# This file is part of lyx2lyx
# Copyright (C) 2011 The LyX team
#
# 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.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA

"""Convert files to the file format generated by lyx 2.0"""

import re

from lyx2lyx_tools import (
    add_to_preamble,
    hex2ratio,
    insert_to_preamble,
    latex_length,
    lyx2latex,
    put_cmd_in_ert,
    revert_flex_inset,
    revert_font_attrs,
    revert_language,
    str2bool,
)
from parser_tools import (
    check_token,
    del_complete_lines,
    del_token,
    find_end_of,
    find_end_of_inset,
    find_end_of_layout,
    find_token,
    find_token_backwards,
    find_token_exact,
    get_option_value,
    get_quoted_value,
    get_value,
    is_in_inset,
)

####################################################################
# Private helper functions


def remove_option(lines, m, option):
    """removes option from line m. returns whether we did anything"""
    l = lines[m].find(option)
    if l == -1:
        return False
    val = lines[m][l:].split('"')[1]
    lines[m] = lines[m][: l - 1] + lines[m][l + len(option + '="' + val + '"') :]
    return True


###############################################################################
###
### Conversion and reversion routines
###
###############################################################################


def revert_swiss(document):
    "Set language german-ch to ngerman"
    i = 0
    if document.language == "german-ch":
        document.language = "ngerman"
        i = find_token(document.header, "\\language", 0)
        if i != -1:
            document.header[i] = "\\language ngerman"
    j = 0
    while True:
        j = find_token(document.body, "\\lang german-ch", j)
        if j == -1:
            return
        document.body[j] = document.body[j].replace("\\lang german-ch", "\\lang ngerman")
        j = j + 1


def revert_tabularvalign(document):
    "Revert the tabular valign option"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Tabular", i)
        if i == -1:
            return
        end = find_end_of_inset(document.body, i)
        if end == -1:
            document.warning("Can't find end of inset at line " + str(i))
            i += 1
            continue
        fline = find_token(document.body, "<features", i, end)
        if fline == -1:
            document.warning("Can't find features for inset at line " + str(i))
            i += 1
            continue
        p = document.body[fline].find("islongtable")
        if p != -1:
            q = document.body[fline].find("tabularvalignment")
            if q != -1:
                document.body[fline] = re.sub(
                    r" tabularvalignment=\"[a-z]+\"", "", document.body[fline]
                )
            i += 1
            continue

        # no longtable
        tabularvalignment = "c"
        # which valignment is specified?
        m = document.body[fline].find('tabularvalignment="top"')
        if m != -1:
            tabularvalignment = "t"
        m = document.body[fline].find('tabularvalignment="bottom"')
        if m != -1:
            tabularvalignment = "b"
        # delete tabularvalignment
        q = document.body[fline].find("tabularvalignment")
        if q != -1:
            document.body[fline] = re.sub(
                r" tabularvalignment=\"[a-z]+\"", "", document.body[fline]
            )

        # don't add a box when centered
        if tabularvalignment == "c":
            i = end
            continue
        subst = ["\\end_inset", "\\end_layout"]
        document.body[end:end] = subst  # just inserts those lines
        subst = [
            "\\begin_inset Box Frameless",
            'position "' + tabularvalignment + '"',
            'hor_pos "c"',
            "has_inner_box 1",
            'inner_pos "c"',
            "use_parbox 0",
            # we don't know the width, assume 50%
            'width "50col%"',
            'special "none"',
            'height "1in"',
            'height_special "totalheight"',
            "status open",
            "",
            "\\begin_layout Plain Layout",
        ]
        document.body[i:i] = subst  # this just inserts the array at i
        # since there could be a tabular inside a tabular, we cannot
        # jump to end
        i += len(subst)


def revert_phantom_types(document, ptype, cmd):
    "Reverts phantom to ERT"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Phantom " + ptype, i)
        if i == -1:
            return
        end = find_end_of_inset(document.body, i)
        if end == -1:
            document.warning("Can't find end of inset at line " + str(i))
            i += 1
            continue
        blay = find_token(document.body, "\\begin_layout Plain Layout", i, end)
        if blay == -1:
            document.warning("Can't find layout for inset at line " + str(i))
            i = end
            continue
        bend = find_end_of_layout(document.body, blay)
        if bend == -1:
            document.warning(
                "Malformed LyX document: Could not find end of Phantom inset's layout."
            )
            i = end
            continue
        substi = [
            "\\begin_inset ERT",
            "status collapsed",
            "",
            "\\begin_layout Plain Layout",
            "",
            "",
            "\\backslash",
            cmd + "{",
            "\\end_layout",
            "",
            "\\end_inset",
        ]
        substj = [
            "\\size default",
            "",
            "\\begin_inset ERT",
            "status collapsed",
            "",
            "\\begin_layout Plain Layout",
            "",
            "}",
            "\\end_layout",
            "",
            "\\end_inset",
        ]
        # do the later one first so as not to mess up the numbering
        document.body[bend : end + 1] = substj
        document.body[i : blay + 1] = substi
        i = end + len(substi) + len(substj) - (end - bend) - (blay - i) - 2


def revert_phantom(document):
    revert_phantom_types(document, "Phantom", "phantom")


def revert_hphantom(document):
    revert_phantom_types(document, "HPhantom", "hphantom")


def revert_vphantom(document):
    revert_phantom_types(document, "VPhantom", "vphantom")


def revert_xetex(document):
    "Reverts documents that use XeTeX"

    i = find_token(document.header, "\\use_xetex", 0)
    if i == -1:
        document.warning("Malformed LyX document: Missing \\use_xetex.")
        return
    if not str2bool(get_value(document.header, "\\use_xetex", i)):
        del document.header[i]
        return
    del document.header[i]

    # 1.) set doc encoding to utf8-plain
    i = find_token(document.header, "\\inputencoding", 0)
    if i == -1:
        document.warning("Malformed LyX document: Missing \\inputencoding.")
    else:
        document.header[i] = "\\inputencoding utf8-plain"

    # 2.) check font settings
    # defaults
    roman = sans = typew = "default"
    osf = False
    sf_scale = tt_scale = 100.0

    i = find_token(document.header, "\\font_roman", 0)
    if i == -1:
        document.warning("Malformed LyX document: Missing \\font_roman.")
    else:
        roman = get_value(document.header, "\\font_roman", i)
        document.header[i] = "\\font_roman default"

    i = find_token(document.header, "\\font_sans", 0)
    if i == -1:
        document.warning("Malformed LyX document: Missing \\font_sans.")
    else:
        sans = get_value(document.header, "\\font_sans", i)
        document.header[i] = "\\font_sans default"

    i = find_token(document.header, "\\font_typewriter", 0)
    if i == -1:
        document.warning("Malformed LyX document: Missing \\font_typewriter.")
    else:
        typew = get_value(document.header, "\\font_typewriter", i)
        document.header[i] = "\\font_typewriter default"

    i = find_token(document.header, "\\font_osf", 0)
    if i == -1:
        document.warning("Malformed LyX document: Missing \\font_osf.")
    else:
        osf = str2bool(get_value(document.header, "\\font_osf", i))
        document.header[i] = "\\font_osf false"

    i = find_token(document.header, "\\font_sc", 0)
    if i == -1:
        document.warning("Malformed LyX document: Missing \\font_sc.")
    else:
        # we do not need this value.
        document.header[i] = "\\font_sc false"

    i = find_token(document.header, "\\font_sf_scale", 0)
    if i == -1:
        document.warning("Malformed LyX document: Missing \\font_sf_scale.")
    else:
        val = get_value(document.header, "\\font_sf_scale", i)
        try:
            # float() can throw
            sf_scale = float(val)
        except:
            document.warning("Invalid font_sf_scale value: " + val)
        document.header[i] = "\\font_sf_scale 100"

    i = find_token(document.header, "\\font_tt_scale", 0)
    if i == -1:
        document.warning("Malformed LyX document: Missing \\font_tt_scale.")
    else:
        val = get_value(document.header, "\\font_tt_scale", i)
        try:
            # float() can throw
            tt_scale = float(val)
        except:
            document.warning("Invalid font_tt_scale value: " + val)
        document.header[i] = "\\font_tt_scale 100"

    # 3.) set preamble stuff
    pretext = ["%% This document must be processed with xelatex!"]
    pretext.append("\\usepackage{fontspec}")
    if roman != "default":
        pretext.append("\\setmainfont[Mapping=tex-text]{" + roman + "}")
    if sans != "default":
        sf = "\\setsansfont["
        if sf_scale != 100.0:
            sf += "Scale=" + str(sf_scale / 100.0) + ","
        sf += "Mapping=tex-text]{" + sans + "}"
        pretext.append(sf)
    if typew != "default":
        tw = "\\setmonofont"
        if tt_scale != 100.0:
            tw += "[Scale=" + str(tt_scale / 100.0) + "]"
        tw += "{" + typew + "}"
        pretext.append(tw)
    if osf:
        pretext.append("\\defaultfontfeatures{Numbers=OldStyle}")
    pretext.append("\\usepackage{xunicode}")
    pretext.append("\\usepackage{xltxtra}")
    insert_to_preamble(document, pretext)


def revert_outputformat(document):
    "Remove default output format param"

    if not del_token(document.header, "\\default_output_format", 0):
        document.warning("Malformed LyX document: Missing \\default_output_format.")


def revert_backgroundcolor(document):
    "Reverts background color to preamble code"
    i = find_token(document.header, "\\backgroundcolor", 0)
    if i == -1:
        return
    colorcode = get_value(document.header, "\\backgroundcolor", i)
    del document.header[i]
    # don't clutter the preamble if backgroundcolor is not set
    if colorcode == "#ffffff":
        return
    red = hex2ratio(colorcode[1:3])
    green = hex2ratio(colorcode[3:5])
    blue = hex2ratio(colorcode[5:7])
    insert_to_preamble(
        document,
        [
            "% To set the background color",
            "\\@ifundefined{definecolor}{\\usepackage{color}}{}",
            "\\definecolor{page_backgroundcolor}{rgb}{" + red + "," + green + "," + blue + "}",
            "\\pagecolor{page_backgroundcolor}",
        ],
    )


def add_use_indices(document):
    "Add \\use_indices if it is missing"
    i = find_token(document.header, "\\use_indices", 0)
    if i != -1:
        return i
    i = find_token(document.header, "\\use_bibtopic", 0)
    if i == -1:
        i = find_token(document.header, "\\cite_engine", 0)
    if i == -1:
        i = find_token(document.header, "\\use_mathdots", 0)
    if i == -1:
        i = find_token(document.header, "\\use_mhchem", 0)
    if i == -1:
        i = find_token(document.header, "\\use_esint", 0)
    if i == -1:
        i = find_token(document.header, "\\use_amsmath", 0)
    if i == -1:
        document.warning("Malformed LyX document: Missing \\use_indices.")
        return -1
    document.header.insert(i + 1, "\\use_indices 0")
    return i + 1


def revert_splitindex(document):
    "Reverts splitindex-aware documents"
    i = add_use_indices(document)
    if i == -1:
        return
    useindices = str2bool(get_value(document.header, "\\use_indices", i))
    del document.header[i]
    preamble = []
    if useindices:
        preamble.append("\\usepackage{splitidx})")

