/**
* 通过关键字获取相关文章id
* @params int $docId 根据用户浏览的文章ID
* @params int $num 需要返回的文章id的数量，最多是36（此参数暂时不用）
* return 返回以带有文章数据和论坛数据的数组, array('article'=>array(5315078=>43,...),'bbs'=>array())
*/
function get_arti_by_word($docIdArr, $num = 12)
{
if (!$docIdArr) {
return false;
}
global $db_guess, $randNum, $needNum, $propertyArr;
/* $classArr = get_class_id_by_docid(array(
'articleIdArr'=>$docIdArr,
));
$classStr = implode(',',array_unique($classArr));
$classNum = count($classArr);
if($classNum>1){
$sqlClassStr = ' class_id in('.$classStr.') ';
}else if($classNum==1){
$sqlClassStr = ' class_id = '.$classStr.' ';
}else{
$sqlClassStr = '';
} */
$wordArr = get_word_results($docIdArr);
$wordStr = '"' . implode('","', $wordArr) . '"';
$sqlWordStr = $wordArr ? ' AND t.word in(' . $wordStr . ') ' : '';
#var_dump($classArr,$classStr);exit('#667-1#');
if (is_array($docIdArr)) {
$docIdStr = implode(',', $docIdArr);
$docIdStr = ' article_id in (' . $docIdStr . ') ';
} else {
$docIdStr = ' article_id=' . $docIdStr . ' ';
}
# 查询的字段
$fields1 = ' t.article_id,t.title,t.uv,count(article_id) as cnt,t.bbsUrl,t.flag ';
$fields2 = ' t.article_id,t.title,t.uv,count(t.bbsUrl) as cnt,t.bbsUrl,t.flag ';
$wheres1 = $sqlWordStr . ' AND t.bbsUrl="" ';
$wheres2 = $sqlWordStr . ' AND t.bbsUrl<>"" AND page_type_id<>4 ';
$order1 = $order2 = ' ORDER BY uv desc ';
# 将小结果集放前边（小结果集驱动大结果集）
# $wordStr = '"京东","苹果","手机","笔记本"';
$sql = '(SELECT ' . $fields1 . ' from tongji_article_title_words t where 1 ' . $wheres1 . ' GROUP BY t.article_id,t.flag ' . $order1 . ' limit 120)
UNION
(SELECT ' . $fields1 . ' from tongji_article_title_words t where 1 ' . $wheres2 . ' ' . $order2 . ' limit 2)';
/**
* @desc START 杨叔说搞一个缓存 add by 任新强 2015-12-30 11:20:21
*/
if (!$sqlWordStr) {
return array();
$mongokey = 'zol:cms:guess:you:like:union:sql:mongo:nb:key';
$mongoDate = ZOL_Api::run("Kv.MongoCenter.get", array('module' => 'cms', 'key' => $mongokey));
if (!$mongoDate) {
$resArr1 = $db_guess->get_results($sql);
ZOL_Api::run("Kv.MongoCenter.set", array('module' => 'cms', 'key' => $mongokey, 'data' => $resArr1, 'life' => 1800));
} else {
$resArr1 = $mongoDate;
}
} else {
$resArr1 = $db_guess->get_results($sql);
}
/**
* @desc END
*/
# 统计词频 + 分词权重
$tmpArr1 = $tmpArr2 = $bbsData = $articleData = array();
$resArr2 = $resArr1;
$bbsDataEnough = false;
# 方案1_1
if ($resArr1) {
foreach ($resArr1 as $k => $v) {
# 只需要取1条论坛数据
if ($v['bbsUrl'] && !$bbsDataEnough) {
$bbsData[] = $v;
if (count($bbsData) > 1) {
$bbsDataEnough = true;
}
}
if ($v['article_id'] == 1) {
continue;
}
if (!array_key_exists($v['article_id'], $tmpArr1)) {
$tmpArr1[$v['article_id']]['word_power_val'] = $propertyArr[$v['flag']] * $v['cnt'];
} else {
$tmpArr1[$v['article_id']]['num']++;
$tmpArr1[$v['article_id']]['word_power_val'] += $propertyArr[$v['flag']] * $v['cnt'];
}
$tmpArr1[$v['article_id']]['uv'] = $v['uv'];
$tmpArr1[$v['article_id']]['article_id'] = $v['article_id'];
}
# 排除用于查找相关文章的文章id
foreach ($docIdArr as $k => $v) {
if (isset($tmpArr1[$v])) {
unset($tmpArr1[$v]);
}
}
# 对数据按照“分词权重”进行倒序
$tmpArr1 = multi_array_sort($tmpArr1, 'word_power_val', SORT_DESC);
$i = 1;
# 每种相似度一个数组，存储“相似度相同”的数据
foreach ($tmpArr1 as $k => $v) {
$newKey = $v['word_power_val'] * 10000;
$tmpArr2[$newKey][] = $v;
}
$tmpArr1 = array();
# 相同相似度的数据按照uv倒序
foreach ($tmpArr2 as $k => $v) {
$tmpArr2[$k] = multi_array_sort($v, 'uv', SORT_DESC);
$tmpArr1 = array_merge($tmpArr1, $tmpArr2[$k]);
}
$tmpArr1 = array_slice($tmpArr1, 0, 12, true);
$tmpArr2 = array();
foreach ($tmpArr1 as $k => $v) {
$tmpArr2[$v['article_id']] = $v;
}
$articleData = $tmpArr2;
}
#var_dump($articleData,$bbsData);
#echo $sql; exit('759-5');
# 文章属性，优先展示第一类属性
#$propertyArr1 = array('nproduct','nmanu','nsubcat','eng','nproperty','ntype','nbooktitle');
#$propertyArr2 = array('n','nr','nz');
