MyShaarli/application/bookmark/BookmarkFilter.php

468 lines
16 KiB
PHP

<?php
namespace Shaarli\Bookmark;
use Exception;
use Shaarli\Bookmark\Exception\BookmarkNotFoundException;
/**
* Class LinkFilter.
*
* Perform search and filter operation on link data list.
*/
class BookmarkFilter
{
/**
* @var string permalinks.
*/
public static $FILTER_HASH = 'permalink';
/**
* @var string text search.
*/
public static $FILTER_TEXT = 'fulltext';
/**
* @var string tag filter.
*/
public static $FILTER_TAG = 'tags';
/**
* @var string filter by day.
*/
public static $FILTER_DAY = 'FILTER_DAY';
/**
* @var string filter by day.
*/
public static $DEFAULT = 'NO_FILTER';
/** @var string Visibility: all */
public static $ALL = 'all';
/** @var string Visibility: public */
public static $PUBLIC = 'public';
/** @var string Visibility: private */
public static $PRIVATE = 'private';
/**
* @var string Allowed characters for hashtags (regex syntax).
*/
public static $HASHTAG_CHARS = '\p{Pc}\p{N}\p{L}\p{Mn}';
/**
* @var Bookmark[] all available bookmarks.
*/
private $bookmarks;
/**
* @param Bookmark[] $bookmarks initialization.
*/
public function __construct($bookmarks)
{
$this->bookmarks = $bookmarks;
}
/**
* Filter bookmarks according to parameters.
*
* @param string $type Type of filter (eg. tags, permalink, etc.).
* @param mixed $request Filter content.
* @param bool $casesensitive Optional: Perform case sensitive filter if true.
* @param string $visibility Optional: return only all/private/public bookmarks
* @param bool $untaggedonly Optional: return only untagged bookmarks. Applies only if $type includes FILTER_TAG
*
* @return Bookmark[] filtered bookmark list.
*
* @throws BookmarkNotFoundException
*/
public function filter($type, $request, $casesensitive = false, $visibility = 'all', $untaggedonly = false)
{
if (!in_array($visibility, ['all', 'public', 'private'])) {
$visibility = 'all';
}
switch ($type) {
case self::$FILTER_HASH:
return $this->filterSmallHash($request);
case self::$FILTER_TAG | self::$FILTER_TEXT: // == "vuotext"
$noRequest = empty($request) || (empty($request[0]) && empty($request[1]));
if ($noRequest) {
if ($untaggedonly) {
return $this->filterUntagged($visibility);
}
return $this->noFilter($visibility);
}
if ($untaggedonly) {
$filtered = $this->filterUntagged($visibility);
} else {
$filtered = $this->bookmarks;
}
if (!empty($request[0])) {
$filtered = (new BookmarkFilter($filtered))->filterTags($request[0], $casesensitive, $visibility);
}
if (!empty($request[1])) {
$filtered = (new BookmarkFilter($filtered))->filterFulltext($request[1], $visibility);
}
return $filtered;
case self::$FILTER_TEXT:
return $this->filterFulltext($request, $visibility);
case self::$FILTER_TAG:
if ($untaggedonly) {
return $this->filterUntagged($visibility);
} else {
return $this->filterTags($request, $casesensitive, $visibility);
}
case self::$FILTER_DAY:
return $this->filterDay($request);
default:
return $this->noFilter($visibility);
}
}
/**
* Unknown filter, but handle private only.
*
* @param string $visibility Optional: return only all/private/public bookmarks
*
* @return Bookmark[] filtered bookmarks.
*/
private function noFilter($visibility = 'all')
{
if ($visibility === 'all') {
return $this->bookmarks;
}
$out = array();
foreach ($this->bookmarks as $key => $value) {
if ($value->isPrivate() && $visibility === 'private') {
$out[$key] = $value;
} elseif (!$value->isPrivate() && $visibility === 'public') {
$out[$key] = $value;
}
}
return $out;
}
/**
* Returns the shaare corresponding to a smallHash.
*
* @param string $smallHash permalink hash.
*
* @return array $filtered array containing permalink data.
*
* @throws \Shaarli\Bookmark\Exception\BookmarkNotFoundException if the smallhash doesn't match any link.
*/
private function filterSmallHash($smallHash)
{
foreach ($this->bookmarks as $key => $l) {
if ($smallHash == $l->getShortUrl()) {
// Yes, this is ugly and slow
return [$key => $l];
}
}
throw new BookmarkNotFoundException();
}
/**
* Returns the list of bookmarks corresponding to a full-text search
*
* Searches:
* - in the URLs, title and description;
* - are case-insensitive;
* - terms surrounded by quotes " are exact terms search.
