eric6/UI/Previewers/PreviewerHTML.py

# -*- coding: utf-8 -*-

# Copyright (c) 2014 - 2019 Detlev Offenbach <detlev@die-offenbachs.de>
#

"""
Module implementing a previewer widget for HTML, Markdown and ReST files.
"""


import os
import threading
import re
import shutil
import tempfile
import sys
import io

from PyQt5.QtCore import pyqtSlot, pyqtSignal, Qt, QUrl, QThread
from PyQt5.QtGui import QCursor
from PyQt5.QtWidgets import (
    QWidget, QVBoxLayout, QLabel, QCheckBox, QSizePolicy, QToolTip
)

from E5Gui.E5Application import e5App

import Utilities
import Preferences


class PreviewerHTML(QWidget):
    """
    Class implementing a previewer widget for HTML, Markdown and ReST files.
    """
    def __init__(self, parent=None):
        """
        Constructor
        
        @param parent reference to the parent widget (QWidget)
        """
        super(PreviewerHTML, self).__init__(parent)
        
        self.__layout = QVBoxLayout(self)
        
        self.titleLabel = QLabel(self)
        self.titleLabel.setWordWrap(True)
        self.titleLabel.setTextInteractionFlags(Qt.NoTextInteraction)
        self.__layout.addWidget(self.titleLabel)
        
        self.__previewAvailable = True
        
        try:
            from PyQt5.QtWebEngineWidgets import QWebEngineView
            self.previewView = QWebEngineView(self)
            self.previewView.page().linkHovered.connect(self.__showLink)
        except ImportError:
            self.__previewAvailable = False
            self.titleLabel.setText(self.tr(
                "<b>HTML Preview is not available!<br/>"
                "Install QtWebEngine.</b>"))
            self.titleLabel.setAlignment(Qt.AlignHCenter)
            self.__layout.addStretch()
            return
        
        sizePolicy = QSizePolicy(QSizePolicy.Preferred, QSizePolicy.Expanding)
        sizePolicy.setHorizontalStretch(0)
        sizePolicy.setVerticalStretch(0)
        sizePolicy.setHeightForWidth(
            self.previewView.sizePolicy().hasHeightForWidth())
        self.previewView.setSizePolicy(sizePolicy)
        self.previewView.setContextMenuPolicy(Qt.NoContextMenu)
        self.previewView.setUrl(QUrl("about:blank"))
        self.__layout.addWidget(self.previewView)
        
        self.jsCheckBox = QCheckBox(self.tr("Enable JavaScript"), self)
        self.jsCheckBox.setToolTip(self.tr(
            "Select to enable JavaScript for HTML previews"))
        self.__layout.addWidget(self.jsCheckBox)
        
        self.ssiCheckBox = QCheckBox(self.tr("Enable Server Side Includes"),
                                     self)
        self.ssiCheckBox.setToolTip(self.tr(
            "Select to enable support for Server Side Includes"))
        self.__layout.addWidget(self.ssiCheckBox)
        
        self.jsCheckBox.clicked[bool].connect(self.on_jsCheckBox_clicked)
        self.ssiCheckBox.clicked[bool].connect(self.on_ssiCheckBox_clicked)
        self.previewView.titleChanged.connect(self.on_previewView_titleChanged)
        
        self.jsCheckBox.setChecked(
            Preferences.getUI("ShowFilePreviewJS"))
        self.ssiCheckBox.setChecked(
            Preferences.getUI("ShowFilePreviewSSI"))
        
        self.__scrollBarPositions = {}
        self.__vScrollBarAtEnd = {}
        self.__hScrollBarAtEnd = {}
        
        self.__processingThread = PreviewProcessingThread()
        self.__processingThread.htmlReady.connect(self.__setHtml)

        self.__previewedPath = None
        self.__previewedEditor = None
    
    def shutdown(self):
        """
        Public method to perform shutdown actions.
        """
        if self.__previewAvailable:
            self.__processingThread.wait()
    
    @pyqtSlot(bool)
    def on_jsCheckBox_clicked(self, checked):
        """
        Private slot to enable/disable JavaScript.
        
        @param checked state of the checkbox (boolean)
        """
        Preferences.setUI("ShowFilePreviewJS", checked)
        self.__setJavaScriptEnabled(checked)
    
    def __setJavaScriptEnabled(self, enable):
        """
        Private method to enable/disable JavaScript.
        
