Changed google URL + response in UTF-8
khorpyakov - May 15, 2009 - 07:29
| Project: | Translation Framework |
| Version: | 6.x-1.x-dev |
| Component: | Code |
| Category: | feature request |
| Priority: | minor |
| Assigned: | Unassigned |
| Status: | needs work |
Description
Recently I had noticed that google response with russian translations is in old KOI8-R character encoding. So I have to modify headers in request with this patch so now response is always in UTF-8.
Also new url provided by google return multiply translation variants encoded in comma-separeted arrays with square brackets if the input was a single word.
| Attachment | Size |
|---|---|
| google_translation_module_head.patch | 2.67 KB |
#1
Can you apply this patch against the current DRUPAL-6--1 tag in CVS since the one it is applied against is older than what is currently in that branch.
#2
#3
I have tested the patch in #2 and can confirm that it works now when I try to translate content using Google. The old URL wasn't working at all.
#4
Note that when I used the old URL, I wasn't getting back a response at all.
While this patch works (i.e., the content is translated), the $data->result returned by the drupal_http_request is in JSON format. This means it looks odd in the browser, having all the JSON escape characters in it.
It looks like there are two alternatives:
1) Make the HTTP request in Javascript & then send the result (i.e., the finished translation) to Drupal (so it can be parsed properly in one go using PHP's parse_json function). I haven't tried this.
2) Use drupal_http_request but then parse the JSON, or at least do str_replace to clean up the response. This is what I've done, but my technique is rather poor. I would appreciate any suggestions for improvement.
In _google_translation_split_string:
<?phpif ($result->code == 200) {
$translation_decoded = str_replace('\x3c', '<', $result->data); // escaping for <
$translation_decoded = str_replace('\x3e', '>', $translation_decoded); // escaping for >
$translation_decoded = str_replace('\"', '"', $translation_decoded); // escaping for double quotes - this line is causing some weird output
$translated_string .= $translation_decoded;
?>
In google_translation_postprocess:
<?phpfunction google_translation_postprocess($translate) {
$translate->translation = trim($translate->translation, '"'); // strip leading and trailing quotes (added by JSON response)
return $translate;
}
?>
#5
Never mind. The original Google Translate URL is working for me now. drupal_http_request() was acting flaky for me before, I guess.
I do have a serious issue with the output it's giving me, which I'll post as a separate issue.