| 1 |
<?php
|
| 2 |
/**
|
| 3 |
* @file
|
| 4 |
* Additional per-module functionality used by import_html.
|
| 5 |
*
|
| 6 |
* Each available module MAY impliment a hook_import_html() function to add or manipulate
|
| 7 |
* its own data as the node gets saved.
|
| 8 |
*
|
| 9 |
* @ingroup import_html Import HTML
|
| 10 |
* @author Dan Morrison http://coders.co.nz/
|
| 11 |
* @version $Id: import_html_modules.inc,v 1.11.2.2 2009/03/14 03:18:51 dman Exp $
|
| 12 |
*/
|
| 13 |
|
| 14 |
/**
|
| 15 |
* hook_import_html - prototype function hook
|
| 16 |
*
|
| 17 |
* invoke this when creating a node out of nothing
|
| 18 |
*
|
| 19 |
* @param $profile array of settings for the import_html process
|
| 20 |
* @param $node node object with lots of extra data added to it, only some of which we want to save.
|
| 21 |
* @param $datadoc XML doc of the 'simple' node source that we may investigate for extra info if the cooked node object is not good enough
|
| 22 |
*/
|
| 23 |
function hook_import_html($profile, &$node, $datadoc = NULL) {
|
| 24 |
|
| 25 |
}
|
| 26 |
/**
|
| 27 |
* hook_import_html_after_save - prototype function hook
|
| 28 |
*
|
| 29 |
* Some functions (alias) can only operate when the nid is already known.
|
| 30 |
* Invoke this when the node has been created
|
| 31 |
*
|
| 32 |
* @param $profile array of settings for the import_html process
|
| 33 |
* @param $node node object with lots of extra data added to it, only some of which we want to save.
|
| 34 |
* @param $datadoc XML doc of the 'simple' node source that we may investigate for extra info if the cooked node object is not good enough
|
| 35 |
*/
|
| 36 |
function hook_import_html_after_save($profile, &$node, $datadoc = NULL) {
|
| 37 |
|
| 38 |
}
|
| 39 |
|
| 40 |
|
| 41 |
|
| 42 |
/**
|
| 43 |
* Implementation of hook_import_html
|
| 44 |
*
|
| 45 |
* Add appropriate normal node attributes - title, body and teaser
|
| 46 |
* - to the node
|
| 47 |
*/
|
| 48 |
function node_import_html($profile, &$node, $datadoc = NULL) {
|
| 49 |
// Title
|
| 50 |
// Note, internally the title should always be plaintext, encoded characters and no newlines
|
| 51 |
// A meta 'title' may have been set,
|
| 52 |
// but I'll use the real visible one by preference.
|
| 53 |
$h1 = xml_get_element_content($datadoc, 'h1');
|
| 54 |
if ($h1) {
|
| 55 |
$node->title = $h1;
|
| 56 |
}
|
| 57 |
else {
|
| 58 |
$title_elem = xml_get_element_content($datadoc, 'title');
|
| 59 |
if ($title_elem) {
|
| 60 |
$node->title = $title_elem;
|
| 61 |
}
|
| 62 |
}
|
| 63 |
|
| 64 |
//
|
| 65 |
// BODY is the thing with id=content or whatever is set in the prefs
|
| 66 |
//
|
| 67 |
debug_pre(array('XML DOM being scanned by node_import_html() for XPATH data extraction to find the body content' => xml_tostring($datadoc)), 3);
|
| 68 |
$content_element = xml_getelementbyid($datadoc, 'content');
|
| 69 |
|
| 70 |
if (!$content_element) {
|
| 71 |
drupal_set_message("Failed to find a body, anything with id='content' in this page");
|
| 72 |
// TODO some cck may be valid without a 'content' body?
|
| 73 |
}
|
| 74 |
|
| 75 |
if ($content_element) {
|
| 76 |
$node->body = trim(xml_tostring($content_element, TRUE));
|
| 77 |
|
| 78 |
// The XSL may give us further hints about how it found this body data
|
| 79 |
$precision = xml_getattribute($content_element, 'precision');
|
| 80 |
if ($precision != NULL) {
|
| 81 |
$node->import_html_precision = $precision;
|
| 82 |
$explanation = xml_getattribute($content_element, 'title');
|
| 83 |
debug(t("I'm %precision % confident that we found the right body element. !explanation", array('%precision' => 100*$precision, '!explanation' => $explanation)));
|
| 84 |
}
|
| 85 |
}
|
| 86 |
|
| 87 |
// It's possible that our input was content-encoded (if it came from RSS or the old import-xml node)
|
| 88 |
// If so, the entities should be unwrapped.