    # deal with index declarations in the preamble
    i = 0
    while True:
        i = find_token(document.header, "\\index", i)
        if i == -1:
            break
        k = find_token(document.header, "\\end_index", i)
        if k == -1:
            document.warning("Malformed LyX document: Missing \\end_index.")
            return
        if useindices:
            line = document.header[i]
            l = re.compile(r"\\index (.*)$")
            m = l.match(line)
            iname = m.group(1)
            ishortcut = get_value(document.header, "\\shortcut", i, k)
            if ishortcut != "":
                preamble.append("\\newindex[" + iname + "]{" + ishortcut + "}")
        del document.header[i : k + 1]
    if preamble:
        insert_to_preamble(document, preamble)

    # deal with index insets
    # these need to have the argument removed
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Index", i)
        if i == -1:
            break
        line = document.body[i]
        l = re.compile(r"\\begin_inset Index (.*)$")
        m = l.match(line)
        itype = m.group(1)
        if itype == "idx" or useindices == "false":
            document.body[i] = "\\begin_inset Index"
        else:
            k = find_end_of_inset(document.body, i)
            if k == -1:
                document.warning("Can't find end of index inset!")
                i += 1
                continue
            content = lyx2latex(document, document.body[i:k])
            # escape quotes
            content = content.replace('"', r"\"")
            subst = put_cmd_in_ert("\\sindex[" + itype + "]{" + content + "}")
            document.body[i : k + 1] = subst
        i = i + 1

    # deal with index_print insets
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset CommandInset index_print", i)
        if i == -1:
            return
        k = find_end_of_inset(document.body, i)
        ptype = get_quoted_value(document.body, "type", i, k)
        if ptype == "idx":
            j = find_token(document.body, "type", i, k)
            del document.body[j]
        elif not useindices:
            del document.body[i : k + 1]
        else:
            subst = put_cmd_in_ert("\\printindex[" + ptype + "]{}")
            document.body[i : k + 1] = subst
        i = i + 1


def convert_splitindex(document):
    "Converts index and printindex insets to splitindex-aware format"
    add_use_indices(document)
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Index", i)
        if i == -1:
            break
        document.body[i] = document.body[i].replace(
            "\\begin_inset Index", "\\begin_inset Index idx"
        )
        i = i + 1
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset CommandInset index_print", i)
        if i == -1:
            return
        if document.body[i + 1].find("LatexCommand printindex") == -1:
            document.warning("Malformed LyX document: Incomplete printindex inset.")
            return
        subst = ["LatexCommand printindex", 'type "idx"']
        document.body[i + 1 : i + 2] = subst
        i = i + 1


def revert_subindex(document):
    "Reverts \\printsubindex CommandInset types"
    i = add_use_indices(document)
    if i == -1:
        return
    useindices = str2bool(get_value(document.header, "\\use_indices", i))
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset CommandInset index_print", i)
        if i == -1:
            return
        k = find_end_of_inset(document.body, i)
        ctype = get_value(document.body, "LatexCommand", i, k)
        if ctype != "printsubindex":
            i = k + 1
            continue
        ptype = get_quoted_value(document.body, "type", i, k)
        if not useindices:
            del document.body[i : k + 1]
        else:
            subst = put_cmd_in_ert("\\printsubindex[" + ptype + "]{}")
            document.body[i : k + 1] = subst
        i = i + 1


def revert_printindexall(document):
    "Reverts \\print[sub]index* CommandInset types"
    i = add_use_indices(document)
    if i == -1:
        return
    useindices = str2bool(get_value(document.header, "\\use_indices", i))
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset CommandInset index_print", i)
        if i == -1:
            return
        k = find_end_of_inset(document.body, i)
        ctype = get_value(document.body, "LatexCommand", i, k)
        if ctype != "printindex*" and ctype != "printsubindex*":
            i = k
            continue
        if not useindices:
            del document.body[i : k + 1]
        else:
            subst = put_cmd_in_ert("\\" + ctype + "{}")
            document.body[i : k + 1] = subst
        i = i + 1


strikeout_preamble = [
    "%  for proper underlining",
    r"\PassOptionsToPackage{normalem}{ulem}",
    r"\usepackage{ulem}",
]


def convert_strikeout(document):
    "Remove preamble code loading 'ulem' package."
    del_complete_lines(document.preamble, ["% Added by lyx2lyx"] + strikeout_preamble)


def revert_strikeout(document):
    "Reverts \\strikeout font attribute"
    changed = revert_font_attrs(document.body, "\\uuline", "\\uuline")
    changed = revert_font_attrs(document.body, "\\uwave", "\\uwave") or changed
    changed = revert_font_attrs(document.body, "\\strikeout", "\\sout") or changed
    if changed == True:
        insert_to_preamble(document, strikeout_preamble)


ulinelatex_preamble = [
    "% fix underbar in citations",
    r"\let\cite@rig\cite",
    r"\newcommand{\b@xcite}[2][\%]{\def\def@pt{\%}\def\pas@pt{#1}",
    r"  \mbox{\ifx\def@pt\pas@pt\cite@rig{#2}\else\cite@rig[#1]{#2}\fi}}",
    r"\renewcommand{\underbar}[1]{{\let\cite\b@xcite\uline{#1}}}",
]


def convert_ulinelatex(document):
    "Remove preamble code for \\uline font attribute."
    del_complete_lines(document.preamble, ["% Added by lyx2lyx"] + ulinelatex_preamble)


def revert_ulinelatex(document):
    "Add preamble code for \\uline font attribute in citations."
    i = find_token(document.body, "\\bar under", 0)
    if i == -1:
        return
    try:
        document.preamble.index(r"\usepackage{ulem}")
    except ValueError:
        insert_to_preamble(document, strikeout_preamble)
    insert_to_preamble(document, ulinelatex_preamble)


def revert_custom_processors(document):
    "Remove bibtex_command and index_command params"

    if not del_token(document.header, "\\bibtex_command", 0):
        document.warning("Malformed LyX document: Missing \\bibtex_command.")

    if not del_token(document.header, "\\index_command", 0):
        document.warning("Malformed LyX document: Missing \\index_command.")


def convert_nomencl_width(document):
    "Add set_width param to nomencl_print"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset CommandInset nomencl_print", i)
        if i == -1:
            break
        document.body.insert(i + 2, 'set_width "none"')
        i = i + 1


def revert_nomencl_width(document):
    "Remove set_width param from nomencl_print"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset CommandInset nomencl_print", i)
        if i == -1:
            break
        j = find_end_of_inset(document.body, i)
        if not del_token(document.body, "set_width", i, j):
            document.warning("Can't find set_width option for nomencl_print!")
        i = j


def revert_nomencl_cwidth(document):
    "Remove width param from nomencl_print"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset CommandInset nomencl_print", i)
        if i == -1:
            break
        j = find_end_of_inset(document.body, i)
        l = find_token(document.body, "width", i, j)
        if l == -1:
            i = j
            continue
        width = get_quoted_value(document.body, "width", i, j)
        del document.body[l]
        insert_to_preamble(document, ["\\setlength{\\nomlabelwidth}{" + width + "}"])
        i = j - 1


def revert_applemac(document):
    "Revert applemac encoding to auto"
    if document.encoding != "applemac":
        return
    document.encoding = "auto"
    i = find_token(document.header, "\\encoding", 0)
    if i != -1:
        document.header[i] = "\\encoding auto"


def revert_longtable_align(document):
    "Remove longtable alignment setting"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Tabular", i)
        if i == -1:
            break
        end = find_end_of_inset(document.body, i)
        if end == -1:
            document.warning("Can't find end of inset at line " + str(i))
            i += 1
            continue
        fline = find_token(document.body, "<features", i, end)
        if fline == -1:
            document.warning("Can't find features for inset at line " + str(i))
            i += 1
            continue
        j = document.body[fline].find("longtabularalignment")
        if j == -1:
            i += 1
            continue
        # FIXME Is this correct? It wipes out everything after the
        # one we found.
        document.body[fline] = document.body[fline][: j - 1] + ">"
        # since there could be a tabular inside this one, we
        # cannot jump to end.
        i += 1


def revert_branch_filename(document):
    "Remove \\filename_suffix parameter from branches"
    i = 0
    while True:
        i = find_token(document.header, "\\filename_suffix", i)
        if i == -1:
            return
        del document.header[i]


def revert_paragraph_indentation(document):
    "Revert custom paragraph indentation to preamble code"
    i = find_token(document.header, "\\paragraph_indentation", 0)
    if i == -1:
        return
    length = get_value(document.header, "\\paragraph_indentation", i)
    # we need only remove the line if indentation is default
    if length != "default":
        # handle percent lengths
        length = latex_length(length)[1]
        insert_to_preamble(document, ["\\setlength{\\parindent}{" + length + "}"])
    del document.header[i]


def revert_percent_skip_lengths(document):
    "Revert relative lengths for paragraph skip separation to preamble code"
    i = find_token(document.header, "\\defskip", 0)
    if i == -1:
        return
    length = get_value(document.header, "\\defskip", i)
    # only revert when a custom length was set and when
    # it used a percent length
    if length in ("smallskip", "medskip", "bigskip"):
        return
    # handle percent lengths
    percent, length = latex_length(length)
    if percent:
        insert_to_preamble(document, ["\\setlength{\\parskip}{" + length + "}"])
        # set defskip to medskip as default
        document.header[i] = "\\defskip medskip"


def revert_percent_vspace_lengths(document):
    "Revert relative VSpace lengths to ERT"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset VSpace", i)
        if i == -1:
            break
        # only revert if a custom length was set and if
        # it used a percent length
        r = re.compile(r"\\begin_inset VSpace (.*)$")
        m = r.match(document.body[i])
        length = m.group(1)
        if length in ("defskip", "smallskip", "medskip", "bigskip", "vfill"):
            i += 1
            continue
        # check if the space has a star (protected space)
        protected = document.body[i].rfind("*") != -1
        if protected:
            length = length.rstrip("*")
        # handle percent lengths
        percent, length = latex_length(length)
        # revert the VSpace inset to ERT
        if percent:
            if protected:
                subst = put_cmd_in_ert("\\vspace*{" + length + "}")
            else:
                subst = put_cmd_in_ert("\\vspace{" + length + "}")
            document.body[i : i + 2] = subst
        i += 1


