view frontends/src/jp/cmd_blog.py @ 1886:f3db27508b31

jp (blog/preview): inotify improvments: - InotifyError are catched (can happen if file is deleted during preview for instance) - handle IN_DELETE_SELF event which happen when editor write content in a new file and switch to it (tested with vim)
author Goffi <goffi@goffi.org>
date Sat, 05 Mar 2016 19:24:10 +0100
parents edd8dc8df1b9
children 16527dd5a81b
line wrap: on
line source

#! /usr/bin/python
# -*- coding: utf-8 -*-

# jp: a SàT command line tool
# Copyright (C) 2009-2016 Jérôme Poisson (goffi@goffi.org)

# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as published by
# the Free Software Foundation, either version 3 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 Affero General Public License for more details.

# You should have received a copy of the GNU Affero General Public License
# along with this program.  If not, see <http://www.gnu.org/licenses/>.


import base
from sat.core.i18n import _
from sat.core.constants import Const as C
from sat.tools import config
from ConfigParser import NoSectionError, NoOptionError
import json
import os.path
import os
import time
import tempfile
import subprocess
import shlex
import glob
from sat.tools import common

__commands__ = ["Blog"]

# extensions to use with known syntaxes
SYNTAX_EXT = {
    '': 'txt', # used when the syntax is not found
    "XHTML": "xhtml",
    "markdown": "md"
    }

# defaut arguments used for some known editors
VIM_SPLIT_ARGS = "-c 'vsplit|wincmd w|next|wincmd w'"
EMACS_SPLIT_ARGS = '--eval "(split-window-horizontally)"'
EDITOR_ARGS_MAGIC = {
    'vim': VIM_SPLIT_ARGS + ' {content_file} {metadata_file}',
    'gvim': VIM_SPLIT_ARGS + ' --nofork {content_file} {metadata_file}',
    'emacs': EMACS_SPLIT_ARGS + ' {content_file} {metadata_file}',
    'xemacs': EMACS_SPLIT_ARGS + ' {content_file} {metadata_file}',
    'nano': ' -F {content_file} {metadata_file}',
    }

CONF_SYNTAX_EXT = 'syntax_ext_dict'
BLOG_TMP_DIR="blog"
METADATA_SUFF = '_metadata.json'
# key to remove from metadata tmp file if they exist
KEY_TO_REMOVE_METADATA = ('id','content', 'content_xhtml', 'comments_node', 'comments_service')

URL_REDIRECT_PREFIX = 'url_redirect_'
INOTIFY_INSTALL = '"pip install inotify"'


class BlogCommon(object):
    def getTmpDir(self, sat_conf):
        """Return directory used to store temporary files

        @param sat_conf(ConfigParser.ConfigParser): instance opened on sat configuration
        @return (str): path to the dir
        """
        local_dir = config.getConfig(sat_conf, '', 'local_dir', Exception)
        return os.path.join(local_dir, BLOG_TMP_DIR)

    def getCurrentFile(self, sat_conf):
        # we guess the blog item currently edited by choosing
        # the most recent file corresponding to temp file pattern
        # in tmp_dir, excluding metadata files
        tmp_dir = self.getTmpDir(sat_conf)
        available = [path for path in glob.glob(os.path.join(tmp_dir, 'blog_*')) if not path.endswith(METADATA_SUFF)]
        if not available:
            self.disp(u"Counldn't find any content draft in {path}".format(path=tmp_dir), error=True)
            self.host.quit(1)
        return max(available, key=lambda path: os.stat(path).st_mtime)

    def guessSyntaxFromPath(self, sat_conf, path):
        """Return syntax guessed according to filename extension

        @param sat_conf(ConfigParser.ConfigParser): instance opened on sat configuration
        @param path(str): path to the content file
        @return(unicode): syntax to use
        """
        # we first try to guess syntax with extension
        ext = os.path.splitext(path)[1][1:] # we get extension without the '.'
        if ext:
            for k,v in SYNTAX_EXT.iteritems():
                if ext == v:
                    return k

