first commit

mod/book/tool/importhtml/classes/privacy/provider.php

@@ -0,0 +1,46 @@
+<?php
+// This file is part of Moodle - http://moodle.org/
+//
+// Moodle 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 3 of the License, or
+// (at your option) any later version.
+//
+// Moodle 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 Moodle.  If not, see <http://www.gnu.org/licenses/>.
+
+/**
+ * Privacy provider implementation for booktool_importhtml.
+ *
+ * @package    booktool_importhtml
+ * @copyright  2018 Mihail Geshoski <mihail@moodle.com>
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+
+namespace booktool_importhtml\privacy;
+
+defined('MOODLE_INTERNAL') || die();
+
+/**
+ * Privacy provider implementation for booktool_importhtml.
+ *
+ * @copyright  2018 Mihail Geshoski <mihail@moodle.com>
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+class provider implements \core_privacy\local\metadata\null_provider {
+
+    /**
+     * Get the language string identifier with the component's language
+     * file to explain why this plugin stores no data.
+     *
+     * @return  string
+     */
+    public static function get_reason(): string {
+        return 'privacy:metadata';
+    }
+}

mod/book/tool/importhtml/db/access.php

@@ -0,0 +1,37 @@
+<?php
+// This file is part of Moodle - http://moodle.org/
+//
+// Moodle 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 3 of the License, or
+// (at your option) any later version.
+//
+// Moodle 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 Moodle.  If not, see <http://www.gnu.org/licenses/>.
+
+/**
+ * Book import capability definition
+ *
+ * @package    booktool_importhtml
+ * @copyright  2011 Petr Skoda {@link http://skodak.org}
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+
+defined('MOODLE_INTERNAL') || die;
+
+$capabilities = array(
+    'booktool/importhtml:import' => array(
+        'riskbitmask' => RISK_XSS,
+        'captype' => 'write',
+        'contextlevel' => CONTEXT_MODULE,
+        'archetypes' => array(
+            'editingteacher' => CAP_ALLOW,
+            'manager' => CAP_ALLOW,
+        )
+    ),
+);

mod/book/tool/importhtml/import_form.php

@@ -0,0 +1,88 @@
+<?php
+// This file is part of Moodle - http://moodle.org/
+//
+// Moodle 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 3 of the License, or
+// (at your option) any later version.
+//
+// Moodle 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 Moodle.  If not, see <http://www.gnu.org/licenses/>.
+
+/**
+ * Book import form
+ *
+ * @package    booktool_importhtml
+ * @copyright  2004-2011 Petr Skoda {@link http://skodak.org}
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+
+defined('MOODLE_INTERNAL') || die;
+
+require_once($CFG->libdir.'/formslib.php');
+
+class booktool_importhtml_form extends moodleform {
+
+    function definition() {
+        $mform = $this->_form;
+        $data  = $this->_customdata;
+
+        $mform->addElement('header', 'general', get_string('import', 'booktool_importhtml'));
+
+        $options = array(
+                // '0'=>get_string('typeonefile', 'booktool_importhtml'),
+                '1'=>get_string('typezipdirs', 'booktool_importhtml'),
+                '2'=>get_string('typezipfiles', 'booktool_importhtml'),
+        );
+        $mform->addElement('select', 'type', get_string('type', 'booktool_importhtml'), $options);
+        $mform->setDefault('type', 2);
+
+        $mform->addElement('filepicker', 'importfile', get_string('ziparchive', 'booktool_importhtml'));
+        $mform->addHelpButton('importfile', 'ziparchive', 'booktool_importhtml');
+        $mform->addRule('importfile', null, 'required');
+
+        $mform->addElement('hidden', 'id');
+        $mform->setType('id', PARAM_INT);
+
+        $mform->addElement('hidden', 'chapterid');
+        $mform->setType('chapterid', PARAM_INT);
+
+        $this->add_action_buttons(true, get_string('doimport', 'booktool_importhtml'));
+
+        $this->set_data($data);
+    }
+
+    function validation($data, $files) {
+        global $USER;
+
+        if ($errors = parent::validation($data, $files)) {
+            return $errors;
+        }
+
+        $usercontext = context_user::instance($USER->id);
+        $fs = get_file_storage();
+
+        if (!$files = $fs->get_area_files($usercontext->id, 'user', 'draft', $data['importfile'], 'id', false)) {
+            $errors['importfile'] = get_string('required');
+            return $errors;
+        } else {
+            $file = reset($files);
+            if ($file->get_mimetype() != 'application/zip') {
+                $errors['importfile'] = get_string('invalidfiletype', 'error', $file->get_filename());
+                // better delete current file, it is not usable anyway
+                $fs->delete_area_files($usercontext->id, 'user', 'draft', $data['importfile']);
+            } else {
+                if (!$chpterfiles = toolbook_importhtml_get_chapter_files($file, $data['type'])) {
+                    $errors['importfile'] = get_string('errornochapters', 'booktool_importhtml');
+                }
+            }
+        }
+
+        return $errors;
+    }
+}