|
| 89 |
// Other (HTML) imports should not require this
|
| 90 |
// @TODO when was $node->content_encoded being flagged?
|
| 91 |
$node->body = ($node->content_encoded) ? html_entity_decode($node->body):$node->body;
|
| 92 |
|
| 93 |
//
|
| 94 |
// Teaser
|
| 95 |
//
|
| 96 |
$teaser = xml_textcontent( xml_getelementbyid($datadoc, 'teaser') );
|
| 97 |
$description = xml_textcontent( xml_getelementbyid($datadoc, 'description') );
|
| 98 |
if (!$teaser) $teaser = $description ;
|
| 99 |
// No custom teaser found, but
|
| 100 |
// Imported HTML is highly likely to contain formatting that
|
| 101 |
// will ruin teasers. Do our own cut-down version that assumes
|
| 102 |
// FILTERED HTML
|
| 103 |
$teaser = filter_filter('process', 0, 1, node_teaser($node->body));
|
| 104 |
if ($teaser) $node->teaser = $teaser;
|
| 105 |
}
|
| 106 |
|
| 107 |
|
| 108 |
/**
|
| 109 |
* Implementation of hook_import_html
|
| 110 |
*
|
| 111 |
* Add appropriate user identification to the new node.
|
| 112 |
*/
|
| 113 |
function user_import_html($profile, &$node, $datadoc = NULL) {
|
| 114 |
global $user;
|
| 115 |
$node->uid = $user->uid;
|
| 116 |
$node->name = $profile['import_user'];
|
| 117 |
if ($import_user = user_load(array('name' => $node->name))) {
|
| 118 |
$node->uid = $import_user->uid;
|
| 119 |
}
|
| 120 |
}
|
| 121 |
|
| 122 |
|
| 123 |
/**
|
| 124 |
* Set alias and navigation for the given node
|
| 125 |
* Helper for import_html_import_files()
|
| 126 |
*/
|
| 127 |
function menu_import_html_after_save($profile, &$node, $datadoc = NULL) {
|
| 128 |
$normal_path = "node/{$node->nid}";
|
| 129 |
|
| 130 |
// Create Menu item for this new item
|
| 131 |
if ($profile['create_menus']) {
|
| 132 |
$node->label = $node->title;
|
| 133 |
|
| 134 |
// Need to beware of stupid long titles - they can't fit in breadcrumbs and menus
|
| 135 |
if (strlen($node->title) > IMPORT_HTML_MAX_LABEL_LENGTH) {
|
| 136 |
$path_bits = split('/', $node->path);
|
| 137 |
$node->label=array_pop($path_bits);
|
| 138 |
$node->label=preg_replace('/\?.*$/', '?', $node->label); // messiness from mirrored URLs with args in
|
| 139 |
$node->label=str_replace('_', ' ', $node->label);
|
| 140 |
// @TODO maybe adjust this title-munging algoritm to make better guesses
|
| 141 |
drupal_set_message(t("When creating menu item, the title '%node_title is too long to fit into a menu label (MAX=%MAX chars), truncating it to '%node_label' . This is not terrible, just unaesthetic. You may want to fix this later.", array('%node_title' => $node->title, '%MAX' => IMPORT_HTML_MAX_LABEL_LENGTH, '%node_label' => $node->label)));
|
| 142 |
}
|
| 143 |
|
| 144 |
// Check if a placeholder for the alias exists.
|
| 145 |
// This can happen if the child items were created earlier and required a placeholder
|
| 146 |
// Try both canonic and aliased paths
|
| 147 |
if( ! $menu = menu_get_item_by_path($normal_path) ) {
|
| 148 |
$menu = menu_get_item_by_path($node->path);
|
| 149 |
}
|
| 150 |
if ($menu) {
|
| 151 |
drupal_set_message(t("Discovered that a placeholder menu item for %path already exists. Filling that one out with the full details now", array('%path' => $node->path)));
|
| 152 |
}
|
| 153 |
else {
|
| 154 |
// New menu item
|
| 155 |
debug("Looking for parent menu of $node->path", 2);
|
| 156 |
$pid = import_html_create_menu_path(dirname($node->path), '');
|
| 157 |
if (!$pid) { $pid = $profile['menu_parent_id']; }
|
| 158 |
$menu = array(
|
| 159 |
'pid' => $pid,
|
| 160 |
'type' => MENU_CREATED_BY_ADMIN | MENU_MODIFIABLE_BY_ADMIN | MENU_VISIBLE_IN_TREE | MENU_VISIBLE_IN_BREADCRUMB,
|
| 161 |
);
|
| 162 |
}
|
| 163 |
$menu['title'] = $node->label;
|
| 164 |
$menu['path'] = $normal_path;
|
| 165 |
$node->menu = $menu;
|
| 166 |
menu_edit_item_save($node->menu);
|
| 167 |
}
|
| 168 |
}
|
| 169 |
|
| 170 |
/**
|
| 171 |
* Given a traditional URL path, return the parent menu item
|
| 172 |
*
|
| 173 |
* A helper function to menu_import_html_after_save()
|
| 174 |
*
|
| 175 |
* This is a recursive function that slices the given
|
| 176 |
* (alias) path up until it finds a menu item it recognises.