        @param enable flag indicating the enable state (boolean)
        """
        self.jsCheckBox.setChecked(enable)
        
        settings = self.previewView.settings()
        settings.setAttribute(settings.JavascriptEnabled, enable)
        
        self.processEditor()
    
    @pyqtSlot(bool)
    def on_ssiCheckBox_clicked(self, checked):
        """
        Private slot to enable/disable SSI.
        
        @param checked state of the checkbox (boolean)
        """
        Preferences.setUI("ShowFilePreviewSSI", checked)
        self.processEditor()
    
    @pyqtSlot(str)
    def __showLink(self, urlStr):
        """
        Private slot to show the hovered link in a tooltip.
        
        @param urlStr hovered URL
        @type str
        """
        QToolTip.showText(QCursor.pos(), urlStr, self.previewView)
    
    def processEditor(self, editor=None):
        """
        Public slot to process an editor's text.
        
        @param editor editor to be processed (Editor)
        """
        if not self.__previewAvailable:
            return
        
        if editor is None:
            editor = self.__previewedEditor
        else:
            self.__previewedEditor = editor
        
        if editor is not None:
            fn = editor.getFileName()
            
            if fn:
                extension = os.path.normcase(os.path.splitext(fn)[1][1:])
            else:
                extension = ""
            if (
                extension in Preferences.getEditor(
                    "PreviewHtmlFileNameExtensions") or
                editor.getLanguage() == "HTML"
            ):
                language = "HTML"
            elif (
                extension in Preferences.getEditor(
                    "PreviewMarkdownFileNameExtensions") or
                editor.getLanguage().lower() == "markdown"
            ):
                language = "Markdown"
            elif (
                extension in Preferences.getEditor(
                    "PreviewRestFileNameExtensions") or
                editor.getLanguage().lower() == "restructuredtext"
            ):
                language = "ReST"
            else:
                self.__setHtml(fn, self.tr(
                    "<p>No preview available for this type of file.</p>"))
                return
            
            if fn:
                project = e5App().getObject("Project")
                if project.isProjectFile(fn):
                    rootPath = project.getProjectPath()
                else:
                    rootPath = os.path.dirname(os.path.abspath(fn))
            else:
                rootPath = ""
            
            if bool(editor.text()):
                self.__processingThread.process(
                    fn, language, editor.text(),
                    self.ssiCheckBox.isChecked(), rootPath,
                    Preferences.getEditor("PreviewRestUseSphinx"),
                    Preferences.getEditor("PreviewMarkdownNLtoBR"),
                    Preferences.getEditor(
                        "PreviewMarkdownUsePyMdownExtensions"),
                    Preferences.getEditor("PreviewMarkdownHTMLFormat"),
                    Preferences.getEditor("PreviewRestDocutilsHTMLFormat"))

    def __setHtml(self, filePath, html, rootPath):
        """
        Private method to set the HTML to the view and restore the scroll bars
        positions.
        
        @param filePath file path of the previewed editor
        @type str
        @param html processed HTML text ready to be shown
        @type str
        @param rootPath path of the web site root
        @type str
        """
        self.__previewedPath = Utilities.normcasepath(
            Utilities.fromNativeSeparators(filePath))
        self.__saveScrollBarPositions()
        self.previewView.page().loadFinished.connect(
            self.__restoreScrollBarPositions)
        if not filePath:
            filePath = "/"
        if rootPath:
            baseUrl = QUrl.fromLocalFile(rootPath + "/index.html")
        else:
            baseUrl = QUrl.fromLocalFile(filePath)
        self.previewView.setHtml(html, baseUrl=baseUrl)
        if self.__previewedEditor:
            self.__previewedEditor.setFocus()
    
    @pyqtSlot(str)
    def on_previewView_titleChanged(self, title):
        """
        Private slot to handle a change of the title.
        