* - terms starting with a dash - are excluded (except exact terms).
*
* Example:
* print_r($mydb->filterFulltext('hollandais'));
*
* mb_convert_case($val, MB_CASE_LOWER, 'UTF-8')
* - allows to perform searches on Unicode text
* - see https://github.com/shaarli/Shaarli/issues/75 for examples
*
* @param string $searchterms search query.
* @param string $visibility Optional: return only all/private/public bookmarks.
*
* @return array search results.
*/
private function filterFulltext($searchterms, $visibility = 'all')
{
if (empty($searchterms)) {
return $this->noFilter($visibility);
}
$filtered = array();
$search = mb_convert_case(html_entity_decode($searchterms), MB_CASE_LOWER, 'UTF-8');
$exactRegex = '/"([^"]+)"/';
// Retrieve exact search terms.
preg_match_all($exactRegex, $search, $exactSearch);
$exactSearch = array_values(array_filter($exactSearch[1]));
// Remove exact search terms to get AND terms search.
$explodedSearchAnd = explode(' ', trim(preg_replace($exactRegex, '', $search)));
$explodedSearchAnd = array_values(array_filter($explodedSearchAnd));
// Filter excluding terms and update andSearch.
$excludeSearch = array();
$andSearch = array();
foreach ($explodedSearchAnd as $needle) {
if ($needle[0] == '-' && strlen($needle) > 1) {
$excludeSearch[] = substr($needle, 1);
} else {
$andSearch[] = $needle;
}
}
// Iterate over every stored link.
foreach ($this->bookmarks as $id => $link) {
// ignore non private bookmarks when 'privatonly' is on.
if ($visibility !== 'all') {
if (!$link->isPrivate() && $visibility === 'private') {
continue;
} elseif ($link->isPrivate() && $visibility === 'public') {
continue;
}
}
// Concatenate link fields to search across fields.
// Adds a '\' separator for exact search terms.
$content = mb_convert_case($link->getTitle(), MB_CASE_LOWER, 'UTF-8') .'\\';
$content .= mb_convert_case($link->getDescription(), MB_CASE_LOWER, 'UTF-8') .'\\';
$content .= mb_convert_case($link->getUrl(), MB_CASE_LOWER, 'UTF-8') .'\\';
$content .= mb_convert_case($link->getTagsString(), MB_CASE_LOWER, 'UTF-8') .'\\';
// Be optimistic
$found = true;
// First, we look for exact term search
for ($i = 0; $i < count($exactSearch) && $found; $i++) {
$found = strpos($content, $exactSearch[$i]) !== false;
}
// Iterate over keywords, if keyword is not found,
// no need to check for the others. We want all or nothing.
for ($i = 0; $i < count($andSearch) && $found; $i++) {
$found = strpos($content, $andSearch[$i]) !== false;
}
// Exclude terms.
for ($i = 0; $i < count($excludeSearch) && $found; $i++) {
$found = strpos($content, $excludeSearch[$i]) === false;
}
if ($found) {
$filtered[$id] = $link;
}
}
return $filtered;
}
/**
* generate a regex fragment out of a tag
*
* @param string $tag to to generate regexs from. may start with '-' to negate, contain '*' as wildcard
*
* @return string generated regex fragment
*/
private static function tag2regex($tag)
{
$len = strlen($tag);
if (!$len || $tag === "-" || $tag === "*") {
// nothing to search, return empty regex
return '';
}
if ($tag[0] === "-") {
// query is negated
$i = 1; // use offset to start after '-' character
$regex = '(?!'; // create negative lookahead
} else {
$i = 0; // start at first character
$regex = '(?='; // use positive lookahead
}
$regex .= '.*(?:^| )'; // before tag may only be a space or the beginning
// iterate over string, separating it into placeholder and content
for (; $i < $len; $i++) {
if ($tag[$i] === '*') {
// placeholder found
$regex .= '[^ ]*?';
} else {
// regular characters
$offset = strpos($tag, '*', $i);
if ($offset === false) {
// no placeholder found, set offset to end of string
$offset = $len;
}
// subtract one, as we want to get before the placeholder or end of string
$offset -= 1;
// we got a tag name that we want to search for. escape any regex characters to prevent conflicts.
$regex .= preg_quote(substr($tag, $i, $offset - $i + 1), '/');
// move $i on
$i = $offset;
}
}
$regex .= '(?:$| ))'; // after the tag may only be a space or the end
return $regex;
}
/**
* Returns the list of bookmarks associated with a given list of tags
*
* You can specify one or more tags, separated by space or a comma, e.g.