        # if not found, we use current syntax
        return self.host.bridge.getParamA("Syntax", "Composition", "value", self.profile)

    def parse_args(self, cmd_line, **format_kw):
        """Parse command arguments

        @param cmd_line(unicode): command line as found in sat.conf
        @param format_kw: keywords used for formmating
        @return (list(unicode)): list of arguments to pass to subprocess function
        """
        try:
            # we split the arguments and add the known fields
            # we split arguments first to avoid escaping issues in file names
            return [a.format(**format_kw) for a in shlex.split(cmd_line)]
        except ValueError as e:
            self.disp(u"Couldn't parse editor cmd [{cmd}]: {reason}".format(cmd=cmd_line, reason=e))
            return []


class Edit(base.CommandBase, BlogCommon):

    def __init__(self, host):
        super(Edit, self).__init__(host, 'edit', use_verbose=True, help=_(u'edit an existing or new blog post'))

    def add_parser_options(self):
        self.parser.add_argument("item", type=base.unicode_decoder, nargs='?', default=u'new', help=_(u"URL of the item to edit, or keyword"))
        self.parser.add_argument("-T", '--title', type=base.unicode_decoder, help=_(u"title of the item"))
        self.parser.add_argument("-t", '--tag', type=base.unicode_decoder, action='append', help=_(u"tag (category) of your item"))
        self.parser.add_argument("--no-comment", action='store_true', help=_(u"disable comments"))

    def getTmpFile(self, sat_conf, tmp_suff):
        """Create a temporary file to received blog item body

        @param sat_conf(ConfigParser.ConfigParser): instance opened on sat configuration
        @param tmp_suff (str): suffix to use for the filename
        @return (tuple(file, str)): opened (w+b) file object and file path
        """
        tmp_dir = self.getTmpDir(sat_conf)
        if not os.path.exists(tmp_dir):
            try:
                os.makedirs(tmp_dir)
            except OSError as e:
                self.disp(u"Can't create {path} directory: {reason}".format(
                    path=tmp_dir, reason=e), error=True)
                self.host.quit(1)
        try:
            fd, path = tempfile.mkstemp(suffix=tmp_suff,
                prefix=time.strftime('blog_%Y-%m-%d_%H:%M:%S_'),
                dir=tmp_dir, text=True)
            return os.fdopen(fd, 'w+b'), path
        except OSError as e:
            self.disp(u"Can't create temporary file: {reason}".format(reason=e), error=True)
            self.host.quit(1)

    def buildMetadataFile(self, content_file_path, mb_data=None):
        """Build a metadata file using json

        The file is named after content_file_path, with extension replaced by _metadata.json
        @param content_file_path(str): path to the temporary file which will contain the body
        @param mb_data(dict, None): microblog metadata (for existing items)
        @return (tuple[dict, str]): merged metadata put originaly in metadata file
            and path to temporary metadata file
        """
        # we first construct metadata from edited item ones and CLI argumments
        # or re-use the existing one if it exists
        meta_file_path = os.path.splitext(content_file_path)[0] + METADATA_SUFF
        if os.path.exists(meta_file_path):
            self.disp(u"Metadata file already exists, we re-use it")
            try:
                with open(meta_file_path, 'rb') as f:
                    mb_data = json.load(f)
            except (OSError, IOError, ValueError) as e:
                self.disp(u"Can't read existing metadata file at {path}, aborting: {reason}".format(
                    path=meta_file_path, reason=e), error=True)
                self.host.quit(1)
        else:
            mb_data = {} if mb_data is None else mb_data.copy()

        # in all cases, we want to remove unwanted keys
        for key in KEY_TO_REMOVE_METADATA:
            try:
                del mb_data[key]
            except KeyError:
                pass
        # and override metadata with command-line arguments
        mb_data['allow_comments'] = C.boolConst(not self.args.no_comment)
        if self.args.tag:
            common.iter2dict('tag', self.args.tag, mb_data)
        if self.args.title is not None:
            mb_data['title'] = self.args.title