|
| 177 |
* Once found, it comes back down again building a path
|
| 178 |
* until we end up with the new parent menu id.
|
| 179 |
*
|
| 180 |
* @param $path the path (aliased URL) we intend to construct
|
| 181 |
* @param $title optional Display title of the menu item at that path point, otherwist it'll use the last fragment of the path as a display title
|
| 182 |
* @return a menu id. FALSE if it hit the top unexpectedly
|
| 183 |
* */
|
| 184 |
function import_html_create_menu_path($path, $title = '') {
|
| 185 |
debug("Looking for a menu item matching path '$path' ", 3);
|
| 186 |
if ((!module_exists('menu')) || (!module_exists('path'))) {
|
| 187 |
return;
|
| 188 |
}
|
| 189 |
|
| 190 |
// check parent exists
|
| 191 |
if ($pid = menu_get_mid_by_path($path)) {
|
| 192 |
return $pid;
|
| 193 |
}
|
| 194 |
|
| 195 |
debug("Failed to find a menu item '$path', making it.", 2);
|
| 196 |
|
| 197 |
$p_path = dirname($path); // parent path
|
| 198 |
|
| 199 |
if ($p_path == dirname($p_path)) {
|
| 200 |
return import_html_variable('menu_parent_id');
|
| 201 |
/*
|
| 202 |
debug("
|
| 203 |
Looping menu path, this is either a recursion error (illegal paths),
|
| 204 |
or caused by creating a new menu tree from the bottom up and we just
|
| 205 |
reached the top.
|
| 206 |
Although a menu structure for this page is now built, it may be hard
|
| 207 |
to navigate to child pages if the parent doesn't exist.
|
| 208 |
This can be fixed by importing the parents first or creating section header content.
|
| 209 |
", 2);
|
| 210 |
return FALSE;
|
| 211 |
*/
|
| 212 |
}
|
| 213 |
|
| 214 |
if (!$pid = menu_get_mid_by_path($p_path)) {
|
| 215 |
// Parent not found,
|
| 216 |
// Need to make it up...
|
| 217 |
$pid = import_html_create_menu_path($p_path);
|
| 218 |
// That will have climbed up and come back down. with its nearest new ancestor
|
| 219 |
}
|
| 220 |
// and attach self to it
|
| 221 |
|
| 222 |
if ($pid) {
|
| 223 |
if (!$title) {
|
| 224 |
$title = basename($path);
|
| 225 |
if (strstr(".", $title)) {
|
| 226 |
$title = substr($title, 0, strrpos($title, "."));
|
| 227 |
}
|
| 228 |
}
|
| 229 |
$new = array(
|
| 230 |
'title' => $title,
|
| 231 |
'path' => $path,
|
| 232 |
'pid' => $pid,
|
| 233 |
'type' => MENU_CUSTOM_ITEM
|
| 234 |
);
|
| 235 |
debug("import_html_create_menu_path : Defining new menu item '$path' as a child of $pid", 3);
|
| 236 |
|
| 237 |
if (!form_get_errors() && (!menu_get_mid_by_path($path))) {
|
| 238 |
menu_edit_item_save($new);
|
| 239 |
}
|
| 240 |
else {
|
| 241 |
debug_pre(form_get_errors());
|
| 242 |
debug_pre(menu_get_menu(), 4);
|
| 243 |
return;
|
| 244 |
}
|
| 245 |
$mid = menu_get_mid_by_path($path);
|
| 246 |
|
| 247 |
# // need to clear the menu cache for the new item to show up
|
| 248 |
# menu_rebuild();
|
| 249 |
|
| 250 |
debug("New Menu Item '". $new['title'] ."' id is :". $mid, 2);
|
| 251 |
return $mid;
|
| 252 |
}
|
| 253 |
else {
|
| 254 |
debug("Failed to regressively build menu for '$path'. This should not happen. Dunno why we are here.");
|
| 255 |
}
|
| 256 |
}
|
| 257 |
|
| 258 |
|
| 259 |
|
| 260 |
/**
|
| 261 |
* Add path alias support to the import_html process
|
| 262 |
*
|
| 263 |
* Adds an old-style (legacy) alias to the node path if required.