def revert_percent_hspace_lengths(document):
    "Revert relative HSpace lengths to ERT"
    i = 0
    while True:
        i = find_token_exact(document.body, "\\begin_inset space \\hspace", i)
        if i == -1:
            break
        j = find_end_of_inset(document.body, i)
        if j == -1:
            document.warning("Can't find end of inset at line " + str(i))
            i += 1
            continue
        # only revert if a custom length was set...
        length = get_value(document.body, "\\length", i + 1, j)
        if length == "":
            document.warning("Malformed lyx document: Missing '\\length' in Space inset.")
            i = j
            continue
        protected = ""
        if document.body[i].find("\\hspace*{}") != -1:
            protected = "*"
        # ...and if it used a percent length
        percent, length = latex_length(length)
        # revert the HSpace inset to ERT
        if percent:
            subst = put_cmd_in_ert("\\hspace" + protected + "{" + length + "}")
            document.body[i : j + 1] = subst
        # if we did a substitution, this will still be ok
        i = j


def revert_hspace_glue_lengths(document):
    "Revert HSpace glue lengths to ERT"
    i = 0
    while True:
        i = find_token_exact(document.body, "\\begin_inset space \\hspace", i)
        if i == -1:
            break
        j = find_end_of_inset(document.body, i)
        if j == -1:
            document.warning("Can't find end of inset at line " + str(i))
            i += 1
            continue
        length = get_value(document.body, "\\length", i + 1, j)
        if length == "":
            document.warning("Malformed lyx document: Missing '\\length' in Space inset.")
            i = j
            continue
        protected = ""
        if document.body[i].find("\\hspace*{}") != -1:
            protected = "*"
        # only revert if the length contains a plus or minus at pos != 0
        if length.find("-", 1) != -1 or length.find("+", 1) != -1:
            # handle percent lengths
            length = latex_length(length)[1]
            # revert the HSpace inset to ERT
            subst = put_cmd_in_ert("\\hspace" + protected + "{" + length + "}")
            document.body[i : j + 1] = subst
        i = j


def convert_author_id(document):
    "Add the author_id to the \\author definition and make sure 0 is not used"
    i = 0
    anum = 1
    re_author = re.compile(r"(\\author) (\".*\")\s*(.*)$")

    while True:
        i = find_token(document.header, "\\author", i)
        if i == -1:
            break
        m = re_author.match(document.header[i])
        if m:
            name = m.group(2)
            email = m.group(3)
            document.header[i] = "\\author %i %s %s" % (anum, name, email)
        anum += 1
        i += 1

    i = 0
    while True:
        i = find_token(document.body, "\\change_", i)
        if i == -1:
            break
        change = document.body[i].split(" ")
        if len(change) == 3:
            type = change[0]
            author_id = int(change[1])
            time = change[2]
            document.body[i] = "%s %i %s" % (type, author_id + 1, time)
        i += 1


def revert_author_id(document):
    "Remove the author_id from the \\author definition"
    i = 0
    anum = 0
    rx = re.compile(r"(\\author)\s+(-?\d+)\s+(\".*\")\s*(.*)$")
    idmap = dict()

    while True:
        i = find_token(document.header, "\\author", i)
        if i == -1:
            break
        m = rx.match(document.header[i])
        if m:
            author_id = int(m.group(2))
            idmap[author_id] = anum
            name = m.group(3)
            email = m.group(4)
            document.header[i] = f"\\author {name} {email}"
        i += 1
        # FIXME Should this be incremented if we didn't match?
        anum += 1

    i = 0
    while True:
        i = find_token(document.body, "\\change_", i)
        if i == -1:
            break
        change = document.body[i].split(" ")
        if len(change) == 3:
            type = change[0]
            author_id = int(change[1])
            time = change[2]
            document.body[i] = "%s %i %s" % (type, idmap[author_id], time)
        i += 1


def revert_suppress_date(document):
    "Revert suppressing of default document date to preamble code"
    i = find_token(document.header, "\\suppress_date", 0)
    if i == -1:
        return
    # remove the preamble line and write to the preamble
    # when suppress_date was true
    date = str2bool(get_value(document.header, "\\suppress_date", i))
    if date:
        add_to_preamble(document, ["\\date{}"])
    del document.header[i]


mhchem_preamble = [r"\PassOptionsToPackage{version=3}{mhchem}", r"\usepackage{mhchem}"]


def convert_mhchem(document):
    "Set mhchem to off for versions older than 1.6.x"
    if document.initial_format < 277:
        # LyX 1.5.x and older did never load mhchem.
        # Therefore we must switch it off: Documents that use mhchem have
        # a manual \usepackage anyway, and documents not using mhchem but
        # custom macros with the same names as mhchem commands might get
        # corrupted if mhchem is automatically loaded.
        mhchem = 0  # off
    else:
        # LyX 1.6.x did always load mhchem automatically.
        mhchem = 1  # auto
    i = find_token(document.header, "\\use_esint", 0)
    if i == -1:
        # pre-1.5.x document
        i = find_token(document.header, "\\use_amsmath", 0)
    if i == -1:
        document.warning("Malformed LyX document: " "Could not find amsmath or esint setting.")
        return
    document.header.insert(i + 1, "\\use_mhchem %d" % mhchem)
    # remove LyX-inserted preamble
    if mhchem != 0:
        del_complete_lines(document.preamble, ["% Added by lyx2lyx"] + mhchem_preamble)


def revert_mhchem(document):
    "Revert mhchem loading to preamble code."

    mhchem = get_value(document.header, "\\use_mhchem", delete=True)
    try:
        mhchem = int(mhchem)
    except ValueError:
        document.warning("Malformed LyX document: " "Could not find mhchem setting.")
        mhchem = 1  # "auto"
    # mhchem in {0: "off", 1: "auto", 2: "on"}

    if mhchem == 1:  # "auto"
        i = 0
        while i != 1 and mhchem == 1:
            i = find_token(document.body, "\\begin_inset Formula", i)
            j = find_end_of_inset(document.body, i)
            if j == -1:
                break
            if (True for line in document.body[i:j] if r"\ce{" in line or r"\cf{" in line):
                mhchem = 2
                break
            i += 1

    if (
        mhchem == 2  # on
        and find_token(document.preamble, r"\usepackage{mhchem}") == -1
    ):
        insert_to_preamble(document, mhchem_preamble)


def revert_fontenc(document):
    "Remove fontencoding param"
    if not del_token(document.header, "\\fontencoding", 0):
        document.warning("Malformed LyX document: Missing \\fontencoding.")


def merge_gbrief(document):
    "Merge g-brief-en and g-brief-de to one class"

    if document.textclass != "g-brief-de":
        if document.textclass == "g-brief-en":
            document.textclass = "g-brief"
            document.set_textclass()
        return

    obsoletedby = {
        "Brieftext": "Letter",
        "Unterschrift": "Signature",
        "Strasse": "Street",
        "Zusatz": "Addition",
        "Ort": "Town",
        "Land": "State",
        "RetourAdresse": "ReturnAddress",
        "MeinZeichen": "MyRef",
        "IhrZeichen": "YourRef",
        "IhrSchreiben": "YourMail",
        "Telefon": "Phone",
        "BLZ": "BankCode",
        "Konto": "BankAccount",
        "Postvermerk": "PostalComment",
        "Adresse": "Address",
        "Datum": "Date",
        "Betreff": "Reference",
        "Anrede": "Opening",
        "Anlagen": "Encl.",
        "Verteiler": "cc",
        "Gruss": "Closing",
    }
    i = 0
    while True:
        i = find_token(document.body, "\\begin_layout", i)
        if i == -1:
            break

        layout = document.body[i][14:]
        if layout in obsoletedby:
            document.body[i] = "\\begin_layout " + obsoletedby[layout]

        i += 1

    document.textclass = "g-brief"
    document.set_textclass()


def revert_gbrief(document):
    "Revert g-brief to g-brief-en"
    if document.textclass == "g-brief":
        document.textclass = "g-brief-en"
        document.set_textclass()


def revert_html_options(document):
    "Remove html options"
    del_token(document.header, "\\html_use_mathml", 0)
    del_token(document.header, "\\html_be_strict", 0)


def revert_includeonly(document):
    i = 0
    while True:
        i = find_token(document.header, "\\begin_includeonly", i)
        if i == -1:
            return
        j = find_end_of(document.header, i, "\\begin_includeonly", "\\end_includeonly")
        if j == -1:
            document.warning("Unable to find end of includeonly section!!")
            break
        document.header[i : j + 1] = []


def convert_includeall(document):
    "Add maintain_unincluded_children param"

    i = 0
    i = find_token(document.header, "\\maintain_unincluded_children", 0)
    if i == -1:
        i = find_token(document.header, "\\textclass", 0)
        if i == -1:
            document.warning("Malformed LyX document! Missing \\textclass header.")
            return
        document.header.insert(i, "\\maintain_unincluded_children false")
        return


def revert_includeall(document):
    "Remove maintain_unincluded_children param"
    del_token(document.header, "\\maintain_unincluded_children", 0)


def revert_multirow(document):
    "Revert multirow cells in tables to TeX-code"

    # first, let's find out if we need to do anything
    # cell type 3 is multirow begin cell
    i = find_token(document.body, '<cell multirow="3"', 0)
    if i == -1:
        return

    add_to_preamble(document, ["\\usepackage{multirow}"])

    begin_table = 0
    while True:
        # find begin/end of table
        begin_table = find_token(document.body, "<lyxtabular version=", begin_table)
        if begin_table == -1:
            break
        end_table = find_end_of(document.body, begin_table, "<lyxtabular", "</lyxtabular>")
        if end_table == -1:
            document.warning("Malformed LyX document: Could not find end of table.")
            begin_table += 1
            continue
        # does this table have multirow?
        i = find_token(document.body, '<cell multirow="3"', begin_table, end_table)
        if i == -1:
            begin_table = end_table
            continue