        # then we create the file and write metadata there, as JSON dict
        # XXX: if we port jp one day on Windows, O_BINARY may need to be added here
        if os.path.exists(meta_file_path):
            self.disp(u"metadata file {} already exists, this should not happen! Cancelling...", error=True)
        with os.fdopen(os.open(meta_file_path, os.O_RDWR | os.O_CREAT ,0o600), 'w+b') as f:
            # we need to use an intermediate unicode buffer to write to the file unicode without escaping characters
            unicode_dump = json.dumps(mb_data, ensure_ascii=False, indent=4, separators=(',', ': '), sort_keys=True)
            f.write(unicode_dump.encode('utf-8'))

        return mb_data, meta_file_path

    def edit(self, sat_conf, content_file_path, content_file_obj, mb_data=None):
        """Edit the file contening the content using editor, and publish it"""
        item_ori_mb_data = mb_data
        # we first create metadata file
        meta_ori, meta_file_path = self.buildMetadataFile(content_file_path, item_ori_mb_data)

        # then we calculate hashes to check for modifications
        import hashlib
        content_file_obj.seek(0)
        tmp_ori_hash = hashlib.sha1(content_file_obj.read()).digest()
        content_file_obj.close()

        # then we launch editor
        editor = config.getConfig(sat_conf, 'jp', 'editor') or os.getenv('EDITOR', 'vi')
        try:
            # is there custom arguments in sat.conf ?
            editor_args = config.getConfig(sat_conf, 'jp', 'blog_editor_args', Exception)
        except (NoOptionError, NoSectionError):
            # no, we check if we know the editor and have special arguments
            editor_args = EDITOR_ARGS_MAGIC.get(os.path.basename(editor), '')
        args = self.parse_args(editor_args, content_file=content_file_path, metadata_file=meta_file_path)
        if not args:
            args = [content_file_path]
        editor_exit = subprocess.call([editor] + args)

        # we send the file if edition was a success
        if editor_exit != 0:
            self.disp(u"Editor exited with an error code, so temporary file has not be deleted, and blog item is not published.\nTou can find temporary file at {path}".format(
                path=content_file_path), error=True)
        else:
            try:
                with open(content_file_path, 'rb') as f:
                    content = f.read()
                with open(meta_file_path, 'rb') as f:
                    mb_data = json.load(f)
            except (OSError, IOError):
                self.disp(u"Can read files at {content_path} and/or {meta_path}, have they been deleted?\nCancelling edition".format(
                    content_path=content_file_path, meta_path=meta_file_path), error=True)
                self.host.quit(1)
            except ValueError:
                self.disp(u"Can't parse metadata, please check it is correct JSON format. Cancelling edition.\n" +
                    "You can find tmp file at {content_path} and temporary meta file at {meta_path}.".format(
                    content_path=content_file_path, meta_path=meta_file_path), error=True)
                self.host.quit(1)

            if not C.bool(mb_data.get('publish', "true")):
                self.disp(u'Publication blocked by "publish" key in metadata, cancelling edition.\n\n' +
                    "temporary file path:\t{content_path}\nmetadata file path:\t{meta_path}".format(
                    content_path=content_file_path, meta_path=meta_file_path), error=True)
                self.host.quit(0)

            if len(content) == 0:
                self.disp(u"Content is empty, cancelling the blog edition")

            # time to re-check the hash
            elif (tmp_ori_hash == hashlib.sha1(content).digest() and
                  meta_ori == mb_data):
                self.disp(u"The content has not been modified, cancelling the blog edition")

            else:
                # we can now send the blog
                mb_data['content_rich'] =  content.decode('utf-8')

                if item_ori_mb_data is not None:
                    mb_data['id'] = item_ori_mb_data['id']

                try:
                    self.host.bridge.mbSend('', '', mb_data, self.profile)
                except Exception as e:
                    self.disp(u"Error while sending your blog, the temporary files have been kept at {content_path} and {meta_path}: {reason}".format(
                        content_path=content_file_path, meta_path=meta_file_path, reason=e), error=True)
                    self.host.quit(1)