if ($resArr1 && is_array($resArr1)) {
# 数量是否足够
$num = count($articleData);
$newArr2 = $articleData;
if ($num >= $needNum) {
return array('article' => $newArr2, 'bbs' => $bbsData);
//return get_from_rand($resArr3);
} else {
//exit('821_1');
return array('article' => $newArr2, 'bbs' => $bbsData);
}
} else {
//mail('*****@*****.**','【ZOL首页自"猜你喜欢"查出的数据不是数组】',"get_arti_by_word\r\n".'查出的数据不是数组'.$sql);
return array();
}
}
/**
* 通过关键字获取相关文章id
* 文章页的要排除：370（Z超值）
* @params int $docId 根据用户浏览的文章ID
* @params int $num 需要返回的文章id的数量，最多是36（此参数暂时不用）
*/
function get_arti_by_word($docIdArr, $num2 = 36)
{
global $randNum, $classId, $db_doc_read, $propertyArr;
if (!$docIdArr) {
return array();
}
$wordArr = get_word_results($docIdArr);
$wordStr = '"' . implode('","', $wordArr) . '"';
$sqlWordStr = $wordArr ? ' AND t.word in(' . $wordStr . ') ' : '';
# 查询的字段
$fields1 = ' t.article_id,t.title,t.uv,count(t.article_id) as cnt,t.bbsUrl,t.flag ';
# 将小结果集放前边（小结果集驱动大结果集）
$wheres1 = $sqlWordStr . ' AND t.bbsUrl="" ';
if (!$sqlWordStr) {
return array();
}
$wheres2 = $sqlWordStr . ' AND t.bbsUrl<>"" AND page_type_id<>4 ';
$order1 = $order2 = ' ORDER BY uv desc ';
$sql = '(SELECT ' . $fields1 . ' from tongji_article_title_words t where 1 ' . $wheres1 . ' GROUP BY t.article_id,t.flag ' . $order1 . ' limit 120)
';
#echo $sql;exit();
$resArr1 = $db_doc_read->get_results($sql);
//201601062133 suhy
/**
* @desc START 杨叔说搞一个缓存 add by 任新强 2015-12-29 20:37:21
*/
// if(!$sqlWordStr) {
// $mongokey = 'zol:cms:keyword:relevance:get:docid:by:ry';
// $mongoDate = ZOL_Api::run("Kv.MongoCenter.get" , array(
// 'module' => 'cms', #业务名
// 'key' => $mongokey, #key
// ));
// if(!$mongoDate){
// $resArr1 = $db_doc_read->get_results($sql);
// ZOL_Api::run("Kv.MongoCenter.set" , array(
// 'module' => 'cms', #业务名
// 'key' => $mongokey, #key
// 'data' => $resArr1, #数据
// 'life' => 60*60*2, #生命期
// ));
// } else {
// $resArr1 = $mongoDate;
// }
// }
/**
* @desc END
*/
# 统计词频 + 分词权重
$tmpArr1 = $tmpArr2 = $bbsData = $articleData = array();
$resArr2 = $resArr1;
$bbsDataEnough = false;
# 方案1_1
if ($resArr1) {
foreach ($resArr1 as $k => $v) {
# 只需要取1条论坛数据
if ($v['bbsUrl'] && !$bbsDataEnough) {
$bbsData[] = $v;
if (count($bbsData) > 1) {
$bbsDataEnough = true;
}
}
if ($v['article_id'] == 1) {
continue;
}
if (!array_key_exists($v['article_id'], $tmpArr1)) {
$tmpArr1[$v['article_id']]['word_power_val'] = $propertyArr[$v['flag']] * $v['cnt'];
} else {
$tmpArr1[$v['article_id']]['num']++;
$tmpArr1[$v['article_id']]['word_power_val'] += $propertyArr[$v['flag']] * $v['cnt'];
}
$tmpArr1[$v['article_id']]['uv'] = $v['uv'];
$tmpArr1[$v['article_id']]['article_id'] = $v['article_id'];
}
# 排除用于查找相关文章的文章id
foreach ($docIdArr as $k => $v) {
if (isset($tmpArr1[$v])) {
unset($tmpArr1[$v]);
}
}
# 对数据按照“分词权重”进行倒序
$tmpArr1 = multi_array_sort($tmpArr1, 'word_power_val', SORT_DESC);
$i = 1;
# 每种相似度一个数组，存储“相似度相同”的数据
foreach ($tmpArr1 as $k => $v) {
$newKey = $v['word_power_val'] * 10000;
$tmpArr2[$newKey][] = $v;
}
$tmpArr1 = array();
# 相同相似度的数据按照uv倒序
foreach ($tmpArr2 as $k => $v) {
$tmpArr2[$k] = multi_array_sort($v, 'uv', SORT_DESC);
$tmpArr1 = array_merge($tmpArr1, $tmpArr2[$k]);
}
$tmpArr1 = array_slice($tmpArr1, 0, 12, true);
$tmpArr2 = array();
foreach ($tmpArr1 as $k => $v) {
$tmpArr2[$v['article_id']] = $v;
}
$articleData = $tmpArr2;
}
#var_dump($articleData);exit('#868-1#');
if ($resArr1 && is_array($resArr1)) {
# 数量是否足够
$num = count($articleData);
$newArr2 = $articleData;
if ($num >= $needNum) {
return array('article' => $newArr2, 'bbs' => $bbsData);
//return get_from_rand($resArr3);
} else {
//exit('821_1');
return array('article' => $newArr2, 'bbs' => $bbsData);
}
} else {
// mail('*****@*****.**','【ZOL首页自"猜你喜欢"查出的数据不是数组】',"get_arti_by_word\r\n".'查出的数据不是数组'.$sql);
return array();
}
}
