servers/ttrss
1
0
Fork 0
mirror of https://tt-rss.org/git/tt-rss.git synced 2024-06-24 11:56:36 +02:00
Code Issues Releases Wiki Activity

limit maximum data length for bayesian processing

Browse Source
This commit is contained in:
Andrew Dolgov 2015-06-18 19:02:39 +03:00
parent bc0e565a77
commit 132e42a9a7
2 changed files with 8 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
plugins/af_sort_bayes/init.php
View File

@ -8,6 +8,7 @@ class Af_Sort_Bayes extends Plugin {
private $score_modifier = 50; private $score_modifier = 50;
private $sql_prefix = "ttrss_plugin_af_sort_bayes"; private $sql_prefix = "ttrss_plugin_af_sort_bayes";
private $auto_categorize_threshold = 10000; private $auto_categorize_threshold = 10000;
private $max_document_length = 3000; // classifier can't rescale output for very long strings apparently
function about() { function about() {
return array(1.0, return array(1.0,
@ -47,7 +48,7 @@ class Af_Sort_Bayes extends Plugin {
if ($this->dbh->num_rows($result) != 0) { if ($this->dbh->num_rows($result) != 0) {
$guid = $this->dbh->fetch_result($result, 0, "guid"); $guid = $this->dbh->fetch_result($result, 0, "guid");
$title = $this->dbh->fetch_result($result, 0, "title"); $title = $this->dbh->fetch_result($result, 0, "title");
$content = mb_strtolower($title . " " . strip_tags($this->dbh->fetch_result($result, 0, "content"))); $content = mb_substr(mb_strtolower($title . " " . strip_tags($this->dbh->fetch_result($result, 0, "content"))), 0, $this->max_document_length);
$score = $this->dbh->fetch_result($result, 0, "score"); $score = $this->dbh->fetch_result($result, 0, "score");
$this->dbh->query("BEGIN"); $this->dbh->query("BEGIN");
@ -302,7 +303,7 @@ class Af_Sort_Bayes extends Plugin {
$dst_category = $id_ugly; $dst_category = $id_ugly;
$bayes_content = mb_strtolower($article["title"] . " " . strip_tags($article["content"])); $bayes_content = mb_substr(mb_strtolower($article["title"] . " " . strip_tags($article["content"])), 0, $this->max_document_length);
if ($count_neutral >= $this->auto_categorize_threshold) { if ($count_neutral >= $this->auto_categorize_threshold) {
// enable automatic categorization // enable automatic categorization
@ -358,7 +359,8 @@ class Af_Sort_Bayes extends Plugin {
if ($this->dbh->num_rows($result) != 0) { if ($this->dbh->num_rows($result) != 0) {
$guid = $this->dbh->fetch_result($result, 0, "guid"); $guid = $this->dbh->fetch_result($result, 0, "guid");
$title = $this->dbh->fetch_result($result, 0, "title"); $title = $this->dbh->fetch_result($result, 0, "title");
$content = mb_strtolower($title . " " . strip_tags($this->dbh->fetch_result($result, 0, "content")));
$content = mb_substr(mb_strtolower($title . " " . strip_tags($this->dbh->fetch_result($result, 0, "content"))), 0, $this->max_document_length);
print "<h2>" . $title . "</h2>"; print "<h2>" . $title . "</h2>";

4
plugins/af_sort_bayes/lib/class.naivebayesianstorage.php
View File

@ -47,6 +47,7 @@
class NaiveBayesianStorage { class NaiveBayesianStorage {
var $con = null; var $con = null;
var $owner_uid = null; var $owner_uid = null;
var $max_document_length = 3000; // classifier can't rescale output for very long strings apparently
function NaiveBayesianStorage($owner_uid) { function NaiveBayesianStorage($owner_uid) {
$this->con = Db::get(); $this->con = Db::get();
@ -239,7 +240,8 @@
$this->con->escape_string($ref['document_id']) . "'"); $this->con->escape_string($ref['document_id']) . "'");
if ($this->con->num_rows($rs) != 0) { if ($this->con->num_rows($rs) != 0) {
$ref['content'] = mb_strtolower($this->con->fetch_result($rs, 0, 'title') . ' ' . strip_tags($this->con->fetch_result($rs, 0, 'content'))); $ref['content'] = mb_substr(mb_strtolower($this->con->fetch_result($rs, 0, 'title') . ' ' . strip_tags($this->con->fetch_result($rs, 0, 'content'))), 0,
$this->max_document_length);
} }
} }