = author_match
                                            fullname = re.sub(" +", " ", fullname)
                                            given_names, surname = fullname.rsplit(None, 1)
                                            if " " in given_names:
                                                first, middle = given_names.split(None, 1)
                                            else:
                                                first = given_names
                                                middle = None
                                            names = (first, middle, surname, suffix)
                                            if suffix:
                                                fullname = fullname+" "+suffix
                                            if not " ".join([ n for n in names if n ]) == fullname:
                                                _err("Author tuple doesn't match text in draft: %s: %s %s" % (authors[i], names, fullname))
                                            authors[i] = (fullname, first, middle, surname, suffix)
                                        #_debug( "Author: %s: %s" % (author_match, authors[author_match]))
                                        break
                        except AssertionError, e:
                            sys.stderr.write("filename: "+self.filename+"\n")
                            sys.stderr.write("authpat: "+authpat+"\n")
                            raise
                    if start and col != None:
                        break
                if start and col != None:
                    break
            if not authors[i]:
                continue
            _debug("2: authors[%s]: %s" % (i, authors[i]))
            if start and col != None:
                _debug("\n * %s" % (authors[i], ))
                done = False
                nonblank_count = 0
                keyword = False
                blanklines = 0
                email = None
                for line in self.lines[start+1:]:
                    _debug( "       " + line.strip())
                    # Break on the second blank line
                    if not line:
                        blanklines += 1
                        if blanklines >= 3:
                            _debug( " - Break on blanklines")
                            break
                        else:
                            continue
                    else:
                        nonblank_count += 1                    

                    # Maybe break on author name
    #                 _debug("Line: %s"%line.strip())
    #                 for a in authors:
    #                     if a and a not in companies:
    #                         _debug("Search for: %s"%(r"(^|\W)"+re.sub("\.? ", ".* ", a)+"(\W|$)"))
                    authmatch = [ a for a in authors[i+1:] if a and not a.lower() in companies and (re.search((r"(?i)(^|\W)"+re.sub("\.? ", ".* ", a)+"(\W|$)"), line.strip()) or acronym_match(a, line.strip()) )]
                    if authmatch:
                        _debug("     ? Other author or company ?  : %s" % authmatch)
                        _debug("     Line: "+line.strip())
                        if nonblank_count == 1 or (nonblank_count == 2 and not blanklines):
                            # First line after an author -- this is a company
                            companies += [ c.lower() for c in authmatch ]
                            companies += [ line.strip().lower() ] # XXX fix this for columnized author list
                            companies = list(set(companies))
                            _debug("       -- Companies: " + ", ".join(companies))
                            for k in range(i+1, len(authors)):
                                if authors[k] and authors[k].lower() in companies:
                                    authors[k] = None
                        elif not "@" in line:
                            # Break on an author name
                            _debug( " - Break on other author name")
                            break
                        else:
                            pass

                    try:
                        column = line[beg:end].strip()
                    except:
                        column = line
                    column = re.sub(" *\(at\) *", "@", column)
                    column = re.sub(" *\(dot\) *", ".", column)
                    column = re.sub(" +at +", "@", column)
                    column = re.sub(" +dot +", ".", column)

    #                 if re.search("^\w+: \w+", column):
    #                     keyword = True
    #                 else:
    #                     if keyword:
    #                         # Break on transition from keyword line to something else
    #                         _debug( " - Break on end of keywords")
    #                         break

                    #_debug( "  Column text :: " + column)
                    _debug("3: authors[%s]: %s" % (i, authors[i]))

                    emailmatch = re.search("[-A-Za-z0-9_.+]+@[-A-Za-z0-9_.]+", column)
                    if emailmatch and not "@" in author:
                        email = emailmatch.group(0).lower()
                        break
                authors[i] = authors[i] + ( email, )
            else:
                authors[i] = None
                if not author in ignore:
                    _debug("Not an author? '%s'" % (author))

        authors = [ a for a in authors if a != None ]
        _debug(" * Final author tuples: %s" % (authors,))
        self._author_info = authors        
        self._authors = [ "%s <%s>"%(full,email) if email else full for full,first,middle,last,suffix,email in authors ]
        self._authors.sort()
        _debug(" * Final author list: " + ", ".join(self._authors))
        _debug("-"*72)