        # store the number of rows and columns
        numrows = get_option_value(document.body[begin_table], "rows")
        numcols = get_option_value(document.body[begin_table], "columns")
        try:
            numrows = int(numrows)
            numcols = int(numcols)
        except:
            document.warning("Unable to determine rows and columns!")
            begin_table = end_table
            continue

        mrstarts = []
        multirows = []
        # collect info on rows and columns of this table.
        begin_row = begin_table
        for row in range(numrows):
            begin_row = find_token(document.body, "<row>", begin_row, end_table)
            if begin_row == -1:
                document.warning("Can't find row " + str(row + 1))
                break
            end_row = find_end_of(document.body, begin_row, "<row>", "</row>")
            if end_row == -1:
                document.warning("Can't find end of row " + str(row + 1))
                break
            begin_cell = begin_row
            multirows.append([])
            for column in range(numcols):
                begin_cell = find_token(document.body, "<cell ", begin_cell, end_row)
                if begin_cell == -1:
                    document.warning(
                        "Can't find column " + str(column + 1) + "in row " + str(row + 1)
                    )
                    break
                # NOTE
                # this will fail if someone puts "</cell>" in a cell, but
                # that seems fairly unlikely.
                end_cell = find_end_of(document.body, begin_cell, "<cell", "</cell>")
                if end_cell == -1:
                    document.warning(
                        "Can't find end of column " + str(column + 1) + "in row " + str(row + 1)
                    )
                    break
                multirows[row].append([begin_cell, end_cell, 0])
                if document.body[begin_cell].find('multirow="3"') != -1:
                    multirows[row][column][2] = 3  # begin multirow
                    mrstarts.append([row, column])
                elif document.body[begin_cell].find('multirow="4"') != -1:
                    multirows[row][column][2] = 4  # in multirow
                begin_cell = end_cell
            begin_row = end_row
        # end of table info collection

        # work from the back to avoid messing up numbering
        mrstarts.reverse()
        for m in mrstarts:
            row = m[0]
            col = m[1]
            # get column width
            col_width = get_option_value(document.body[begin_table + 2 + col], "width")
            # "0pt" means that no width is specified
            if not col_width or col_width == "0pt":
                col_width = "*"
            # determine the number of cells that are part of the multirow
            nummrs = 1
            for r in range(row + 1, numrows):
                if multirows[r][col][2] != 4:
                    break
                nummrs += 1
                # take the opportunity to revert this line
                lineno = multirows[r][col][0]
                document.body[lineno] = (
                    document.body[lineno]
                    .replace(' multirow="4" ', " ")
                    .replace('valignment="middle"', 'valignment="top"')
                    .replace(' topline="true" ', " ")
                )
                # remove bottom line of previous multirow-part cell
                lineno = multirows[r - 1][col][0]
                document.body[lineno] = document.body[lineno].replace(
                    ' bottomline="true" ', " "
                )
            # revert beginning cell
            bcell = multirows[row][col][0]
            ecell = multirows[row][col][1]
            document.body[bcell] = (
                document.body[bcell]
                .replace(' multirow="3" ', " ")
                .replace('valignment="middle"', 'valignment="top"')
            )
            blay = find_token(document.body, "\\begin_layout", bcell, ecell)
            if blay == -1:
                document.warning("Can't find layout for cell!")
                continue
            bend = find_end_of_layout(document.body, blay)
            if bend == -1:
                document.warning("Can't find end of layout for cell!")
                continue
            # do the later one first, so as not to mess up the numbering
            # we are wrapping the whole cell in this ert
            # so before the end of the layout...
            document.body[bend:bend] = put_cmd_in_ert("}")
            # ...and after the beginning
            document.body[blay + 1 : blay + 1] = put_cmd_in_ert(
                "\\multirow{" + str(nummrs) + "}{" + col_width + "}{"
            )

        begin_table = end_table


def convert_math_output(document):
    r"Convert \html_use_mathml to \html_math_output"
    i = find_token(document.header, "\\html_use_mathml", 0)
    if i == -1:
        return
    rgx = re.compile(r"\\html_use_mathml\s+(\w+)")
    m = rgx.match(document.header[i])
    newval = "0"  # MathML
    if m:
        val = str2bool(m.group(1))
        if not val:
            newval = "2"  # Images
    else:
        document.warning("Can't match " + document.header[i])
    document.header[i] = "\\html_math_output " + newval


def revert_math_output(document):
    r"Revert \html_math_output to \html_use_mathml"
    i = find_token(document.header, "\\html_math_output", 0)
    if i == -1:
        return
    rgx = re.compile(r"\\html_math_output\s+(\d)")
    m = rgx.match(document.header[i])
    newval = "true"
    if m:
        val = m.group(1)
        if val == "1" or val == "2":
            newval = "false"
    else:
        document.warning("Unable to match " + document.header[i])
    document.header[i] = "\\html_use_mathml " + newval


def revert_inset_preview(document):
    "Dissolves the preview inset"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Preview", i)
        if i == -1:
            return
        iend = find_end_of_inset(document.body, i)
        if iend == -1:
            document.warning("Malformed LyX document: Could not find end of Preview inset.")
            i += 1
            continue

        # This has several issues.
        # We need to do something about the layouts inside InsetPreview.
        # If we just leave the first one, then we have something like:
        # \begin_layout Standard
        # ...
        # \begin_layout Standard
        # and we get a "no \end_layout" error. So something has to be done.
        # Ideally, we would check if it is the same as the layout we are in.
        # If so, we just remove it; if not, we end the active one. But it is
        # not easy to know what layout we are in, due to depth changes, etc,
        # and it is not clear to me how much work it is worth doing. In most
        # cases, the layout will probably be the same.
        #
        # For the same reason, we have to remove the \end_layout tag at the
        # end of the last layout in the inset. Again, that will sometimes be
        # wrong, but it will usually be right. To know what to do, we would
        # again have to know what layout the inset is in.

        blay = find_token(document.body, "\\begin_layout", i, iend)
        if blay == -1:
            document.warning("Can't find layout for preview inset!")
            # always do the later one first...
            del document.body[iend]
            del document.body[i]
            # deletions mean we do not need to reset i
            continue

        # This is where we would check what layout we are in.
        # The check for Standard is definitely wrong.
        #
        # lay = document.body[blay].split(None, 1)[1]
        # if lay != oldlayout:
        #     # record a boolean to tell us what to do later....
        #     # better to do it later, since (a) it won't mess up
        #     # the numbering and (b) we only modify at the end.

        # we want to delete the last \\end_layout in this inset, too.
        # note that this may not be the \\end_layout that goes with blay!!
        bend = find_end_of_layout(document.body, blay)
        while True:
            tmp = find_token(document.body, "\\end_layout", bend + 1, iend)
            if tmp == -1:
                break
            bend = tmp
        if bend == blay:
            document.warning("Unable to find last layout in preview inset!")
            del document.body[iend]
            del document.body[i]
            # deletions mean we do not need to reset i
            continue
        # always do the later one first...
        del document.body[iend]
        del document.body[bend]
        del document.body[i : blay + 1]
        # we do not need to reset i


def revert_equalspacing_xymatrix(document):
    "Revert a Formula with xymatrix@! to an ERT inset"
    i = 0
    has_preamble = False
    has_equal_spacing = False

    while True:
        i = find_token(document.body, "\\begin_inset Formula", i)
        if i == -1:
            break
        j = find_end_of_inset(document.body, i)
        if j == -1:
            document.warning("Malformed LyX document: Could not find end of Formula inset.")
            i += 1
            continue

        for curline in range(i, j):
            found = document.body[curline].find("\\xymatrix@!")
            if found != -1:
                break

        if found != -1:
            has_equal_spacing = True
            content = [document.body[i][21:]]
            content += document.body[i + 1 : j]
            subst = put_cmd_in_ert(content)
            document.body[i : j + 1] = subst
            i += len(subst) - (j - i) + 1
        else:
            for curline in range(i, j):
                l = document.body[curline].find("\\xymatrix")
                if l != -1:
                    has_preamble = True
                    break
            i = j + 1

    if has_equal_spacing and not has_preamble:
        add_to_preamble(document, ["\\usepackage[all]{xy}"])


def revert_notefontcolor(document):
    "Reverts greyed-out note font color to preamble code"

    i = find_token(document.header, "\\notefontcolor", 0)
    if i == -1:
        return

    colorcode = get_value(document.header, "\\notefontcolor", i)
    del document.header[i]

    # are there any grey notes?
    if find_token(document.body, "\\begin_inset Note Greyedout", 0) == -1:
        # no need to do anything else, and \renewcommand will throw
        # an error since lyxgreyedout will not exist.
        return

    # the color code is in the form #rrggbb where every character denotes a hex number
    red = hex2ratio(colorcode[1:3])
    green = hex2ratio(colorcode[3:5])
    blue = hex2ratio(colorcode[5:7])
    # write the preamble
    insert_to_preamble(
        document,
        [
            "%  for greyed-out notes",
            "\\@ifundefined{definecolor}{\\usepackage{color}}{}"
            "\\definecolor{note_fontcolor}{rgb}{%s,%s,%s}" % (red, green, blue),
            "\\renewenvironment{lyxgreyedout}",
            " {\\textcolor{note_fontcolor}\\bgroup}{\\egroup}",
        ],
    )


def revert_turkmen(document):
    "Set language Turkmen to English"

    revert_language(document, "turkmen", "turkmen", "turkmen")


def revert_fontcolor(document):
    "Reverts font color to preamble code"
    i = find_token(document.header, "\\fontcolor", 0)
    if i == -1:
        return
    colorcode = get_value(document.header, "\\fontcolor", i)
    del document.header[i]
    # don't clutter the preamble if font color is not set
    if colorcode == "#000000":
        return
    # the color code is in the form #rrggbb where every character denotes a hex number
    red = hex2ratio(colorcode[1:3])
    green = hex2ratio(colorcode[3:5])
    blue = hex2ratio(colorcode[5:7])
    # write the preamble
    insert_to_preamble(
        document,
        [
            "%  Set the font color",
            "\\@ifundefined{definecolor}{\\usepackage{color}}{}",
            f"\\definecolor{{document_fontcolor}}{{rgb}}{{{red},{green},{blue}}}",
            "\\color{document_fontcolor}",
        ],
    )


def revert_shadedboxcolor(document):
    "Reverts shaded box color to preamble code"
    i = find_token(document.header, "\\boxbgcolor", 0)
    if i == -1:
        return
    colorcode = get_value(document.header, "\\boxbgcolor", i)
    del document.header[i]
    # the color code is in the form #rrggbb
    red = hex2ratio(colorcode[1:3])
    green = hex2ratio(colorcode[3:5])
    blue = hex2ratio(colorcode[5:7])
    # write the preamble
    insert_to_preamble(
        document,
        [
            "%  Set the color of boxes with shaded background",
            "\\@ifundefined{definecolor}{\\usepackage{color}}{}",
            f"\\definecolor{{shadecolor}}{{rgb}}{{{red},{green},{blue}}}",
        ],
    )


def revert_lyx_version(document):
    "Reverts LyX Version information from Inset Info"
    version = "LyX version"
    try:
        import lyx2lyx_version

        version = lyx2lyx_version.version
    except:
        pass

    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Info", i)
        if i == -1:
            return
        j = find_end_of_inset(document.body, i + 1)
        if j == -1:
            document.warning("Malformed LyX document: Could not find end of Info inset.")
            i += 1
            continue