            os.unlink(content_file_path)
            os.unlink(meta_file_path)

    def start(self):
        item_lower = self.args.item.lower()
        sat_conf = config.parseMainConf()
        # if there are user defined extension, we use them
        SYNTAX_EXT.update(config.getConfig(sat_conf, 'jp', CONF_SYNTAX_EXT, {}))
        current_syntax = None

        if item_lower == 'current':
            # use wants to continue current draft
            content_file_path = self.getCurrentFile(sat_conf)
            content_file_obj = open(content_file_path, 'r+b')
            current_syntax = self.guessSyntaxFromPath(sat_conf, content_file_path)
        else:
            # we get current syntax to determine file extension
            current_syntax = self.host.bridge.getParamA("Syntax", "Composition", "value", self.profile)
            # we now create a temporary file
            tmp_suff = '.' + SYNTAX_EXT.get(current_syntax, SYNTAX_EXT[''])
            content_file_obj, content_file_path = self.getTmpFile(sat_conf, tmp_suff)

        self.disp(u"Syntax used: {}".format(current_syntax), 1)

        if item_lower == 'new':
            self.disp(u'Editing a new blog item', 2)
            mb_data = None
        elif item_lower == 'current':
            self.disp(u'Continuing edition of current draft', 2)
            mb_data = None
        elif item_lower == 'last':
            self.disp(u'Editing last published item', 2)
            try:
                mb_data = self.host.bridge.mbGet('', '', 1, [], {}, self.profile)[0][0]
            except Exception as e:
                self.disp(u"Error while retrieving last comment: {}".format(e))
                self.host.quit(1)

            content = mb_data['content_xhtml']
            if content and current_syntax != 'XHTML':
                content = self.host.bridge.syntaxConvert(content, 'XHTML', current_syntax, False, self.profile)
            content_file_obj.write(content.encode('utf-8'))
            content_file_obj.seek(0)

        self.edit(sat_conf, content_file_path, content_file_obj, mb_data=mb_data)


class Preview(base.CommandBase, BlogCommon):

    def __init__(self, host):
        super(Preview, self).__init__(host, 'preview', use_verbose=True, help=_(u'preview a blog content'))

    def add_parser_options(self):
        self.parser.add_argument("--inotify", type=str, choices=('auto', 'true', 'false'), default=u'auto', help=_(u"use inotify to handle preview"))
        self.parser.add_argument("file", type=base.unicode_decoder, nargs='?', default=u'current', help=_(u"path to the content file"))

    def showPreview(self):
        # we implement showPreview here so we don't have to import webbroser and urllib
        # when preview is not used
        url = 'file:{}'.format(self.urllib.quote(self.preview_file_path))
        self.webbrowser.open_new_tab(url)

    def _launchPreviewExt(self, cmd_line, opt_name):
        url = 'file:{}'.format(self.urllib.quote(self.preview_file_path))
        args = self.parse_args(cmd_line, url=url, preview_file=self.preview_file_path)
        if not args:
            self.disp(u"Couln't find command in \"{name}\", abording".format(name=opt_name), error=True)
            self.host.quit(1)
        subprocess.Popen(args)

    def openPreviewExt(self):
        self._launchPreviewExt(self.open_cb_cmd, "blog_preview_open_cmd")

    def updatePreviewExt(self):
        self._launchPreviewExt(self.update_cb_cmd, "blog_preview_update_cmd")

    def updateContent(self):
        with open(self.content_file_path, 'rb') as f:
            content = f.read().decode('utf-8')
            if content and self.syntax != 'XHTML':
                # we use safe=True because we want to have a preview as close as possible to what the
                # people will see
                content = self.host.bridge.syntaxConvert(content, self.syntax, 'XHTML', True, self.profile)

        xhtml = (u'<html xmlns="http://www.w3.org/1999/xhtml">' +
                 u'<head><meta http-equiv="Content-Type" content="text/html;charset=utf-8" /></head>'+
                 '<body>{}</body>' +
                 u'</html>').format(content)