|
| 264 |
*/
|
| 265 |
function path_import_html_after_save($profile, &$node, $datadoc = NULL) {
|
| 266 |
#dpm(array('INVOKED path_import_html_after_save for '. $node->path));
|
| 267 |
if ($profile['legacy_aliases'] && ($node->old_path != $node->path)) {
|
| 268 |
debug("Setting up navigation links for this item now. Directing {$node->old_path} to go to the system path ". drupal_get_normal_path($node->path), 2);
|
| 269 |
|
| 270 |
$normal_path = "node/{$node->nid}";
|
| 271 |
|
| 272 |
if ($normal_path == $node->path) {
|
| 273 |
drupal_set_message(t("Failed to resolve %node_path into a system path. Cannot create alias at this time.", array('%node_path' => $node->path)));
|
| 274 |
}
|
| 275 |
else {
|
| 276 |
path_set_alias($normal_path, $node->old_path);
|
| 277 |
drupal_set_message(t("This document (known internally as '%normal_path' ) should now be accessible via aliases as both '!main_alias' and '!legacy_alias' ",
|
| 278 |
array(
|
| 279 |
'%normal_path' => $normal_path,
|
| 280 |
'!main_alias' => l($node->path, $node->path),
|
| 281 |
'!legacy_alias' => l($node->old_path, $node->old_path)
|
| 282 |
)
|
| 283 |
));
|
| 284 |
}
|
| 285 |
}
|
| 286 |
|
| 287 |
}
|
| 288 |
|
| 289 |
|
| 290 |
/**
|
| 291 |
* Absorbs elements in the import document we recognise as being (probably)
|
| 292 |
* taxonomy terms.
|
| 293 |
*
|
| 294 |
*
|
| 295 |
* Any link rel="tag" is a start. The plaintext content of such a tag is
|
| 296 |
* searched for as a term.
|
| 297 |
*
|
| 298 |
* both <a href rel='tag' > and <link rel="tag"> are good for me.
|
| 299 |
*
|
| 300 |
* @TODO this still needs to be expanded with real-world cases
|
| 301 |
*/
|
| 302 |
function taxonomy_import_html($profile, &$node, $datadoc = NULL) {
|
| 303 |
|
| 304 |
$active_vocabs = taxonomy_get_vocabularies($node->type);
|
| 305 |
if(!$active_vocabs) { return; }
|
| 306 |
// Scan to see if any allow freetagging. After that it's guesswork which one to target. The one with the lowset number;
|
| 307 |
$prime_vocab = 0;
|
| 308 |
foreach($active_vocabs as $vocabulary) {
|
| 309 |
if($vocabulary->tags) {
|
| 310 |
$prime_vocabulary = $vocabulary;
|
| 311 |
break;
|
| 312 |
}
|
| 313 |
}
|
| 314 |
|
| 315 |
// Translate links with rel=tag into taxonomy terms
|
| 316 |
$relationships = xml_query($datadoc, './/*[@rel]');
|
| 317 |
foreach ($relationships as $link) {
|
| 318 |
if (empty($link)) {continue;}
|
| 319 |
$reltype = xml_getattribute($link, 'rel');
|
| 320 |
// rels can be multiple - like classes. Split on space
|
| 321 |
$reltypes = explode(' ', xml_getAttribute($link, 'rel'));
|
| 322 |
$rellink = xml_getattribute($link, 'href');
|
| 323 |
$label = trim(xml_textcontent($link));
|
| 324 |
// Microformats may put their value into a 'title' attribute instead
|
| 325 |
if (empty($label)) {
|
| 326 |
$label = xml_getAttribute($link, 'title');
|
| 327 |
}
|
| 328 |
|
| 329 |
if (!empty($label)) {
|
| 330 |
foreach($reltypes as $reltype) {
|
| 331 |
if ($reltype == 'tag') {
|
| 332 |
// Some of this magic should probably shift into absorb_metas()
|
| 333 |
// Still thinking of the best way to encode tags into the raw HTML
|
| 334 |
|
| 335 |
if (strstr($label, ":")) {
|
| 336 |
// allow an extended format for this value
|
| 337 |
// to support freetext vocab additions
|
| 338 |
// <a rel='tag'>Author:Joe Brown</a> will create a term in the 'Author' vocab called "Joe Brown".