        # We expect:
        # \begin_inset Info
        # type  "lyxinfo"
        # arg   "version"
        # \end_inset
        typ = get_quoted_value(document.body, "type", i, j)
        arg = get_quoted_value(document.body, "arg", i, j)
        if arg != "version" or typ != "lyxinfo":
            i = j + 1
            continue

        # We do not actually know the version of LyX used to produce the document.
        # But we can use our version, since we are reverting.
        s = [version]
        # Now we want to check if the line after "\end_inset" is empty. It normally
        # is, so we want to remove it, too.
        lastline = j + 1
        if document.body[j + 1].strip() == "":
            lastline = j + 2
        document.body[i:lastline] = s
        i = i + 1


def revert_math_scale(document):
    "Remove math scaling and LaTeX options"
    del_token(document.header, "\\html_math_img_scale", 0)
    del_token(document.header, "\\html_latex_start", 0)
    del_token(document.header, "\\html_latex_end", 0)


def revert_pagesizes(document):
    "Revert page sizes to default"
    i = find_token(document.header, "\\papersize", 0)
    if i != -1:
        size = document.header[i][11:]
        if (
            size == "a0paper"
            or size == "a1paper"
            or size == "a2paper"
            or size == "a6paper"
            or size == "b0paper"
            or size == "b1paper"
            or size == "b2paper"
            or size == "b6paper"
            or size == "b0j"
            or size == "b1j"
            or size == "b2j"
            or size == "b3j"
            or size == "b4j"
            or size == "b5j"
            or size == "b6j"
        ):
            del document.header[i]


def revert_DIN_C_pagesizes(document):
    "Revert DIN C page sizes to default"
    i = find_token(document.header, "\\papersize", 0)
    if i != -1:
        size = document.header[i][11:]
        if (
            size == "c0paper"
            or size == "c1paper"
            or size == "c2paper"
            or size == "c3paper"
            or size == "c4paper"
            or size == "c5paper"
            or size == "c6paper"
        ):
            del document.header[i]


def convert_html_quotes(document):
    "Remove quotes around html_latex_start and html_latex_end"

    i = find_token(document.header, "\\html_latex_start", 0)
    if i != -1:
        line = document.header[i]
        l = re.compile(r'\\html_latex_start\s+"(.*)"')
        m = l.match(line)
        if m:
            document.header[i] = "\\html_latex_start " + m.group(1)

    i = find_token(document.header, "\\html_latex_end", 0)
    if i != -1:
        line = document.header[i]
        l = re.compile(r'\\html_latex_end\s+"(.*)"')
        m = l.match(line)
        if m:
            document.header[i] = "\\html_latex_end " + m.group(1)


def revert_html_quotes(document):
    "Remove quotes around html_latex_start and html_latex_end"

    i = find_token(document.header, "\\html_latex_start", 0)
    if i != -1:
        line = document.header[i]
        l = re.compile(r"\\html_latex_start\s+(.*)")
        m = l.match(line)
        if not m:
            document.warning("Weird html_latex_start line: " + line)
            del document.header[i]
        else:
            document.header[i] = '\\html_latex_start "' + m.group(1) + '"'

    i = find_token(document.header, "\\html_latex_end", 0)
    if i != -1:
        line = document.header[i]
        l = re.compile(r"\\html_latex_end\s+(.*)")
        m = l.match(line)
        if not m:
            document.warning("Weird html_latex_end line: " + line)
            del document.header[i]
        else:
            document.header[i] = '\\html_latex_end "' + m.group(1) + '"'


def revert_output_sync(document):
    "Remove forward search options"
    del_token(document.header, "\\output_sync_macro", 0)
    del_token(document.header, "\\output_sync", 0)


def revert_align_decimal(document):
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Tabular", i)
        if i == -1:
            return
        j = find_end_of_inset(document.body, i)
        if j == -1:
            document.warning("Unable to find end of Tabular inset at line " + str(i))
            i += 1
            continue
        cell = find_token(document.body, "<cell", i, j)
        if cell == -1:
            document.warning("Can't find any cells in Tabular inset at line " + str(i))
            i = j
            continue
        k = i + 1
        while True:
            k = find_token(document.body, "<column", k, cell)
            if k == -1:
                return
            if document.body[k].find('alignment="decimal"') == -1:
                k += 1
                continue
            remove_option(document.body, k, "decimal_point")
            document.body[k] = document.body[k].replace(
                'alignment="decimal"', 'alignment="center"'
            )
            k += 1


def convert_optarg(document):
    "Convert \\begin_inset OptArg to \\begin_inset Argument"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset OptArg", i)
        if i == -1:
            return
        document.body[i] = "\\begin_inset Argument"
        i += 1


def revert_argument(document):
    "Convert \\begin_inset Argument to \\begin_inset OptArg"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Argument", i)
        if i == -1:
            return
        document.body[i] = "\\begin_inset OptArg"
        i += 1


def revert_makebox(document):
    "Convert \\makebox to TeX code"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Box", i)
        if i == -1:
            break
        z = find_end_of_inset(document.body, i)
        if z == -1:
            document.warning("Malformed LyX document: Can't find end of box inset.")
            i += 1
            continue
        blay = find_token(document.body, "\\begin_layout", i, z)
        if blay == -1:
            document.warning("Malformed LyX document: Can't find layout in box.")
            i = z
            continue
        j = find_token(document.body, "use_makebox", i)
        if j == -1 or j != i + 6:
            document.warning("Malformed LyX document: Can't find use_makebox statement in box.")
            i = z
            continue
        # delete use_makebox
        if (
            not check_token(document.body[i], "\\begin_inset Box Frameless")
            or get_value(document.body, "use_makebox", j) != 1
        ):
            del document.body[j]
            i += 1
            continue
        bend = find_end_of_layout(document.body, blay)
        if bend == -1 or bend > z:
            document.warning("Malformed LyX document: Can't find end of layout in box.")
            i = z
            continue
        # determine the alignment
        align = get_quoted_value(document.body, "hor_pos", i, blay, "c")
        # determine the width
        length = get_quoted_value(document.body, "width", i, blay, "50col%")
        length = latex_length(length)[1]
        # remove the \end_layout \end_inset pair
        document.body[bend : z + 1] = put_cmd_in_ert("}")
        subst = "\\makebox[" + length + "][" + align + "]{"
        document.body[i : blay + 1] = put_cmd_in_ert(subst)
        i += 1


def convert_use_makebox(document):
    "Adds use_makebox option for boxes"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Box", i)
        if i == -1:
            return
        k = find_token(document.body, "use_parbox", i)
        if k == -1 or k != i + 5:
            document.warning("Malformed LyX document: Can't find use_parbox statement in box.")
            i += 1
            continue
        if k == i + 5:
            document.body.insert(k + 1, "use_makebox 0")
        i += 1


def revert_IEEEtran(document):
    "Convert IEEEtran layouts and styles to TeX code"

    if document.textclass != "IEEEtran":
        return

    revert_flex_inset(document, "IEEE membership", "\\IEEEmembership")
    revert_flex_inset(document, "Lowercase", "\\MakeLowercase")

    layouts = (
        "Special Paper Notice",
        "After Title Text",
        "Publication ID",
        "Page headings",
        "Biography without photo",
    )
    latexcmd = {
        "Special Paper Notice": "\\IEEEspecialpapernotice",
        "After Title Text": "\\IEEEaftertitletext",
        "Publication ID": "\\IEEEpubid",
    }
    obsoletedby = {
        "Page headings": "MarkBoth",
        "Biography without photo": "BiographyNoPhoto",
    }

    for layout in layouts:
        i = 0
        while True:
            i = find_token(document.body, "\\begin_layout " + layout, i)
            if i == -1:
                break
            j = find_end_of_layout(document.body, i)
            if j == -1:
                document.warning(
                    "Malformed LyX document: Can't find end of " + layout + " layout."
                )
                i += 1
                continue
            if layout in list(obsoletedby.keys()):
                document.body[i] = "\\begin_layout " + obsoletedby[layout]
                i = j
                continue
            content = lyx2latex(document, document.body[i : j + 1])
            add_to_preamble(document, [latexcmd[layout] + "{" + content + "}"])
            del document.body[i : j + 1]
            # no need to reset i


def convert_prettyref(document):
    "Converts prettyref references to neutral formatted refs"
    re_ref = re.compile('^\\s*reference\\s+"(\\w+):(\\S+)"')
    nm_ref = re.compile('^\\s*name\\s+"(\\w+):(\\S+)"')

    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset CommandInset ref", i)
        if i == -1:
            break
        j = find_end_of_inset(document.body, i)
        if j == -1:
            document.warning("Malformed LyX document: No end of InsetRef!")
            i += 1
            continue
        k = find_token(document.body, "LatexCommand prettyref", i, j)
        if k != -1:
            document.body[k] = "LatexCommand formatted"
        i = j + 1
    document.header.insert(-1, "\\use_refstyle 0")


def revert_refstyle(document):
    "Reverts neutral formatted refs to prettyref"
    re_ref = re.compile('^reference\\s+"(\\w+):(\\S+)"')
    nm_ref = re.compile('^\\s*name\\s+"(\\w+):(\\S+)"')