    # ------------------------------------------------------------------
    def get_title(self):
        if self._title:
            return self._title
        match = re.search('(?:\n\s*\n\s*)((.+\n){0,2}(.+\n*))(\s+<?draft-\S+\s*\n)\s*\n', self.pages[0])
        if not match:
            match = re.search('(?:\n\s*\n\s*)<?draft-\S+\s*\n*((.+\n){1,3})\s*\n', self.pages[0])
        if not match:
            match = re.search('(?i)(.+\n|.+\n.+\n)(\s*status of this memo\s*\n)', self.pages[0])
        if match:
            title = match.group(1)
            title = title.strip()
            title = re.sub('\s*\n\s*', ' ', title)
            title = re.sub(' +', ' ', title)
            self._title = title
            return self._title
        self.errors["title"] = "Could not find the title on the first page."

    # ------------------------------------------------------------------
    def get_refs(self):
        refs = []
        normrefs = []
        rfcrefs = []
        refline = None
        for i in range(len(self.lines)-1, 15, -1):
            if re.search(r"(?i)^ *[0-9.]+ *(((normative|informative|informational|non-normative) )?references|references\W+(normative|informative))", self.lines[i]):
                refline = i
                break
        if refline:
            for i in range(refline, len(self.lines)):
                line = self.lines[i].strip()
                ref_match = re.search(r"(?i)^\[[a-z0-9.-]+( [a-z0-9.-]+)?\].+", line)
                if ref_match:
                    para = line
                    while True:
                        i += 1
                        if i >= len(self.lines):
                            break
                        line = self.lines[i].strip()
                        if not line:
                            break
                        if para[-1] not in ["-", "/"]:
                            para += " "
                        para += line
                    refs += [ para ]
                    rfc_match = re.search("(?i)rfc ?\d+", para)
                    if rfc_match:
                        rfc = rfc_match.group(0).replace(" ","").lower()
                        rfcrefs += [ rfc ]
        normrefs = list(set(normrefs))
        normrefs.sort()
        rfcrefs = list(set(rfcrefs))
        rfcrefs.sort()
        refs = list(set(refs))
        refs.sort()
        return normrefs, rfcrefs, refs


# ----------------------------------------------------------------------

def getmeta(fn):
    # Initial values
    fields = {}
    fields["eventsource"] = "draft"

    if " " in fn or not fn.endswith(".txt"):
        _warn("Skipping unexpected draft name: '%s'" % (fn))
        return

    if os.path.exists(fn):
        filename = fn
        fn = os.path.basename(fn)
    elif fn.lower().startswith('rfc'):
        filename = os.path.join("/www/tools.ietf.org/rfc", fn)
    else:
        filename = os.path.join("/www/tools.ietf.org/id", fn)
    if not os.path.exists(filename):
        _warn("Could not find file: '%s'" % (filename))
        return

    timestamp = time.strftime("%Y-%m-%dT%H:%M:%S+00:00", time.gmtime(os.stat(filename)[stat.ST_MTIME]))
    text = _gettext(filename)
    draft = Draft(text)
    #_debug("\n".join(draft.lines))

    fields["eventdate"] = timestamp
    if draft.filename:
        fields["doctag"] = draft.filename
    fields["docrev"] = draft.revision

    fields["doctitle"] = draft.get_title()
    fields["docpages"] = str(draft.get_pagecount())
    fields["docauthors"] = ", ".join(draft.get_authors())
#    fields["docauthinfo"] = str(draft.get_author_info())
    normrefs, rfcrefs, refs = draft.get_refs()
    fields["docrfcrefs"] = ", ".join(rfcrefs)
    fields["doccreationdate"] = str(draft.get_creation_date())
    deststatus = draft.get_status()
    if deststatus:
        fields["docdeststatus"] = deststatus
    abstract = draft.get_abstract()
    if abstract:
        fields["docabstract"] = abstract