        @param title new title (string)
        """
        if title:
            self.titleLabel.setText(self.tr("Preview - {0}").format(title))
        else:
            self.titleLabel.setText(self.tr("Preview"))
    
    def __saveScrollBarPositions(self):
        """
        Private method to save scroll bar positions for a previewed editor.
        """
        from PyQt5.QtCore import QPoint
        try:
            pos = self.previewView.scrollPosition()
        except AttributeError:
            pos = self.__execJavaScript(
                "(function() {"
                "var res = {"
                "    x: 0,"
                "    y: 0,"
                "};"
                "res.x = window.scrollX;"
                "res.y = window.scrollY;"
                "return res;"
                "})()"
            )
            if pos is not None:
                pos = QPoint(pos["x"], pos["y"])
            else:
                pos = QPoint(0, 0)
        self.__scrollBarPositions[self.__previewedPath] = pos
        self.__hScrollBarAtEnd[self.__previewedPath] = False
        self.__vScrollBarAtEnd[self.__previewedPath] = False

    def __restoreScrollBarPositions(self):
        """
        Private method to restore scroll bar positions for a previewed editor.
        """
        if self.__previewedPath not in self.__scrollBarPositions:
            return
        
        pos = self.__scrollBarPositions[self.__previewedPath]
        self.previewView.page().runJavaScript(
            "window.scrollTo({0}, {1});".format(pos.x(), pos.y()))
    
    def __execJavaScript(self, script):
        """
        Private function to execute a JavaScript function Synchroneously.
        
        @param script JavaScript script source to be executed
        @type str
        @return result of the script
        @rtype depending upon script result
        """
        from PyQt5.QtCore import QEventLoop
        loop = QEventLoop()
        resultDict = {"res": None}
        
        def resultCallback(res, resDict=resultDict):
            if loop and loop.isRunning():
                resDict["res"] = res
                loop.quit()
        
        self.previewView.page().runJavaScript(
            script, resultCallback)
        
        loop.exec_()
        return resultDict["res"]


class PreviewProcessingThread(QThread):
    """
    Class implementing a thread to process some text into HTML usable by the
    previewer view.
    
    @signal htmlReady(str, str, str) emitted with the file name, the processed
        HTML and the web site root path to signal the availability of the
        processed HTML
    """
    htmlReady = pyqtSignal(str, str, str)
    
    def __init__(self, parent=None):
        """
        Constructor
        
        @param parent reference to the parent object (QObject)
        """
        super(PreviewProcessingThread, self).__init__()
        
        self.__lock = threading.Lock()
    
    def process(self, filePath, language, text, ssiEnabled, rootPath,
                useSphinx, convertNewLineToBreak, usePyMdownExtensions,
                markdownHtmlFormat, restDocutilsHtmlFormat):
        """
        Public method to convert the given text to HTML.
        
        @param filePath file path of the text
        @type str
        @param language language of the text
        @type str
        @param text text to be processed
        @type str
        @param ssiEnabled flag indicating to do some (limited) SSI processing
        @type bool
        @param rootPath root path to be used for SSI processing
        @type str
        @param useSphinx flag indicating to use Sphinx to generate the
            ReST preview
        @type bool
        @param convertNewLineToBreak flag indicating to convert new lines
            to HTML break (Markdown only)
        @type bool
        @param usePyMdownExtensions flag indicating to enable the PyMdown
            extensions, if they are available
        @type bool
        @param markdownHtmlFormat HTML format to be generated by markdown
        @type str
        @param restDocutilsHtmlFormat HTML format to be generated by docutils
        @type str
        """
        with self.__lock:
            self.__filePath = filePath
            self.__language = language
            self.__text = text
            self.__ssiEnabled = ssiEnabled
            self.__rootPath = rootPath
            self.__haveData = True
            self.__useSphinx = useSphinx
            self.__convertNewLineToBreak = convertNewLineToBreak
            self.__usePyMdownExtensions = usePyMdownExtensions
            self.__markdownHtmlFormat = markdownHtmlFormat
            self.__restDocutilsHtmlFormat = restDocutilsHtmlFormat
            if not self.isRunning():
                self.start(QThread.LowPriority)
    
    def run(self):
        """
        Public thread method to convert the stored data.
        """
        while True:
            # exits with break
            with self.__lock:
                filePath = self.__filePath
                language = self.__language
                text = self.__text
                ssiEnabled = self.__ssiEnabled
                rootPath = self.__rootPath
                useSphinx = self.__useSphinx
                convertNewLineToBreak = self.__convertNewLineToBreak
                usePyMdownExtensions = self.__usePyMdownExtensions
                markdownHtmlFormat = self.__markdownHtmlFormat
                restDocutilsHtmlFormat = self.__restDocutilsHtmlFormat
            
                self.__haveData = False

            html = self.__getHtml(language, text, ssiEnabled, filePath,
                                  rootPath, useSphinx, convertNewLineToBreak,
                                  usePyMdownExtensions, markdownHtmlFormat,
                                  restDocutilsHtmlFormat)
            
            with self.__lock:
                if not self.__haveData:
                    self.htmlReady.emit(filePath, html, rootPath)
                    break
                # else - next iteration
    
    def __getHtml(self, language, text, ssiEnabled, filePath, rootPath,
                  useSphinx, convertNewLineToBreak, usePyMdownExtensions,
                  markdownHtmlFormat, restDocutilsHtmlFormat):
        """
        Private method to process the given text depending upon the given
        language.
        