    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset CommandInset ref", i)
        if i == -1:
            break
        j = find_end_of_inset(document.body, i)
        if j == -1:
            document.warning("Malformed LyX document: No end of InsetRef")
            i += 1
            continue
        k = find_token(document.body, "LatexCommand formatted", i, j)
        if k != -1:
            document.body[k] = "LatexCommand prettyref"
        i = j + 1
    i = find_token(document.header, "\\use_refstyle", 0)
    if i != -1:
        document.header.pop(i)


def revert_nameref(document):
    "Convert namerefs to regular references"
    cmds = ["Nameref", "nameref"]
    foundone = False
    rx = re.compile(r'reference "(.*)"')
    for cmd in cmds:
        i = 0
        oldcmd = "LatexCommand " + cmd
        while True:
            # It seems better to look for this, as most of the reference
            # insets won't be ones we care about.
            i = find_token(document.body, oldcmd, i)
            if i == -1:
                break
            cmdloc = i
            i += 1
            # Make sure it is actually in an inset!
            # A normal line could begin with "LatexCommand nameref"!
            stins, endins = is_in_inset(document.body, cmdloc, "\\begin_inset CommandInset ref")
            if endins == -1:
                continue
            # ok, so it is in an InsetRef
            refline = find_token(document.body, "reference", stins, endins)
            if refline == -1:
                document.warning("Can't find reference for inset at line " + stinst + "!!")
                continue
            m = rx.match(document.body[refline])
            if not m:
                document.warning("Can't match reference line: " + document.body[ref])
                continue
            foundone = True
            ref = m.group(1)
            newcontent = put_cmd_in_ert("\\" + cmd + "{" + ref + "}")
            document.body[stins : endins + 1] = newcontent

    if foundone:
        add_to_preamble(document, ["\\usepackage{nameref}"])


def remove_Nameref(document):
    "Convert Nameref commands to nameref commands"
    i = 0
    while True:
        # It seems better to look for this, as most of the reference
        # insets won't be ones we care about.
        i = find_token(document.body, "LatexCommand Nameref", i)
        if i == -1:
            break
        cmdloc = i
        i += 1
        # Make sure it is actually in an inset!
        val = is_in_inset(
            document.body, cmdloc, "\\begin_inset CommandInset ref", default=False
        )
        if not val:
            continue
        document.body[cmdloc] = "LatexCommand nameref"


def revert_mathrsfs(document):
    r"Load mathrsfs if \mathrsfs us use in the document"
    i = 0
    for line in document.body:
        if line.find("\\mathscr{") != -1:
            add_to_preamble(document, ["\\usepackage{mathrsfs}"])
            return


def convert_flexnames(document):
    "Convert \\begin_inset Flex Custom:Style to \\begin_inset Flex Style and similarly for CharStyle and Element."

    i = 0
    rx = re.compile(r"^\\begin_inset Flex (?:Custom|CharStyle|Element):(.+)$")
    while True:
        i = find_token(document.body, "\\begin_inset Flex", i)
        if i == -1:
            return
        m = rx.match(document.body[i])
        if m:
            document.body[i] = "\\begin_inset Flex " + m.group(1)
        i += 1


flex_insets = {
    "Alert": "CharStyle:Alert",
    "Code": "CharStyle:Code",
    "Concepts": "CharStyle:Concepts",
    "E-Mail": "CharStyle:E-Mail",
    "Emph": "CharStyle:Emph",
    "Expression": "CharStyle:Expression",
    "Initial": "CharStyle:Initial",
    "Institute": "CharStyle:Institute",
    "Meaning": "CharStyle:Meaning",
    "Noun": "CharStyle:Noun",
    "Strong": "CharStyle:Strong",
    "Structure": "CharStyle:Structure",
    "ArticleMode": "Custom:ArticleMode",
    "Endnote": "Custom:Endnote",
    "Glosse": "Custom:Glosse",
    "PresentationMode": "Custom:PresentationMode",
    "Tri-Glosse": "Custom:Tri-Glosse",
}

flex_elements = {
    "Abbrev": "Element:Abbrev",
    "CCC-Code": "Element:CCC-Code",
    "Citation-number": "Element:Citation-number",
    "City": "Element:City",
    "Code": "Element:Code",
    "CODEN": "Element:CODEN",
    "Country": "Element:Country",
    "Day": "Element:Day",
    "Directory": "Element:Directory",
    "Dscr": "Element:Dscr",
    "Email": "Element:Email",
    "Emph": "Element:Emph",
    "Filename": "Element:Filename",
    "Firstname": "Element:Firstname",
    "Fname": "Element:Fname",
    "GuiButton": "Element:GuiButton",
    "GuiMenu": "Element:GuiMenu",
    "GuiMenuItem": "Element:GuiMenuItem",
    "ISSN": "Element:ISSN",
    "Issue-day": "Element:Issue-day",
    "Issue-months": "Element:Issue-months",
    "Issue-number": "Element:Issue-number",
    "KeyCap": "Element:KeyCap",
    "KeyCombo": "Element:KeyCombo",
    "Keyword": "Element:Keyword",
    "Literal": "Element:Literal",
    "MenuChoice": "Element:MenuChoice",
    "Month": "Element:Month",
    "Orgdiv": "Element:Orgdiv",
    "Orgname": "Element:Orgname",
    "Postcode": "Element:Postcode",
    "SS-Code": "Element:SS-Code",
    "SS-Title": "Element:SS-Title",
    "State": "Element:State",
    "Street": "Element:Street",
    "Surname": "Element:Surname",
    "Volume": "Element:Volume",
    "Year": "Element:Year",
}


def revert_flexnames(document):
    if document.backend == "latex":
        flexlist = flex_insets
    else:
        flexlist = flex_elements

    rx = re.compile(r"^\\begin_inset Flex\s+(.+)$")
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Flex", i)
        if i == -1:
            return
        m = rx.match(document.body[i])
        if not m:
            document.warning("Illegal flex inset: " + document.body[i])
            i += 1
            continue
        style = m.group(1)
        if style in flexlist:
            document.body[i] = "\\begin_inset Flex " + flexlist[style]
        i += 1


def convert_mathdots(document):
    "Load mathdots automatically"
    i = find_token(document.header, "\\use_mhchem", 0)
    if i == -1:
        i = find_token(document.header, "\\use_esint", 0)
    if i == -1:
        document.warning("Malformed LyX document: Can't find \\use_mhchem.")
        return
    j = find_token(document.preamble, "\\usepackage{mathdots}", 0)
    if j == -1:
        document.header.insert(i + 1, "\\use_mathdots 0")
    else:
        document.header.insert(i + 1, "\\use_mathdots 2")
        del document.preamble[j]


def revert_mathdots(document):
    "Load mathdots if used in the document"

    mathdots = find_token(document.header, "\\use_mathdots", 0)
    if mathdots == -1:
        document.warning("No \\use_mathdots line. Assuming auto.")
    else:
        val = get_value(document.header, "\\use_mathdots", mathdots)
        del document.header[mathdots]
        try:
            usedots = int(val)
        except:
            document.warning("Invalid \\use_mathdots value: " + val + ". Assuming auto.")
            # probably usedots has not been changed, but be safe.
            usedots = 1

        if usedots == 0:
            # do not load case
            return
        if usedots == 2:
            # force load case
            add_to_preamble(document, ["\\usepackage{mathdots}"])
            return

    # so we are in the auto case. we want to load mathdots if \iddots is used.
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Formula", i)
        if i == -1:
            return
        j = find_end_of_inset(document.body, i)
        if j == -1:
            document.warning(
                "Malformed LyX document: Can't find end of Formula inset at line " + str(i)
            )
            i += 1
            continue
        code = "\n".join(document.body[i:j])
        if code.find("\\iddots") != -1:
            add_to_preamble(document, ["\\@ifundefined{iddots}{\\usepackage{mathdots}}"])
            return
        i = j


def convert_rule(document):
    "Convert \\lyxline to CommandInset line."
    i = 0

    inset = [
        "\\begin_inset CommandInset line",
        "LatexCommand rule",
        'offset "0.5ex"',
        'width "100line%"',
        'height "1pt"',
        "",
        "\\end_inset",
        "",
        "",
    ]

    # if paragraphs are indented, we may have to unindent to get the
    # line to be full-width.
    indent = get_value(document.header, "\\paragraph_separation", 0)
    have_indent = indent == "indent"

    while True:
        i = find_token(document.body, "\\lyxline", i)
        if i == -1:
            return

        # we need to find out if this line follows other content
        # in its paragraph. find its layout....
        lastlay = find_token_backwards(document.body, "\\begin_layout", i)
        if lastlay == -1:
            document.warning("Can't find layout for line at " + str(i))
            # do the best we can.
            document.body[i : i + 1] = inset
            i += len(inset)
            continue

        # ...and look for other content before it.
        lineisfirst = True
        for line in document.body[lastlay + 1 : i]:
            # is it empty or a paragraph option?
            if not line or line[0] == "\\":
                continue
            lineisfirst = False
            break

        if lineisfirst:
            document.body[i : i + 1] = inset
            if indent:
                # we need to unindent, lest the line be too long
                document.body.insert(lastlay + 1, "\\noindent")
            i += len(inset)
        else:
            # so our line is in the middle of a paragraph
            # we need to add a new line, lest this line follow the
            # other content on that line and run off the side of the page
            document.body[i : i + 1] = inset
            document.body[i:i] = ["\\begin_inset Newline newline", "\\end_inset", ""]
        i += len(inset)


def revert_rule(document):
    "Revert line insets to Tex code"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset CommandInset line", i)
        if i == -1:
            return
        # find end of inset
        j = find_token(document.body, "\\end_inset", i)
        if j == -1:
            document.warning("Malformed LyX document: Can't find end of line inset.")
            return
        # determine the optional offset
        offset = get_quoted_value(document.body, "offset", i, j)
        if offset:
            offset = "[" + offset + "]"
        # determine the width
        width = get_quoted_value(document.body, "width", i, j, "100col%")
        width = latex_length(width)[1]
        # determine the height
        height = get_quoted_value(document.body, "height", i, j, "1pt")
        height = latex_length(height)[1]
        # output the \rule command
        subst = "\\rule[" + offset + "]{" + width + "}{" + height + "}"
        document.body[i : j + 1] = put_cmd_in_ert(subst)
        i += len(subst) - (j - i)


def revert_diagram(document):
    "Add the feyn package if \\Diagram is used in math"
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Formula", i)
        if i == -1:
            return
        j = find_end_of_inset(document.body, i)
        if j == -1:
            document.warning("Malformed LyX document: Can't find end of Formula inset.")
            return
        lines = "\n".join(document.body[i:j])
        if lines.find("\\Diagram") == -1:
            i = j
            continue
        add_to_preamble(document, ["\\usepackage{feyn}"])
        # only need to do it once!
        return


chapters = (
    "amsbook",
    "book",
    "docbook-book",
    "elsart",
    "extbook",
    "extreport",
    "jbook",
    "jreport",
    "jsbook",
    "literate-book",
    "literate-report",
    "memoir",
    "mwbk",
    "mwrep",
    "recipebook",
    "report",
    "scrbook",
    "scrreprt",
    "svmono",
    "svmult",
    "tbook",
    "treport",
    "tufte-book",
)


def convert_bibtex_clearpage(document):
    "insert a clear(double)page before bibliographystyle if bibtotoc option is used"

    if document.textclass not in chapters:
        return

    i = find_token(document.header, "\\papersides", 0)
    sides = 0
    if i == -1:
        document.warning("Malformed LyX document: Can't find papersides definition.")
        document.warning("Assuming single sided.")
        sides = 1
    else:
        val = get_value(document.header, "\\papersides", i)
        try:
            sides = int(val)
        except:
            pass
        if sides != 1 and sides != 2:
            document.warning("Invalid papersides value: " + val)
            document.warning("Assuming single sided.")
            sides = 1

    j = 0
    while True:
        j = find_token(document.body, "\\begin_inset CommandInset bibtex", j)
        if j == -1:
            return

        k = find_end_of_inset(document.body, j)
        if k == -1:
            document.warning("Can't find end of Bibliography inset at line " + str(j))
            j += 1
            continue