    return fields


# ----------------------------------------------------------------------
def _output(docname, fields, outfile=sys.stdout):
    if opt_timestamp:
        outfile.write("%s " % (fields["eventdate"]))
    outfile.write("%s" % (os.path.basename(docname.strip())))

    def outputkey(key, fields):
        outfile.write(" %s='%s'" % ( key.lower(), fields[key].strip().replace("\\", "\\\\" ).replace("'", "\\x27" ).replace("\n", "\\n")))

    keys = fields.keys()
    keys.sort()
    for key in keys:
        if fields[key] and not key in ["eventdate", ]:
            outputkey(key, fields)
    outfile.write("\n")

# ----------------------------------------------------------------------
def _printmeta(timestamp, fn, outfile=sys.stdout):
    if opt_trace:
        t = time.time()
        sys.stderr.write("%-58s" % fn[:-4])

    fields = getmeta(fn)
    _output(fields.get("doctag", fn[:-7]), fields, outfile)

    if opt_trace:
        sys.stderr.write("%5.1f\n" % ((time.time() - t)))


# ----------------------------------------------------------------------
# Main
# ----------------------------------------------------------------------

def _main(outfile=sys.stdout):
    global opt_debug, opt_timestamp, opt_trace, files
    # set default values, if any
    # ----------------------------------------------------------------------
    # Option processing
    # ----------------------------------------------------------------------
    options = ""
    for line in re.findall("\n +(if|elif) +opt in \[(.+)\]:\s+#(.+)\n", open(sys.argv[0]).read()):
        if not options:
            options += "OPTIONS\n"
        options += "        %-16s %s\n" % (line[1].replace('"', ''), line[2])
    options = options.strip()

    # with ' < 1:' on the next line, this is a no-op:
    if len(sys.argv) < 1:
        vars = globals()
        vars.update(locals())
        print __doc__ % vars
        sys.exit(1)

    try:
        opts, files = getopt.gnu_getopt(sys.argv[1:], "dhtTv", ["debug", "help", "timestamp", "trace", "version",])
    except Exception, e:
        print "%s: %s" % (program, e)
        sys.exit(1)

    # parse options
    for opt, value in opts:
        if   opt in ["-d", "--debug"]:  # Output debug information
            opt_debug = True
        elif opt in ["-h", "--help"]:   # Output this help text, then exit
            vars = globals()
            vars.update(locals())
            print __doc__ % vars
            sys.exit(1)
        elif opt in ["-v", "--version"]: # Output version information, then exit
            print program, version
            sys.exit(0)
        elif opt in ["-t", "--timestamp"]: # Emit leading timestamp information 
            opt_timestamp = True
        elif opt in ["-T", "--trace"]: # Emit trace information while working
            opt_trace = True

    if not files:
        files = [ "-" ]

    for file in files:
        _debug( "Reading drafts from '%s'" % file)
        if file == "-":
            file = sys.stdin
        elif file.endswith(".gz"):
            file = gzip.open(file)
        else:
            file = open(file)

        if os.path.exists(file.name):
            timestamp = time.strftime("%Y-%m-%dT%H:%M:%S+00:00", time.gmtime(os.stat(file.name)[stat.ST_MTIME]))
        else:
            timestamp = time.strftime("%Y-%m-%dT%H:%M:%S+00:00", time.gmtime())

        basename = os.path.basename(file.name)
        if basename.startswith("draft-"):
            draft = basename
            _debug( "** Processing '%s'" % draft)
            _printmeta(timestamp, file.name, outfile)
        else:
            for line in file:
                draft = line.strip()
                if draft.startswith("#"):
                    continue
                if draft:
                    _debug( "** Processing '%s'" % draft)
                    _printmeta(timestamp, draft, outfile)

if __name__ == "__main__":
    try:
        _main()
    except KeyboardInterrupt:
        raise
        pass