mod/book/tool/importhtml/index.php

@@ -0,0 +1,93 @@
+<?php
+// This file is part of Moodle - http://moodle.org/
+//
+// Moodle 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 3 of the License, or
+// (at your option) any later version.
+//
+// Moodle 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 Moodle.  If not, see <http://www.gnu.org/licenses/>.
+
+/**
+ * Book import
+ *
+ * @package    booktool_importhtml
+ * @copyright  2004-2011 Petr Skoda {@link http://skodak.org}
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+
+require(__DIR__.'/../../../../config.php');
+require_once(__DIR__.'/locallib.php');
+require_once(__DIR__.'/import_form.php');
+
+$id        = required_param('id', PARAM_INT);           // Course Module ID
+$chapterid = optional_param('chapterid', 0, PARAM_INT); // Chapter ID
+
+$cm = get_coursemodule_from_id('book', $id, 0, false, MUST_EXIST);
+$course = $DB->get_record('course', array('id'=>$cm->course), '*', MUST_EXIST);
+$book = $DB->get_record('book', array('id'=>$cm->instance), '*', MUST_EXIST);
+
+require_login($course, false, $cm);
+
+$context = context_module::instance($cm->id);
+require_capability('booktool/importhtml:import', $context);
+
+$PAGE->set_url('/mod/book/tool/importhtml/index.php', array('id' => $id));
+
+if ($chapterid) {
+    if (!$chapter = $DB->get_record('book_chapters', array('id'=>$chapterid, 'bookid'=>$book->id))) {
+        $chapterid = 0;
+    }
+} else {
+    $chapter = false;
+}
+
+$PAGE->set_title($book->name);
+$PAGE->set_heading($course->fullname);
+$PAGE->activityheader->set_attrs([
+    'hidecompletion' => true,
+    'description' => ''
+]);
+// Prepare the page header.
+$strbook = get_string('modulename', 'mod_book');
+$strbooks = get_string('modulenameplural', 'mod_book');
+
+$mform = new booktool_importhtml_form(null, array('id'=>$id, 'chapterid'=>$chapterid));
+
+// If data submitted, then process and store.
+if ($mform->is_cancelled()) {
+    if (empty($chapter->id)) {
+        redirect($CFG->wwwroot."/mod/book/view.php?id=$cm->id");
+    } else {
+        redirect($CFG->wwwroot."/mod/book/view.php?id=$cm->id&chapterid=$chapter->id");
+    }
+
+} else if ($data = $mform->get_data()) {
+    echo $OUTPUT->header();
+    echo $OUTPUT->heading(get_string('importingchapters', 'booktool_importhtml'), 3);
+
+    // this is a bloody hack - children do not try this at home!
+    $fs = get_file_storage();
+    $draftid = file_get_submitted_draft_itemid('importfile');
+    if (!$files = $fs->get_area_files(context_user::instance($USER->id)->id, 'user', 'draft', $draftid, 'id DESC', false)) {
+        redirect($PAGE->url);
+    }
+    $file = reset($files);
+    toolbook_importhtml_import_chapters($file, $data->type, $book, $context);
+
+    echo $OUTPUT->continue_button(new moodle_url('/mod/book/view.php', array('id'=>$id)));
+    echo $OUTPUT->footer();
+    die;
+}
+
+echo $OUTPUT->header();
+
+$mform->display();
+
+echo $OUTPUT->footer();