        # only act if there is the option "bibtotoc"
        val = get_value(document.body, "options", j, k)
        if not val:
            document.warning("Can't find options for bibliography inset at line " + str(j))
            j = k
            continue

        if val.find("bibtotoc") == -1:
            j = k
            continue

        # so we want to insert a new page right before the paragraph that
        # this bibliography thing is in.
        lay = find_token_backwards(document.body, "\\begin_layout", j)
        if lay == -1:
            document.warning(
                "Can't find layout containing bibliography inset at line " + str(j)
            )
            j = k
            continue

        if sides == 1:
            cmd = "clearpage"
        else:
            cmd = "cleardoublepage"
        subst = [
            "\\begin_layout Standard",
            "\\begin_inset Newpage " + cmd,
            "\\end_inset",
            "",
            "",
            "\\end_layout",
            "",
        ]
        document.body[lay:lay] = subst
        j = k + len(subst)


def revert_bibtex_clearpage(document):
    "remove clear(double)page before bibliographystyle if bibtotoc option is used"

    if document.textclass not in chapters:
        return

    i = find_token(document.header, "\\papersides", 0)
    sides = 0
    if i == -1:
        document.warning("Malformed LyX document: Can't find papersides definition.")
        document.warning("Assuming single sided.")
        sides = 1
    else:
        val = get_value(document.header, "\\papersides", i)
        try:
            sides = int(val)
        except:
            pass
        if sides != 1 and sides != 2:
            document.warning("Invalid papersides value: " + val)
            document.warning("Assuming single sided.")
            sides = 1

    j = 0
    while True:
        j = find_token(document.body, "\\begin_inset CommandInset bibtex", j)
        if j == -1:
            return

        k = find_end_of_inset(document.body, j)
        if k == -1:
            document.warning("Can't find end of Bibliography inset at line " + str(j))
            j += 1
            continue

        # only act if there is the option "bibtotoc"
        val = get_value(document.body, "options", j, k)
        if not val:
            document.warning("Can't find options for bibliography inset at line " + str(j))
            j = k
            continue

        if val.find("bibtotoc") == -1:
            j = k
            continue

        # we had inserted \\clear[double]page right before the paragraph that
        # this bibliography thing is in. Remove this. The older format has the
        # respective command hardcoded.
        lay = find_token_backwards(document.body, "\\begin_layout", j)
        if lay == -1:
            document.warning(
                "Can't find layout containing bibliography inset at line " + str(j)
            )
            j = k
            continue
        # Find the layout before this.
        lay = find_token_backwards(document.body, "\\begin_layout", lay - 1)
        if lay == -1:
            document.warning("Can't find layout before bibliography inset at line " + str(j))
            j = k
            continue

        if sides == 1:
            cmd = "clearpage"
        else:
            cmd = "cleardoublepage"

        if (
            document.body[lay] != "\\begin_layout Standard"
            or document.body[lay + 1] != "\\begin_inset Newpage " + cmd
        ):
            j = k
            continue
        layend = find_end_of_layout(document.body, lay)
        if layend == -1:
            document.warning(
                "Can't find end of layout containg newpage inset at line " + str(layend)
            )
            j += 1
            continue

        del document.body[lay : layend + 1]
        j = lay


def check_passthru(document):
    tc = document.textclass
    ok = tc == "literate-article" or tc == "literate-book" or tc == "literate-report"
    if not ok:
        mods = document.get_module_list()
        for mod in mods:
            if mod == "sweave" or mod == "noweb":
                ok = True
                break
    return ok


def convert_passthru(document):
    "http://www.mail-archive.com/lyx-devel@lists.lyx.org/msg161298.html"
    if not check_passthru:
        return

    rx = re.compile("\\\\begin_layout \\s*(\\w+)")
    beg = 0
    for lay in ["Chunk", "Scrap"]:
        while True:
            beg = find_token(document.body, "\\begin_layout " + lay, beg)
            if beg == -1:
                break
            end = find_end_of_layout(document.body, beg)
            if end == -1:
                document.warning("Can't find end of layout at line " + str(beg))
                beg += 1
                continue

            # we are now going to replace newline insets within this layout
            # by new instances of this layout. so we have repeated layouts
            # instead of newlines.

            # if the paragraph has any customization, however, we do not want to
            # do the replacement.
            if document.body[beg + 1].startswith("\\"):
                beg = end + 1
                continue

            ns = beg
            while True:
                ns = find_token(document.body, "\\begin_inset Newline newline", ns, end)
                if ns == -1:
                    break
                ne = find_end_of_inset(document.body, ns)
                if ne == -1 or ne > end:
                    document.warning("Can't find end of inset at line " + str(ne))
                    ns += 1
                    continue
                if document.body[ne + 1] == "":
                    ne += 1
                subst = ["\\end_layout", "", "\\begin_layout " + lay]
                document.body[ns : ne + 1] = subst
                # now we need to adjust end, in particular, but might as well
                # do ns properly, too
                newlines = (ne - ns) - len(subst)
                ns += newlines + 2
                end += newlines + 2

            # ok, we now want to find out if the next layout is the
            # same as this one. if so, we will insert an extra copy of it
            didit = False
            next = find_token(document.body, "\\begin_layout", end)
            if next != -1:
                m = rx.match(document.body[next])
                if m:
                    nextlay = m.group(1)
                    if nextlay == lay:
                        subst = ["\\begin_layout " + lay, "", "\\end_layout", ""]
                        document.body[next:next] = subst
                        didit = True
            beg = end + 1
            if didit:
                beg += 4  # for the extra layout


def revert_passthru(document):
    "http://www.mail-archive.com/lyx-devel@lists.lyx.org/msg161298.html"
    if not check_passthru:
        return
    rx = re.compile("\\\\begin_layout \\s*(\\w+)")
    beg = 0
    for lay in ["Chunk", "Scrap"]:
        while True:
            beg = find_token(document.body, "\\begin_layout " + lay, beg)
            if beg == -1:
                break
            end = find_end_of_layout(document.body, beg)
            if end == -1:
                document.warning("Can't find end of layout at line " + str(beg))
                beg += 1
                continue

            # we now want to find out if the next layout is the
            # same as this one. but we will need to do this over and
            # over again.
            while True:
                next = find_token(document.body, "\\begin_layout", end)
                if next == -1:
                    break
                m = rx.match(document.body[next])
                if not m:
                    break
                nextlay = m.group(1)
                if nextlay != lay:
                    break
                # so it is the same layout again. we now want to know if it is empty.
                # but first let's check and make sure there is no content between the
                # two layouts. i'm not sure if that can happen or not.
                for l in range(end + 1, next):
                    if document.body[l] != "":
                        document.warning("Found content between adjacent " + lay + " layouts!")
                        break
                nextend = find_end_of_layout(document.body, next)
                if nextend == -1:
                    document.warning("Can't find end of layout at line " + str(next))
                    break
                empty = True
                for l in range(next + 1, nextend):
                    if document.body[l] != "":
                        empty = False
                        break
                if empty:
                    # empty layouts just get removed
                    # should we check if it's before yet another such layout?
                    del document.body[next : nextend + 1]
                    # and we do not want to check again. we know the next layout
                    # should be another Chunk and should be left as is.
                    break
                else:
                    # if it's not empty, then we want to insert a newline in place
                    # of the layout switch
                    subst = ["\\begin_inset Newline newline", "\\end_inset", ""]
                    document.body[end : next + 1] = subst
                    # and now we have to find the end of the new, larger layout
                    newend = find_end_of_layout(document.body, beg)
                    if newend == -1:
                        document.warning("Can't find end of new layout at line " + str(beg))
                        break
                    end = newend
            beg = end + 1


def revert_multirowOffset(document):
    "Revert multirow cells with offset in tables to TeX-code"
    # this routine is the same as the revert_multirow routine except that
    # it checks additionally for the offset

    # first, let's find out if we need to do anything
    i = find_token(document.body, '<cell multirow="3" mroffset=', 0)
    if i == -1:
        return

    add_to_preamble(document, ["\\usepackage{multirow}"])

    rgx = re.compile(r'mroffset="[^"]+?"')
    begin_table = 0

    while True:
        # find begin/end of table
        begin_table = find_token(document.body, "<lyxtabular version=", begin_table)
        if begin_table == -1:
            break
        end_table = find_end_of(document.body, begin_table, "<lyxtabular", "</lyxtabular>")
        if end_table == -1:
            document.warning("Malformed LyX document: Could not find end of table.")
            begin_table += 1
            continue
        # does this table have multirow?
        i = find_token(document.body, '<cell multirow="3"', begin_table, end_table)
        if i == -1:
            begin_table = end_table
            continue

        # store the number of rows and columns
        numrows = get_option_value(document.body[begin_table], "rows")
        numcols = get_option_value(document.body[begin_table], "columns")
        try:
            numrows = int(numrows)
            numcols = int(numcols)
        except:
            document.warning("Unable to determine rows and columns!")
            begin_table = end_table
            continue

        mrstarts = []
        multirows = []
        # collect info on rows and columns of this table.
        begin_row = begin_table
        for row in range(numrows):
            begin_row = find_token(document.body, "<row>", begin_row, end_table)
            if begin_row == -1:
                document.warning("Can't find row " + str(row + 1))
                break
            end_row = find_end_of(document.body, begin_row, "<row>", "</row>")
            if end_row == -1:
                document.warning("Can't find end of row " + str(row + 1))
                break
            begin_cell = begin_row
            multirows.append([])
            for column in range(numcols):
                begin_cell = find_token(document.body, "<cell ", begin_cell, end_row)
                if begin_cell == -1:
                    document.warning(
                        "Can't find column " + str(column + 1) + "in row " + str(row + 1)
                    )
                    break
                # NOTE
                # this will fail if someone puts "</cell>" in a cell, but
                # that seems fairly unlikely.
                end_cell = find_end_of(document.body, begin_cell, "<cell", "</cell>")
                if end_cell == -1:
                    document.warning(
                        "Can't find end of column " + str(column + 1) + "in row " + str(row + 1)
                    )
                    break
                multirows[row].append([begin_cell, end_cell, 0])
                if document.body[begin_cell].find('multirow="3" mroffset=') != -1:
                    multirows[row][column][2] = 3  # begin multirow
                    mrstarts.append([row, column])
                elif document.body[begin_cell].find('multirow="4"') != -1:
                    multirows[row][column][2] = 4  # in multirow
                begin_cell = end_cell
            begin_row = end_row
        # end of table info collection