        with open(self.preview_file_path, 'wb') as f:
            f.write(xhtml.encode('utf-8'))

    def start(self):
        import webbrowser
        import urllib
        self.webbrowser, self.urllib = webbrowser, urllib

        if self.args.inotify != 'false':
            try:
                import inotify.adapters
                import inotify.constants
                from inotify.calls import InotifyError
            except ImportError:
                if self.args.inotify == 'auto':
                    inotify = None
                    self.disp(u'inotify module not found, deactivating feature. You can install it with {install}'.format(install=INOTIFY_INSTALL))
                else:
                    self.disp(u"inotify not found, can't activate the feature! Please install it with {install}".format(install=INOTIFY_INSTALL), error=True)
                    self.host.quit(1)
            else:
                # we deactivate logging in inotify, which is quite annoying
                try:
                    inotify.adapters._LOGGER.setLevel(40)
                except AttributeError:
                    self.disp(u"Logger doesn't exists, inotify may have chanded", error=True)
        else:
            inotify=None

        sat_conf = config.parseMainConf()
        SYNTAX_EXT.update(config.getConfig(sat_conf, 'jp', CONF_SYNTAX_EXT, {}))

        try:
            self.open_cb_cmd = config.getConfig(sat_conf, 'jp', "blog_preview_open_cmd", Exception)
        except (NoOptionError, NoSectionError):
            self.open_cb_cmd = None
            open_cb = self.showPreview
        else:
            open_cb = self.openPreviewExt

        self.update_cb_cmd = config.getConfig(sat_conf, 'jp', "blog_preview_update_cmd", self.open_cb_cmd)
        if self.update_cb_cmd is None:
            update_cb = self.showPreview
        else:
            update_cb = self.updatePreviewExt

        # which file do we need to edit?
        if self.args.file == 'current':
            self.content_file_path = self.getCurrentFile(sat_conf)
        else:
            self.content_file_path = os.path.abspath(self.args.file)

        self.syntax = self.guessSyntaxFromPath(sat_conf, self.content_file_path)


        # at this point the syntax is converted, we can display the preview
        preview_file = tempfile.NamedTemporaryFile(suffix='.xhtml', delete=False)
        self.preview_file_path = preview_file.name
        preview_file.close()
        self.updateContent()

        if inotify is None:
            # XXX: we don't delete file automatically because browser need it (and webbrowser.open can return before it is read)
            self.disp(u'temporary file created at {}\nthis file will NOT BE DELETED AUTOMATICALLY, please delete it yourself when you have finished'.format(self.preview_file_path))
            open_cb()
        else:
            open_cb()
            i = inotify.adapters.Inotify(block_duration_s=60) # no need for 1 s duraction, inotify drive actions here

            def add_watch():
                i.add_watch(self.content_file_path, mask=inotify.constants.IN_CLOSE_WRITE |
                                                         inotify.constants.IN_DELETE_SELF)
            add_watch()

            try:
                for event in i.event_gen():
                    if event is not None:
                        self.disp(u"Content updated", 1)
                        if "IN_DELETE_SELF" in event[1]:
                            self.disp(u"IN_DELETE_SELF event catched, changing the watch", 2)
                            add_watch()
                        self.updateContent()
                        update_cb()
            except InotifyError:
                self.disp(u"Can't catch inotify events, as the file been deleted?", error=True)
            finally:
                os.unlink(self.preview_file_path)
                try:
                    i.remove_watch(self.content_file_path)
                except InotifyError:
                    pass


class Import(base.CommandAnswering):
    def __init__(self, host):
        super(Import, self).__init__(host, 'import', use_progress=True, help=_(u'import an external blog'))
        self.need_loop=True