mod/book/tool/importhtml/lang/en/booktool_importhtml.php

@@ -0,0 +1,41 @@
+<?php
+// This file is part of Moodle - http://moodle.org/
+//
+// Moodle 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 3 of the License, or
+// (at your option) any later version.
+//
+// Moodle 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 Moodle.  If not, see <http://www.gnu.org/licenses/>.
+
+/**
+ * Book import language strings
+ *
+ * @package    booktool_importhtml
+ * @copyright  2011 Petr Skoda {@link http://skodak.org}
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+
+defined('MOODLE_INTERNAL') || die;
+
+$string['doimport'] = 'Import';
+$string['errornochapters'] = 'Cannot find chapters in selected file';
+$string['import'] = 'Import chapter';
+$string['importhtml:import'] = 'Import chapters';
+$string['importing'] = 'Importing';
+$string['importingchapters'] = 'Importing chapters into book';
+$string['pluginname'] = 'Book chapter import';
+$string['privacy:metadata'] = 'The Book chapter import plugin does not store any personal data.';
+$string['relinking'] = 'Relinking';
+$string['type'] = 'Type';
+$string['typeonefile'] = 'One HTML file with headings as chapters';
+$string['typezipfiles'] = 'Each HTML file represents one chapter';
+$string['typezipdirs'] = 'Each folder represents one chapter';
+$string['ziparchive'] = 'Zip file';
+$string['ziparchive_help'] = 'Select a zip file containing HTML files and optional multimedia files and folders. To upload subchapters, add "_sub" to the end of HTML file or folder names.';

mod/book/tool/importhtml/lib.php

@@ -0,0 +1,38 @@
+<?php
+// This file is part of Moodle - http://moodle.org/
+//
+// Moodle 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 3 of the License, or
+// (at your option) any later version.
+//
+// Moodle 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 Moodle.  If not, see <http://www.gnu.org/licenses/>.
+
+/**
+ * HTML import lib
+ *
+ * @package    booktool_importhtml
+ * @copyright  2011 Petr Skoda {@link http://skodak.org}
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+
+defined('MOODLE_INTERNAL') || die;
+
+/**
+ * Adds module specific settings to the settings block
+ *
+ * @param settings_navigation $settings The settings navigation object
+ * @param navigation_node $node The node to add module settings to
+ */
+function booktool_importhtml_extend_settings_navigation(settings_navigation $settings, navigation_node $node) {
+    if (has_capability('booktool/importhtml:import', $settings->get_page()->cm->context)) {
+        $url = new moodle_url('/mod/book/tool/importhtml/index.php', array('id' => $settings->get_page()->cm->id));
+        $node->add(get_string('import', 'booktool_importhtml'), $url, navigation_node::TYPE_SETTING, null, 'importchapter', null);
+    }
+}