        @param language language of the text
        @type str
        @param text to be processed
        @type str
        @param ssiEnabled flag indicating to do some (limited) SSI processing
        @type bool
        @param filePath file path of the text
        @type str
        @param rootPath root path to be used for SSI processing
        @type str
        @param useSphinx flag indicating to use Sphinx to generate the
            ReST preview
        @type bool
        @param convertNewLineToBreak flag indicating to convert new lines
            to HTML break (Markdown only)
        @type bool
        @param usePyMdownExtensions flag indicating to enable the PyMdown
            extensions, if they are available
        @type bool
        @param markdownHtmlFormat HTML format to be generated by markdown
        @type str
        @param restDocutilsHtmlFormat HTML format to be generated by docutils
        @type str
        @return processed HTML text
        @rtype str
        """
        if language == "HTML":
            if ssiEnabled:
                html = self.__processSSI(text, filePath, rootPath)
            else:
                html = text
            return self.__processRootPath(html, rootPath)
        elif language == "Markdown":
            return self.__convertMarkdown(
                text, convertNewLineToBreak, usePyMdownExtensions,
                markdownHtmlFormat)
        elif language == "ReST":
            return self.__convertReST(text, useSphinx, restDocutilsHtmlFormat)
        else:
            return self.tr(
                "<p>No preview available for this type of file.</p>")
    
    def __processSSI(self, txt, filename, root):
        """
        Private method to process the given text for SSI statements.
        
        Note: Only a limited subset of SSI statements are supported.
        
        @param txt text to be processed (string)
        @param filename name of the file associated with the given text
            (string)
        @param root directory of the document root (string)
        @return processed HTML (string)
        """
        if not filename:
            return txt
        
        # SSI include
        incRe = re.compile(
            r"""<!--#include[ \t]+(virtual|file)=[\"']([^\"']+)[\"']\s*-->""",
            re.IGNORECASE)
        baseDir = os.path.dirname(os.path.abspath(filename))
        docRoot = root if root != "" else baseDir
        while True:
            incMatch = incRe.search(txt)
            if incMatch is None:
                break
            
            if incMatch.group(1) == "virtual":
                incFile = Utilities.normjoinpath(docRoot, incMatch.group(2))
            elif incMatch.group(1) == "file":
                incFile = Utilities.normjoinpath(baseDir, incMatch.group(2))
            else:
                incFile = ""
            if os.path.exists(incFile):
                try:
                    f = open(incFile, "r")
                    incTxt = f.read()
                    f.close()
                except (IOError, OSError):
                    # remove SSI include
                    incTxt = ""
            else:
                # remove SSI include
                incTxt = ""
            txt = txt[:incMatch.start(0)] + incTxt + txt[incMatch.end(0):]
        
        return txt
    
    def __processRootPath(self, txt, root):
        """
        Private method to adjust absolute references to the given root path.
        
        @param txt text to be processed
        @type str
        @param root directory of the document root
        @type str
        @return processed HTML
        @rtype str
        """
        if not root:
            return txt
        
        root = Utilities.fromNativeSeparators(root)
        if not root.endswith("/"):
            root += "/"
        rootLen = len(root)
        
        refRe = re.compile(
            r"""(href|src)=[\\"']/([^\\"']+)[\\"']""",
            re.IGNORECASE)
        pos = 0
        while True:
            refMatch = refRe.search(txt, pos)
            if refMatch is None:
                break
            
            txt = (txt[:refMatch.start(0)] + refMatch.group(1) + '="' + root +
                   refMatch.group(2) + '"' + txt[refMatch.end(0):])
            pos = refMatch.end(0) + rootLen
        
        return txt
    
    def __convertReST(self, text, useSphinx, restDocutilsHtmlFormat):
        """
        Private method to convert ReST text into HTML.
        