    def add_parser_options(self):
        self.parser.add_argument("importer", type=base.unicode_decoder, nargs='?', help=_(u"importer name, nothing to display importers list"))
        self.parser.add_argument('--host', type=base.unicode_decoder, help=_(u"original blog host"))
        self.parser.add_argument('--no-images-upload', action='store_true', help=_(u"do *NOT* upload images (default: do upload images)"))
        self.parser.add_argument('--upload-ignore-host', help=_(u"do not upload images from this host (default: upload all images)"))
        self.parser.add_argument("--ignore-tls-errors", action="store_true", help=_("ignore invalide TLS certificate for uploads"))
        self.parser.add_argument('-o', '--option', action='append', nargs=2, default=[], metavar=(u'NAME', u'VALUE'),
            help=_(u"importer specific options (see importer description)"))
        self.parser.add_argument('--service', type=base.unicode_decoder, default=u'', metavar=u'PUBSUB_SERVICE',
            help=_(u"PubSub service where the items must be uploaded (default: server)"))
        self.parser.add_argument("location", type=base.unicode_decoder, nargs='?',
            help=_(u"importer data location (see importer description), nothing to show importer description"))

    def onProgressStarted(self, metadata):
        self.disp(_(u'Blog upload started'),2)

    def onProgressFinished(self, metadata):
        self.disp(_(u'Blog uploaded successfully'),2)
        redirections = {k[len(URL_REDIRECT_PREFIX):]:v for k,v in metadata.iteritems()
            if k.startswith(URL_REDIRECT_PREFIX)}
        if redirections:
            conf = u'\n'.join([
                u'url_redirections_profile = {}'.format(self.profile),
                u"url_redirections_dict = {}".format(
                # we need to add ' ' before each new line and to double each '%' for ConfigParser
                u'\n '.join(json.dumps(redirections, indent=1, separators=(',',': ')).replace(u'%', u'%%').split(u'\n'))),
                ])
            self.disp(_(u'\nTo redirect old URLs to new ones, put the following lines in your sat.conf file, in [libervia] section:\n\n{conf}'.format(conf=conf)))

    def onProgressError(self, error_msg):
        self.disp(_(u'Error while uploading blog: {}').format(error_msg),error=True)

    def error(self, failure):
        self.disp(_("Error while trying to upload a blog: {reason}").format(reason=failure), error=True)
        self.host.quit(1)

    def start(self):
        if self.args.location is None:
            for name in ('option', 'service', 'no_images_upload'):
                if getattr(self.args, name):
                    self.parser.error(_(u"{name} argument can't be used without location argument").format(name=name))
            if self.args.importer is None:
                print u'\n'.join([u'{}: {}'.format(name, desc) for name, desc in self.host.bridge.blogImportList()])
            else:
                try:
                    short_desc, long_desc = self.host.bridge.blogImportDesc(self.args.importer)
                except Exception as e:
                    msg = [l for l in unicode(e).split('\n') if l][-1] # we only keep the last line
                    print msg
                    self.host.quit(1)
                else:
                    print u"{name}: {short_desc}\n\n{long_desc}".format(name=self.args.importer, short_desc=short_desc, long_desc=long_desc)
            self.host.quit()
        else:
            # we have a location, an import is requested
            options = {key: value for key, value in self.args.option}
            if self.args.host:
                options['host'] = self.args.host
            if self.args.ignore_tls_errors:
                options['ignore_tls_errors'] = C.BOOL_TRUE
            if self.args.no_images_upload:
                options['upload_images'] = C.BOOL_FALSE
                if self.args.upload_ignore_host:
                    self.parser.error(u"upload-ignore-host option can't be used when no-images-upload is set")
            elif self.args.upload_ignore_host:
                options['upload_ignore_host'] = self.args.upload_ignore_host
            def gotId(id_):
                self.progress_id = id_
            self.host.bridge.blogImport(self.args.importer, self.args.location, options, self.args.service, self.profile,
                callback=gotId, errback=self.error)


class Blog(base.CommandBase):
    subcommands = (Edit, Preview, Import)

    def __init__(self, host):
        super(Blog, self).__init__(host, 'blog', use_profile=False, help=_('blog/microblog management'))