|
| 339 |
debug("Splitting rel link '$label' up to make it a vocab term entry", 2);
|
| 340 |
$vocab_term = split(":", $label);
|
| 341 |
$vocab = $vocab_term[0];
|
| 342 |
$label = $vocab_term[1];
|
| 343 |
$vocabulary = taxonomy_get_vocabulary_by_name($vocab);
|
| 344 |
if ($vocabulary) {
|
| 345 |
$node->taxonomy['tags'][$vocabulary->vid] .= ",". $label;
|
| 346 |
}
|
| 347 |
}
|
| 348 |
|
| 349 |
// Match against existing terms
|
| 350 |
$terms = taxonomy_get_term_by_name($label);
|
| 351 |
// Allow an optional 'S' on the end of terms when looking for a match. Just to be flexible.
|
| 352 |
$terms = array_merge($terms, taxonomy_get_term_by_name($label ."s"));
|
| 353 |
|
| 354 |
if (!is_array($node->taxonomy)) {$node->taxonomy = array();}
|
| 355 |
if ($terms) {
|
| 356 |
foreach($terms as $term) {
|
| 357 |
// If we did successfully identify the vocab earlier, filter on that. Otherwise let anything go
|
| 358 |
if ($vocabulary && ($term->vid != $vocabulary->vid)) { continue; }
|
| 359 |
$node->taxonomy[$term->tid] = $term;
|
| 360 |
}
|
| 361 |
}
|
| 362 |
else {
|
| 363 |
// Couldn't retrieve it, can we add freetags?
|
| 364 |
$node->taxonomy['tags'][$prime_vocabulary->vid] .= ",". $label;
|
| 365 |
}
|
| 366 |
|
| 367 |
} // is a 'tag'
|
| 368 |
import_html_absorb_properties($node, $reltype, $label);
|
| 369 |
} // inner rel tag type
|
| 370 |
} // has value
|
| 371 |
} // each rel
|
| 372 |
}
|
| 373 |
|
| 374 |
|
| 375 |
|
| 376 |
/**
|
| 377 |
* Utility Functions which should exist but don't yet.
|
| 378 |
* out-of-module names sorry.
|
| 379 |
*/
|
| 380 |
|
| 381 |
|
| 382 |
/**
|
| 383 |
* Return a menu item matching a given path or alias.
|
| 384 |
*
|
| 385 |
* Menu_get_item re-writes the path from alias to 'normal' path when it
|
| 386 |
* initializes. This makes it impossible to retrieve menu items by name using the available API.
|
| 387 |
*
|
| 388 |
* @return a menu ID
|
| 389 |
*/
|
| 390 |
function menu_get_mid_by_path($path) {
|
| 391 |
// Caching works against me when adding lots of aliases in one go. Wipe it :(
|
| 392 |
drupal_lookup_path('wipe'); // This is not working?
|
| 393 |
$alias = db_result(db_query("SELECT src FROM {url_alias} WHERE dst = '%s'", $path));
|
| 394 |
|
| 395 |
$menu = menu_get_menu();
|
| 396 |
if ($mid = $menu['path index'][$path]) {
|
| 397 |
return $mid;
|
| 398 |
}
|
| 399 |
if ($alias && $mid = $menu['path index'][$alias]) {
|
| 400 |
return $mid;
|
| 401 |
}
|
| 402 |
|
| 403 |
// it's bad mojo to mess with the DB directly, but menu doesn't provide a lookup API.
|
| 404 |
// Or a way to avoid caching. Do it by hand if I need a newly added menu
|
| 405 |
$row = db_fetch_object(db_query('SELECT * FROM {menu} WHERE path != "" AND (path = "%s" OR path = "%s") ', $path, $alias ));
|
| 406 |
if ($row) {
|
| 407 |
return $row->mid;
|
| 408 |
}
|
| 409 |
}
|
| 410 |
/**
|
| 411 |
* Return a menu item matching a given path or alias.
|
| 412 |
*/
|
| 413 |
function menu_get_item_by_path($path) {
|
| 414 |
$result = db_query('SELECT * FROM {menu} WHERE path="%s"', $path);
|
| 415 |
return db_fetch_array($result);
|
| 416 |
}
|
| 417 |
|
| 418 |
|
| 419 |
function taxonomy_get_vocabulary_by_name($name) {
|
| 420 |
$vs = taxonomy_get_vocabularies();
|
| 421 |
foreach ($vs as $voc) {
|
| 422 |
if ($voc->name == $name) return $voc;
|
| 423 |
}
|
| 424 |
}
|
Parent Directory
| 
Revision Log
| 
Revision Graph