        @param text text to be processed (string)
        @param useSphinx flag indicating to use Sphinx to generate the
            ReST preview (boolean)
        @param restDocutilsHtmlFormat HTML format to be generated by docutils
            (string)
        @return processed HTML (string)
        """
        if useSphinx:
            return self.__convertReSTSphinx(text)
        else:
            return self.__convertReSTDocutils(text, restDocutilsHtmlFormat)
    
    def __convertReSTSphinx(self, text):
        """
        Private method to convert ReST text into HTML using 'sphinx'.
        
        @param text text to be processed (string)
        @return processed HTML (string)
        """
        try:
            from sphinx.application import Sphinx   # __IGNORE_EXCEPTION__
        except ImportError:
            return self.tr(
                """<p>ReStructuredText preview requires the"""
                """ <b>sphinx</b> package.<br/>Install it with"""
                """ your package manager,'pip install Sphinx' or see"""
                """ <a href="http://pypi.python.org/pypi/Sphinx">"""
                """this page.</a></p>"""
                """<p>Alternatively you may disable Sphinx usage"""
                """ on the Editor, Filehandling configuration page.</p>""")
        
        tempDir = tempfile.mkdtemp(prefix='eric-rest-')
        try:
            filename = 'sphinx_preview'
            basePath = os.path.join(tempDir, filename)
            fh = open(basePath + '.rst', 'w', encoding='utf-8')
            fh.write(text)
            fh.close()

            overrides = {'html_add_permalinks': False,
                         'html_copy_source': False,
                         'html_title': 'Sphinx preview',
                         'html_use_index': False,
                         'html_use_modindex': False,
                         'html_use_smartypants': True,
                         'master_doc': filename}
            app = Sphinx(srcdir=tempDir, confdir=None, outdir=tempDir,
                         doctreedir=tempDir, buildername='html',
                         confoverrides=overrides, status=None,
                         warning=io.StringIO())
            app.build(force_all=True, filenames=None)

            fh = open(basePath + '.html', 'r', encoding='utf-8')
            html = fh.read()
            fh.close()
        finally:
            shutil.rmtree(tempDir)
        
        return html
    
    def __convertReSTDocutils(self, text, htmlFormat):
        """
        Private method to convert ReST text into HTML using 'docutils'.
        
        @param text text to be processed (string)
        @param htmlFormat HTML format to be generated (string)
        @return processed HTML (string)
        """
        if 'sphinx' in sys.modules:
            # Make sure any Sphinx polution of docutils has been removed.
            unloadKeys = [k for k in sys.modules.keys()
                          if k.startswith(('docutils', 'sphinx'))]
            for key in unloadKeys:
                sys.modules.pop(key)
        
        try:
            import docutils.core    # __IGNORE_EXCEPTION__
            import docutils.utils   # __IGNORE_EXCEPTION__
        except ImportError:
            return self.tr(
                """<p>ReStructuredText preview requires the"""
                """ <b>python-docutils</b> package.<br/>Install it with"""
                """ your package manager, 'pip install docutils' or see"""
                """ <a href="http://pypi.python.org/pypi/docutils">"""
                """this page.</a></p>""")
        
        # redirect sys.stderr because we are not interested in it here
        origStderr = sys.stderr
        sys.stderr = io.StringIO()
        try:
            html = docutils.core.publish_string(
                text, writer_name=htmlFormat.lower()).decode("utf-8")
        except docutils.utils.SystemMessage as err:
            errStr = str(err).split(":")[-1].replace("\n", "<br/>")
            return self.tr(
                """<p>Docutils returned an error:</p><p>{0}</p>"""
            ).format(errStr)
        
        sys.stderr = origStderr
        return html
    
    def __convertMarkdown(self, text, convertNewLineToBreak,
                          usePyMdownExtensions, htmlFormat):
        """
        Private method to convert Markdown text into HTML.
        