mod/book/tool/importhtml/locallib.php

@@ -0,0 +1,346 @@
+<?php
+// This file is part of Moodle - http://moodle.org/
+//
+// Moodle 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 3 of the License, or
+// (at your option) any later version.
+//
+// Moodle 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 Moodle.  If not, see <http://www.gnu.org/licenses/>.
+
+/**
+ * HTML import lib
+ *
+ * @package    booktool_importhtml
+ * @copyright  2011 Petr Skoda {@link http://skodak.org}
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+
+defined('MOODLE_INTERNAL') || die;
+
+require_once(__DIR__.'/lib.php');
+require_once($CFG->dirroot.'/mod/book/locallib.php');
+
+/**
+ * Import HTML pages packaged into one zip archive
+ *
+ * @param stored_file $package
+ * @param string $type type of the package ('typezipdirs' or 'typezipfiles')
+ * @param stdClass $book
+ * @param context_module $context
+ * @param bool $verbose
+ */
+function toolbook_importhtml_import_chapters($package, $type, $book, $context, $verbose = true) {
+    global $DB, $OUTPUT;
+
+    $fs = get_file_storage();
+    $chapterfiles = toolbook_importhtml_get_chapter_files($package, $type);
+    $packer = get_file_packer('application/zip');
+    $fs->delete_area_files($context->id, 'mod_book', 'importhtmltemp', 0);
+    $package->extract_to_storage($packer, $context->id, 'mod_book', 'importhtmltemp', 0, '/');
+    // $datafiles = $fs->get_area_files($context->id, 'mod_book', 'importhtmltemp', 0, 'id', false);
+    // echo "<pre>";p(var_export($datafiles, true));
+
+    $chapters = array();
+
+    if ($verbose) {
+        echo $OUTPUT->notification(get_string('importing', 'booktool_importhtml'), 'notifysuccess');
+    }
+    if ($type == 0) {
+        $chapterfile = reset($chapterfiles);
+        if ($file = $fs->get_file_by_hash(sha1("$context->id/mod_book/importhtmltemp/0/$chapterfile->pathname"))) {
+            $htmlcontent = toolbook_importhtml_fix_encoding($file->get_content());
+            $htmlchapters = toolbook_importhtml_parse_headings(toolbook_importhtml_parse_body($htmlcontent));
+            // TODO: process h1 as main chapter and h2 as subchapters
+        }
+    } else {
+        foreach ($chapterfiles as $chapterfile) {
+            if ($file = $fs->get_file_by_hash(sha1("/$context->id/mod_book/importhtmltemp/0/$chapterfile->pathname"))) {
+                $chapter = new stdClass();
+                $htmlcontent = toolbook_importhtml_fix_encoding($file->get_content());
+
+                $chapter->bookid        = $book->id;
+                $chapter->pagenum       = $DB->get_field_sql('SELECT MAX(pagenum) FROM {book_chapters} WHERE bookid = ?', array($book->id)) + 1;
+                $chapter->importsrc     = '/'.$chapterfile->pathname;
+                $chapter->content       = toolbook_importhtml_parse_styles($htmlcontent);
+                $chapter->content       .= toolbook_importhtml_parse_body($htmlcontent);
+                $chapter->title         = toolbook_importhtml_parse_title($htmlcontent, $chapterfile->pathname);
+                $chapter->contentformat = FORMAT_HTML;
+                $chapter->hidden        = 0;
+                $chapter->timecreated   = time();
+                $chapter->timemodified  = time();
+                if (preg_match('/_sub(\/|\.htm)/i', $chapter->importsrc)) { // If filename or directory ends with *_sub treat as subchapters
+                    $chapter->subchapter = 1;
+                } else {
+                    $chapter->subchapter = 0;
+                }
+
+                $chapter->id = $DB->insert_record('book_chapters', $chapter);
+                $chapter = $DB->get_record('book_chapters', array('id' => $chapter->id));
+                $chapters[$chapter->id] = $chapter;
+
+                \mod_book\event\chapter_created::create_from_chapter($book, $context, $chapter)->trigger();
+            }
+        }
+    }
+
+    if ($verbose) {
+        echo $OUTPUT->notification(get_string('relinking', 'booktool_importhtml'), 'notifysuccess');
+    }
+    $allchapters = $DB->get_records('book_chapters', array('bookid'=>$book->id), 'pagenum');
+    foreach ($chapters as $chapter) {
+        // find references to all files and copy them + relink them
+        $matches = null;
+        if (preg_match_all('/(src|codebase|name|href)\s*=\s*"([^"]+)"/i', $chapter->content, $matches)) {
+            $file_record = array('contextid'=>$context->id, 'component'=>'mod_book', 'filearea'=>'chapter', 'itemid'=>$chapter->id);
+            foreach ($matches[0] as $i => $match) {