* print_r($mydb->filterTags('linux programming'));
*
* @param string $tags list of tags separated by commas or blank spaces.
* @param bool $casesensitive ignore case if false.
* @param string $visibility Optional: return only all/private/public bookmarks.
*
* @return array filtered bookmarks.
*/
public function filterTags($tags, $casesensitive = false, $visibility = 'all')
{
// get single tags (we may get passed an array, even though the docs say different)
$inputTags = $tags;
if (!is_array($tags)) {
// we got an input string, split tags
$inputTags = preg_split('/(?:\s+)|,/', $inputTags, -1, PREG_SPLIT_NO_EMPTY);
}
if (!count($inputTags)) {
// no input tags
return $this->noFilter($visibility);
}
// If we only have public visibility, we can't look for hidden tags
if ($visibility === self::$PUBLIC) {
$inputTags = array_values(array_filter($inputTags, function ($tag) {
return ! startsWith($tag, '.');
}));
if (empty($inputTags)) {
return [];
}
}
// build regex from all tags
$re = '/^' . implode(array_map("self::tag2regex", $inputTags)) . '.*$/';
if (!$casesensitive) {
// make regex case insensitive
$re .= 'i';
}
// create resulting array
$filtered = [];
// iterate over each link
foreach ($this->bookmarks as $key => $link) {
// check level of visibility
// ignore non private bookmarks when 'privateonly' is on.
if ($visibility !== 'all') {
if (!$link->isPrivate() && $visibility === 'private') {
continue;
} elseif ($link->isPrivate() && $visibility === 'public') {
continue;
}
}
$search = $link->getTagsString(); // build search string, start with tags of current link
if (strlen(trim($link->getDescription())) && strpos($link->getDescription(), '#') !== false) {
// description given and at least one possible tag found
$descTags = array();
// find all tags in the form of #tag in the description
preg_match_all(
'/(?<![' . self::$HASHTAG_CHARS . '])#([' . self::$HASHTAG_CHARS . ']+?)\b/sm',
$link->getDescription(),
$descTags
);
if (count($descTags[1])) {
// there were some tags in the description, add them to the search string
$search .= ' ' . implode(' ', $descTags[1]);
}
};
// match regular expression with search string
if (!preg_match($re, $search)) {
// this entry does _not_ match our regex
continue;
}
$filtered[$key] = $link;
}
return $filtered;
}
/**
* Return only bookmarks without any tag.
*
* @param string $visibility return only all/private/public bookmarks.
*
* @return array filtered bookmarks.
*/
public function filterUntagged($visibility)
{
$filtered = [];
foreach ($this->bookmarks as $key => $link) {
if ($visibility !== 'all') {
if (!$link->isPrivate() && $visibility === 'private') {
continue;
} elseif ($link->isPrivate() && $visibility === 'public') {
continue;
}
}
if (empty(trim($link->getTagsString()))) {
$filtered[$key] = $link;
}
}
return $filtered;
}
/**
* Returns the list of articles for a given day, chronologically sorted
*
* Day must be in the form 'YYYYMMDD' (e.g. '20120125'), e.g.
* print_r($mydb->filterDay('20120125'));
*
* @param string $day day to filter.
*
* @return array all link matching given day.
*
* @throws Exception if date format is invalid.
*/
public function filterDay($day)
{
if (!checkDateFormat('Ymd', $day)) {
throw new Exception('Invalid date format');
}
$filtered = array();
foreach ($this->bookmarks as $key => $l) {
if ($l->getCreated()->format('Ymd') == $day) {
$filtered[$key] = $l;
}
}
// sort by date ASC
return array_reverse($filtered, true);
}
/**
* Convert a list of tags (str) to an array. Also
* - handle case sensitivity.
* - accepts spaces commas as separator.
*
* @param string $tags string containing a list of tags.
* @param bool $casesensitive will convert everything to lowercase if false.
*
* @return array filtered tags string.
*/
public static function tagsStrToArray($tags, $casesensitive)
{
// We use UTF-8 conversion to handle various graphemes (i.e. cyrillic, or greek)
$tagsOut = $casesensitive ? $tags : mb_convert_case($tags, MB_CASE_LOWER, 'UTF-8');
$tagsOut = str_replace(',', ' ', $tagsOut);
return preg_split('/\s+/', $tagsOut, -1, PREG_SPLIT_NO_EMPTY);
}
}