        @param text text to be processed
        @type str
        @param convertNewLineToBreak flag indicating to convert new lines
            to HTML break (Markdown only)
        @type bool
        @param usePyMdownExtensions flag indicating to enable the PyMdown
            extensions, if they are available
        @type bool
        @param htmlFormat HTML format to be generated by markdown
        @type str
        @return processed HTML
        @rtype str
        """
        try:
            import markdown     # __IGNORE_EXCEPTION__
        except ImportError:
            return self.tr(
                """<p>Markdown preview requires the <b>Markdown</b> """
                """package.<br/>Install it with your package manager,"""
                """ 'pip install Markdown' or see """
                """<a href="http://pythonhosted.org/Markdown/install.html">"""
                """installation instructions.</a></p>""")
        
        try:
            import mdx_mathjax  # __IGNORE_EXCEPTION__ __IGNORE_WARNING__
        except ImportError:
            # mathjax doesn't require import statement if installed
            # as extension
            pass
        
        extensions = []
        if usePyMdownExtensions:
            try:
                import pymdownx     # __IGNORE_EXCEPTION__ __IGNORE_WARNING__
                # PyPI package is 'pymdown-extensions'
                
                extensions = [
                    'pymdownx.extra', 'pymdownx.caret', 'pymdownx.emoji',
                    'pymdownx.mark', 'pymdownx.tilde', 'pymdownx.keys',
                    'pymdownx.tasklist', 'pymdownx.smartsymbols',
                ]
                if convertNewLineToBreak:
                    extensions.append('nl2br')
            except ImportError:
                pass
        if not extensions:
            if convertNewLineToBreak:
                extensions = ['extra', 'toc', 'nl2br']
            else:
                extensions = ['extra', 'toc']
            
            # version 2.0 supports only extension names, not instances
            if (
                markdown.version_info[0] > 2 or
                (markdown.version_info[0] == 2 and
                 markdown.version_info[1] > 0)
            ):
                class _TildeExtension(markdown.Extension):
                    """
                    Class is placed here, because it depends on imported
                    markdown, and markdown import is lazy.
                    
                    (see https://pythonhosted.org/Markdown/extensions/api.html
                    this page for details)
                    """
                    DEL_RE = r'(~~)(.+?)~~'
                    SUB_RE = r'(~)(.+?)~'

                    def extendMarkdown(self, md, md_globals):
                        # Create the sub pattern and insert it into markdown
                        # parser
                        sub_tag = markdown.inlinepatterns.SimpleTagPattern(
                            self.SUB_RE, 'sub')
                        md.inlinePatterns.add('sub', sub_tag, '>not_strong')
                        
                        # Create the del pattern and insert it into markdown
                        # parser
                        del_tag = markdown.inlinepatterns.SimpleTagPattern(
                            self.DEL_RE, 'del')
                        md.inlinePatterns.add('del', del_tag, '>not_strong')
                
                class _CaretExtension(markdown.Extension):
                    """
                    Class is placed here, because it depends on imported
                    markdown, and markdown import is lazy.
                    
                    (see https://pythonhosted.org/Markdown/extensions/api.html
                    this page for details)
                    """
                    INS_RE = r'(\^\^)(.*?)\^\^'
                    SUP_RE = r'(\^)(.*?)\^'
                    
                    def extendMarkdown(self, md, md_globals):
                        # Create the sup pattern and insert it into markdown
                        # parser
                        sup_tag = markdown.inlinepatterns.SimpleTagPattern(
                            self.SUP_RE, 'sup')
                        md.inlinePatterns.add('sup', sup_tag, '>not_strong')
                        
                        # Create the ins pattern and insert it into markdown
                        # parser
                        ins_tag = markdown.inlinepatterns.SimpleTagPattern(
                            self.INS_RE, 'ins')
                        md.inlinePatterns.add('ins', ins_tag, '>not_strong')
                
                class _MarkExtension(markdown.Extension):
                    """
                    Class is placed here, because it depends on imported
                    markdown, and markdown import is lazy.
                    
                    (see https://pythonhosted.org/Markdown/extensions/api.html
                    this page for details)
                    """
                    MARK_RE = r'(==)(.*?)=='
                    
                    def extendMarkdown(self, md, md_globals):
                        # Create the mark pattern and insert it into markdown
                        # parser
                        mark_tag = markdown.inlinepatterns.SimpleTagPattern(
                            self.MARK_RE, 'mark')
                        md.inlinePatterns.add('mark', mark_tag, '>not_strong')
                
                extensions.extend([
                    _TildeExtension(), _CaretExtension(), _MarkExtension()
                ])

        try:
            return markdown.markdown(
                text, extensions=extensions + ['mdx_mathjax'],
                output_format=htmlFormat.lower())
        except (ImportError, ValueError):
            # markdown raises ValueError or ImportError, depends on version
            # It is not clear, how to distinguish missing mathjax from other
            # errors. So keep going without mathjax.
            return markdown.markdown(
                text, extensions=extensions,
                output_format=htmlFormat.lower())