+                $filepath = dirname($chapter->importsrc).'/'.$matches[2][$i];
+                $filepath = toolbook_importhtml_fix_path($filepath);
+
+                if (strtolower($matches[1][$i]) === 'href') {
+                    // skip linked html files, we will try chapter relinking later
+                    foreach ($allchapters as $target) {
+                        if ($target->importsrc === $filepath) {
+                            continue 2;
+                        }
+                    }
+                }
+
+                if ($file = $fs->get_file_by_hash(sha1("/$context->id/mod_book/importhtmltemp/0$filepath"))) {
+                    if (!$oldfile = $fs->get_file_by_hash(sha1("/$context->id/mod_book/chapter/$chapter->id$filepath"))) {
+                        $fs->create_file_from_storedfile($file_record, $file);
+                    }
+                    $chapter->content = str_replace($match, $matches[1][$i].'="@@PLUGINFILE@@'.$filepath.'"', $chapter->content);
+                }
+            }
+            $DB->set_field('book_chapters', 'content', $chapter->content, array('id'=>$chapter->id));
+        }
+    }
+    unset($chapters);
+
+    $allchapters = $DB->get_records('book_chapters', array('bookid'=>$book->id), 'pagenum');
+    foreach ($allchapters as $chapter) {
+        $newcontent = $chapter->content;
+        $matches = null;
+        if (preg_match_all('/(href)\s*=\s*"([^"]+)"/i', $chapter->content, $matches)) {
+            foreach ($matches[0] as $i => $match) {
+                if (strpos($matches[2][$i], ':') !== false or strpos($matches[2][$i], '@') !== false) {
+                    // it is either absolute or pluginfile link
+                    continue;
+                }
+                $chapterpath = dirname($chapter->importsrc).'/'.$matches[2][$i];
+                $chapterpath = toolbook_importhtml_fix_path($chapterpath);
+                foreach ($allchapters as $target) {
+                    if ($target->importsrc === $chapterpath) {
+                        $newcontent = str_replace($match, 'href="'.new moodle_url('/mod/book/view.php',
+                                array('id'=>$context->instanceid, 'chapterid'=>$target->id)).'"', $newcontent);
+                    }
+                }
+            }
+        }
+        if ($newcontent !== $chapter->content) {
+            $DB->set_field('book_chapters', 'content', $newcontent, array('id'=>$chapter->id));
+        }
+    }
+
+    $fs->delete_area_files($context->id, 'mod_book', 'importhtmltemp', 0);
+
+    // update the revision flag - this takes a long time, better to refetch the current value
+    $book = $DB->get_record('book', array('id'=>$book->id));
+    $DB->set_field('book', 'revision', $book->revision+1, array('id'=>$book->id));
+}
+
+/**
+ * Parse the headings of the imported package of type 'typeonefile'
+ * (currently unsupported)
+ *
+ * @param string $html html content to parse
+ * @todo implement this once the type 'typeonefile' is enabled
+ */
+function toolbook_importhtml_parse_headings($html) {
+}
+
+/**
+ * Parse the links to external css sheets of the imported html content
+ *
+ * @param string $html html content to parse
+ * @return string all the links to external css sheets
+ */
+function toolbook_importhtml_parse_styles($html) {
+    $styles = '';
+    if (preg_match('/<head[^>]*>(.+)<\/head>/is', $html, $matches)) {
+        $head = $matches[1];
+        if (preg_match_all('/<link[^>]+rel="stylesheet"[^>]*>/i', $head, $matches)) { // Extract links to css.
+            for ($i=0; $i<count($matches[0]); $i++) {
+                $styles .= $matches[0][$i]."\n";
+            }
+        }
+    }
+    return $styles;
+}
+
+/**
+ * Normalize paths to be absolute
+ *
+ * @param string $path original path with MS/relative separators
+ * @return string the normalized and cleaned absolute path
+ */
+function toolbook_importhtml_fix_path($path) {
+    $path = str_replace('\\', '/', $path); // anti MS hack
+    $path = '/'.ltrim($path, './'); // dirname() produces . for top level files + our paths start with /
+
+    $cnt = substr_count($path, '..');
+    for ($i=0; $i<$cnt; $i++) {
+        $path = preg_replace('|[^/]+/\.\./|', '', $path, 1);
+    }
+
+    $path = clean_param($path, PARAM_PATH);
+    return $path;
+}
+
+/**
+ * Convert some html content to utf8, getting original encoding from html headers
+ *
+ * @param string $html html content to convert
+ * @return string html content converted to utf8
+ */