        # work from the back to avoid messing up numbering
        mrstarts.reverse()
        for m in mrstarts:
            row = m[0]
            col = m[1]
            # get column width
            col_width = get_option_value(document.body[begin_table + 2 + col], "width")
            # "0pt" means that no width is specified
            if not col_width or col_width == "0pt":
                col_width = "*"
            # determine the number of cells that are part of the multirow
            nummrs = 1
            for r in range(row + 1, numrows):
                if multirows[r][col][2] != 4:
                    break
                nummrs += 1
                # take the opportunity to revert this line
                lineno = multirows[r][col][0]
                document.body[lineno] = (
                    document.body[lineno]
                    .replace(' multirow="4" ', " ")
                    .replace('valignment="middle"', 'valignment="top"')
                    .replace(' topline="true" ', " ")
                )
                # remove bottom line of previous multirow-part cell
                lineno = multirows[r - 1][col][0]
                document.body[lineno] = document.body[lineno].replace(
                    ' bottomline="true" ', " "
                )
            # revert beginning cell
            bcell = multirows[row][col][0]
            ecell = multirows[row][col][1]
            offset = get_option_value(document.body[bcell], "mroffset")
            document.body[bcell] = (
                document.body[bcell]
                .replace(' multirow="3" ', " ")
                .replace('valignment="middle"', 'valignment="top"')
            )
            # remove mroffset option
            document.body[bcell] = rgx.sub("", document.body[bcell])

            blay = find_token(document.body, "\\begin_layout", bcell, ecell)
            if blay == -1:
                document.warning("Can't find layout for cell!")
                continue
            bend = find_end_of_layout(document.body, blay)
            if bend == -1:
                document.warning("Can't find end of layout for cell!")
                continue
            # do the later one first, so as not to mess up the numbering
            # we are wrapping the whole cell in this ert
            # so before the end of the layout...
            document.body[bend:bend] = put_cmd_in_ert("}")
            # ...and after the beginning
            document.body[blay + 1 : blay + 1] = put_cmd_in_ert(
                "\\multirow{" + str(nummrs) + "}{" + col_width + "}[" + offset + "]{"
            )

        # on to the next table
        begin_table = end_table


def revert_script(document):
    "Convert subscript/superscript inset to TeX code"
    i = 0
    foundsubscript = False
    while True:
        i = find_token(document.body, "\\begin_inset script", i)
        if i == -1:
            break
        z = find_end_of_inset(document.body, i)
        if z == -1:
            document.warning("Malformed LyX document: Can't find end of script inset.")
            i += 1
            continue
        blay = find_token(document.body, "\\begin_layout", i, z)
        if blay == -1:
            document.warning("Malformed LyX document: Can't find layout in script inset.")
            i = z
            continue

        if check_token(document.body[i], "\\begin_inset script subscript"):
            subst = "\\textsubscript{"
            foundsubscript = True
        elif check_token(document.body[i], "\\begin_inset script superscript"):
            subst = "\\textsuperscript{"
        else:
            document.warning("Malformed LyX document: Unknown type of script inset.")
            i = z
            continue
        bend = find_end_of_layout(document.body, blay)
        if bend == -1 or bend > z:
            document.warning(
                "Malformed LyX document: Can't find end of layout in script inset."
            )
            i = z
            continue
        # remove the \end_layout \end_inset pair
        document.body[bend : z + 1] = put_cmd_in_ert("}")
        document.body[i : blay + 1] = put_cmd_in_ert(subst)
        i += 1
    # these classes provide a \textsubscript command:
    # FIXME: Would be nice if we could use the information of the .layout file here
    classes = ["memoir", "scrartcl", "scrbook", "scrlttr2", "scrreprt"]
    if foundsubscript and find_token_exact(classes, document.textclass, 0) == -1:
        add_to_preamble(document, ["\\usepackage{subscript}"])


def convert_use_xetex(document):
    "convert \\use_xetex to \\use_non_tex_fonts"
    i = find_token(document.header, "\\use_xetex", 0)
    if i == -1:
        document.header.insert(-1, "\\use_non_tex_fonts 0")
    else:
        val = get_value(document.header, "\\use_xetex", 0)
        document.header[i] = "\\use_non_tex_fonts " + val


def revert_use_xetex(document):
    "revert \\use_non_tex_fonts to \\use_xetex"
    i = 0
    i = find_token(document.header, "\\use_non_tex_fonts", 0)
    if i == -1:
        document.warning("Malformed document. No \\use_non_tex_fonts param!")
        return

    val = get_value(document.header, "\\use_non_tex_fonts", 0)
    document.header[i] = "\\use_xetex " + val


def revert_labeling(document):
    koma = (
        "scrartcl",
        "scrarticle-beamer",
        "scrbook",
        "scrlettr",
        "scrlttr2",
        "scrreprt",
    )
    if document.textclass in koma:
        return
    i = 0
    while True:
        i = find_token_exact(document.body, "\\begin_layout Labeling", i)
        if i == -1:
            return
        document.body[i] = "\\begin_layout List"


def revert_langpack(document):
    "revert \\language_package parameter"
    i = 0
    i = find_token(document.header, "\\language_package", 0)
    if i == -1:
        document.warning("Malformed document. No \\language_package param!")
        return

    del document.header[i]


def convert_langpack(document):
    "Add \\language_package parameter"
    i = find_token(document.header, r"\language", 0)
    if i == -1:
        document.warning("Malformed document. No \\language defined!")
        return

    document.header.insert(i + 1, "\\language_package default")


def revert_tabularwidth(document):
    i = 0
    while True:
        i = find_token(document.body, "\\begin_inset Tabular", i)
        if i == -1:
            return
        j = find_end_of_inset(document.body, i)
        if j == -1:
            document.warning("Unable to find end of Tabular inset at line " + str(i))
            i += 1
            continue
        i += 1
        features = find_token(document.body, "<features", i, j)
        if features == -1:
            document.warning("Can't find any features in Tabular inset at line " + str(i))
            i = j
            continue
        if document.body[features].find('alignment="tabularwidth"') != -1:
            remove_option(document.body, features, "tabularwidth")


def revert_html_css_as_file(document):
    if not del_token(document.header, "\\html_css_as_file", 0):
        document.warning("Malformed LyX document: Missing \\html_css_as_file.")


##
# Conversion hub
#

supported_versions = ["2.0.0", "2.0"]
convert = [
    [346, []],
    [347, []],
    [348, []],
    [349, []],
    [350, []],
    [351, []],
    [352, [convert_splitindex]],
    [353, []],
    [354, []],
    [355, [convert_strikeout]],
    [356, []],
    [357, [convert_ulinelatex]],
    [358, []],
    [359, [convert_nomencl_width]],
    [360, []],
    [361, []],
    [362, []],
    [363, []],
    [364, []],
    [365, []],
    [366, []],
    [367, []],
    [368, []],
    [369, [convert_author_id]],
    [370, []],
    [371, [convert_mhchem]],
    [372, []],
    [373, [merge_gbrief]],
    [374, []],
    [375, []],
    [376, [convert_includeall]],
    [377, []],
    [378, []],
    [379, [convert_math_output]],
    [380, []],
    [381, []],
    [382, []],
    [383, []],
    [384, []],
    [385, []],
    [386, []],
    [387, []],
    [388, []],
    [389, [convert_html_quotes]],
    [390, []],
    [391, []],
    [392, []],
    [393, [convert_optarg]],
    [394, [convert_use_makebox]],
    [395, []],
    [396, []],
    [397, [remove_Nameref]],
    [398, []],
    [399, [convert_mathdots]],
    [400, [convert_rule]],
    [401, []],
    [402, [convert_bibtex_clearpage]],
    [403, [convert_flexnames]],
    [404, [convert_prettyref]],
    [405, []],
    [406, [convert_passthru]],
    [407, []],
    [408, []],
    [409, [convert_use_xetex]],
    [410, []],
    [411, [convert_langpack]],
    [412, []],
    [413, []],
]

revert = [
    [412, [revert_html_css_as_file]],
    [411, [revert_tabularwidth]],
    [410, [revert_langpack]],
    [409, [revert_labeling]],
    [408, [revert_use_xetex]],
    [407, [revert_script]],
    [406, [revert_multirowOffset]],
    [405, [revert_passthru]],
    [404, []],
    [403, [revert_refstyle]],
    [402, [revert_flexnames]],
    [401, [revert_bibtex_clearpage]],
    [400, [revert_diagram]],
    [399, [revert_rule]],
    [398, [revert_mathdots]],
    [397, [revert_mathrsfs]],
    [396, []],
    [395, [revert_nameref]],
    [394, [revert_DIN_C_pagesizes]],
    [393, [revert_makebox]],
    [392, [revert_argument]],
    [391, []],
    [390, [revert_align_decimal, revert_IEEEtran]],
    [389, [revert_output_sync]],
    [388, [revert_html_quotes]],
    [387, [revert_pagesizes]],
    [386, [revert_math_scale]],
    [385, [revert_lyx_version]],
    [384, [revert_shadedboxcolor]],
    [383, [revert_fontcolor]],
    [382, [revert_turkmen]],
    [381, [revert_notefontcolor]],
    [380, [revert_equalspacing_xymatrix]],
    [379, [revert_inset_preview]],
    [378, [revert_math_output]],
    [377, []],
    [376, [revert_multirow]],
    [375, [revert_includeall]],
    [374, [revert_includeonly]],
    [373, [revert_html_options]],
    [372, [revert_gbrief]],
    [371, [revert_fontenc]],
    [370, [revert_mhchem]],
    [369, [revert_suppress_date]],
    [368, [revert_author_id]],
    [367, [revert_hspace_glue_lengths]],
    [366, [revert_percent_vspace_lengths, revert_percent_hspace_lengths]],
    [365, [revert_percent_skip_lengths]],
    [364, [revert_paragraph_indentation]],
    [363, [revert_branch_filename]],
    [362, [revert_longtable_align]],
    [361, [revert_applemac]],
    [360, []],
    [359, [revert_nomencl_cwidth]],
    [358, [revert_nomencl_width]],
    [357, [revert_custom_processors]],
    [356, [revert_ulinelatex]],
    [355, []],
    [354, [revert_strikeout]],
    [353, [revert_printindexall]],
    [352, [revert_subindex]],
    [351, [revert_splitindex]],
    [350, [revert_backgroundcolor]],
    [349, [revert_outputformat]],
    [348, [revert_xetex]],
    [347, [revert_phantom, revert_hphantom, revert_vphantom]],
    [346, [revert_tabularvalign]],
    [345, [revert_swiss]],
]


if __name__ == "__main__":
    pass