+function toolbook_importhtml_fix_encoding($html) {
+    if (preg_match('/<head[^>]*>(.+)<\/head>/is', $html, $matches)) {
+        $head = $matches[1];
+        if (preg_match('/charset=([^"]+)/is', $head, $matches)) {
+            $enc = $matches[1];
+            return core_text::convert($html, $enc, 'utf-8');
+        }
+    }
+    return iconv('UTF-8', 'UTF-8//IGNORE', $html);
+}
+
+/**
+ * Extract the body from any html contents
+ *
+ * @param string $html the html to parse
+ * @return string the contents of the body
+ */
+function toolbook_importhtml_parse_body($html) {
+    $matches = null;
+    if (preg_match('/<body[^>]*>(.+)<\/body>/is', $html, $matches)) {
+        return $matches[1];
+    } else {
+        return '';
+    }
+}
+
+/**
+ * Extract the title of any html content, getting it from the title tag
+ *
+ * @param string $html the html to parse
+ * @param string $default default title to apply if no title is found
+ * @return string the resulting title
+ */
+function toolbook_importhtml_parse_title($html, $default) {
+    $matches = null;
+    if (preg_match('/<title>([^<]+)<\/title>/i', $html, $matches)) {
+        return $matches[1];
+    } else {
+        return $default;
+    }
+}
+
+/**
+ * Returns all the html files (chapters) from a file package
+ *
+ * @param stored_file $package file to be processed
+ * @param string $type type of the package ('typezipdirs' or 'typezipfiles')
+ *
+ * @return array the html files found in the package
+ */
+function toolbook_importhtml_get_chapter_files($package, $type) {
+    $packer = get_file_packer('application/zip');
+    $files = $package->list_files($packer);
+    $tophtmlfiles = array();
+    $subhtmlfiles = array();
+    $topdirs = array();
+
+    foreach ($files as $file) {
+        if (empty($file->pathname)) {
+            continue;
+        }
+        if (substr($file->pathname, -1) === '/') {
+            if (substr_count($file->pathname, '/') !== 1) {
+                // skip subdirs
+                continue;
+            }
+            if (!isset($topdirs[$file->pathname])) {
+                $topdirs[$file->pathname] = array();
+            }
+
+        } else {
+            $mime = mimeinfo('icon', $file->pathname);
+            if ($mime !== 'markup') {
+                continue;
+            }
+            $level = substr_count($file->pathname, '/');
+            if ($level === 0) {
+                $tophtmlfiles[$file->pathname] = $file;
+            } else if ($level === 1) {
+                $subhtmlfiles[$file->pathname] = $file;
+                $dir = preg_replace('|/.*$|', '', $file->pathname);
+                $topdirs[$dir][$file->pathname] = $file;
+            } else {
+                // lower levels are not interesting
+                continue;
+            }
+        }
+    }
+
+    core_collator::ksort($tophtmlfiles, core_collator::SORT_NATURAL);
+    core_collator::ksort($subhtmlfiles, core_collator::SORT_NATURAL);
+    core_collator::ksort($topdirs, core_collator::SORT_NATURAL);
+
+    $chapterfiles = array();
+
+    if ($type == 2) {
+        $chapterfiles = $tophtmlfiles;
+
+    } else if ($type == 1) {
+        foreach ($topdirs as $dir => $htmlfiles) {
+            if (empty($htmlfiles)) {
+                continue;
+            }
+            core_collator::ksort($htmlfiles, core_collator::SORT_NATURAL);
+            if (isset($htmlfiles[$dir.'/index.html'])) {
+                $htmlfile = $htmlfiles[$dir.'/index.html'];
+            } else if (isset($htmlfiles[$dir.'/index.htm'])) {
+                $htmlfile = $htmlfiles[$dir.'/index.htm'];
+            } else if (isset($htmlfiles[$dir.'/Default.htm'])) {
+                $htmlfile = $htmlfiles[$dir.'/Default.htm'];
+            } else {
+                $htmlfile = reset($htmlfiles);
+            }
+            $chapterfiles[$htmlfile->pathname] = $htmlfile;
+        }
+    } else if ($type == 0) {
+        if ($tophtmlfiles) {
+            if (isset($tophtmlfiles['index.html'])) {
+                $htmlfile = $tophtmlfiles['index.html'];
+            } else if (isset($tophtmlfiles['index.htm'])) {
+                $htmlfile = $tophtmlfiles['index.htm'];
+            } else if (isset($tophtmlfiles['Default.htm'])) {
+                $htmlfile = $tophtmlfiles['Default.htm'];
+            } else {
+                $htmlfile = reset($tophtmlfiles);
+            }
+        } else {
+            $htmlfile = reset($subhtmlfiles);
+        }
+        $chapterfiles[$htmlfile->pathname] = $htmlfile;
+    }
+
+    return $chapterfiles;
+}

mod/book/tool/importhtml/tests/locallib_test.php

@@ -0,0 +1,78 @@
+<?php
+// This file is part of Moodle - http://moodle.org/
+//
+// Moodle 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 3 of the License, or
+// (at your option) any later version.
+//
+// Moodle 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 Moodle.  If not, see <http://www.gnu.org/licenses/>.
+
+/**
+ * booktool_importhtml tests.
+ *
+ * @package    booktool_importhtml
+ * @category   phpunit
+ * @copyright  2013 Frédéric Massart
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+namespace booktool_importhtml;
+
+defined('MOODLE_INTERNAL') || die();
+global $CFG;
+
+require_once($CFG->dirroot.'/mod/book/tool/importhtml/locallib.php');
+
+/**
+ * booktool_importhtml tests class.
+ *
+ * @package    booktool_importhtml
+ * @category   phpunit
+ * @copyright  2013 Frédéric Massart
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+class locallib_test extends \advanced_testcase {
+
+    public function setUp(): void {
+        $this->resetAfterTest();
+    }
+
+    public function test_import_chapters_events(): void {
+        $course = $this->getDataGenerator()->create_course();
+        $book = $this->getDataGenerator()->create_module('book', array('course' => $course->id));
+        $context = \context_module::instance($book->cmid);
+
+        $record = new \stdClass();
+        $record->contextid = $context->id;
+        $record->component = 'phpunit';
+        $record->filearea = 'test';
+        $record->itemid = 0;
+        $record->filepath = '/';
+        $record->filename = 'chapters.zip';
+
+        $fs = get_file_storage();
+        $file = $fs->create_file_from_pathname($record, __DIR__ . '/fixtures/chapters.zip');
+
+        // Importing the chapters.
+        $sink = $this->redirectEvents();
+        toolbook_importhtml_import_chapters($file, 2, $book, $context, false);
+        $events = $sink->get_events();
+
+        // Checking the results.
+        $this->assertCount(5, $events);
+        foreach ($events as $event) {
+            $this->assertInstanceOf('\mod_book\event\chapter_created', $event);
+            $this->assertEquals($context, $event->get_context());
+            $chapter = $event->get_record_snapshot('book_chapters', $event->objectid);
+            $this->assertNotEmpty($chapter);
+            $this->assertEventContextNotUsed($event);
+        }
+    }
+
+}

mod/book/tool/importhtml/version.php

@@ -0,0 +1,29 @@
+<?php
+// This file is part of Moodle - http://moodle.org/
+//
+// Moodle 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 3 of the License, or
+// (at your option) any later version.
+//
+// Moodle 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 Moodle.  If not, see <http://www.gnu.org/licenses/>.
+
+/**
+ * Book import plugin version info
+ *
+ * @package    booktool_importhtml
+ * @copyright  2011 Petr Skoda {@link http://skodak.org}
+ * @license    http://www.gnu.org/copyleft/gpl.html GNU GPL v3 or later
+ */
+
+defined('MOODLE_INTERNAL') || die;
+
+$plugin->component = 'booktool_importhtml'; // Full name of the plugin (used for diagnostics)
+$plugin->version   = 2024042200; // The current plugin version (Date: YYYYMMDDXX).
+$plugin->requires  = 2024041600; // Requires this Moodle version.