Google News Extension

[cablock]

Has anyone developed a dynamic XML map that lists weblog entries? I already publish to Google news with another site and it requires a particular format as well as (crucially) data lookup to keep it current.

The syntax is along the lines of:

<url>
<loc>(get link)</loc>
<n:news>
<n:publication>
<n:name>My site</n:name>
<n:language>en</n:language>
</n:publication>
<n:genres>UserGenerated</n:genres>
<n:title><%=enc((rs("headline")))%></n:title>
<n:publication_date></n:publication_date>
<n:keywords></n:keywords>
</n:news>
</url>

I'm really struggling to resolve this. Is there a dynamic extension I've missed that might help me generate a weblog map or something close?

[Harm10]

How about http://extensions.pivotx.net/entry/9/xml_sitemap?

[cablock]

Thanks, I got this to work by rewriting the sitemap extension to output as follows:

<urlset>
<url>
<loc>http://www.siteurl.com/newsitem</loc>
<n:news>
<n:publication>
<n:name>News Source Name</n:name>
<n:language>en</n:language>
</n:publication>
<n:genres>UserGenerated</n:genres>
<n:title>Story Headline</n:title>
<n:publication_date>2011-07-21</n:publication_date>
<n:keywords>Some Keywords</n:keywords>
</n:news>
</url>

This is the format Google prefers for news.

However, I haven't managed to find how items are ordered (ascending/descending). If I can figure that out, I can add this extension for anyone else wanting to output entries to Google News. Any help in setting the order of items appreciated. It should start with most recent rather than first entry ever posted.

Thanks

[cablock]

Also, you have to call Google sitemap schemes as in second line I added, to get the parameters needed to display as above.

<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9"
xmlns:n="http://www.google.com/schemas/sitemap-news/0.9">

Forgive my lack of code familiarity (I'm a designer) but I can follow the logic and have at least got it working

Setting a limit on how many entries show would also be useful.

[hansfn]

Since you have the extension almost ready, I'll give you the missing pieces in stead writing it myself. Please contribution the extension to the community when it's complete - see http://extensions.pivotx.net/page/extensions-guidelines (You can send it directly to me if you prefer - hansfn@pivotx.net)

Since the XML file should list news, I assume that you aren't including pages. So basically you are only listing the last N posted entries, right? If so, you can remove the "Iterate through the chapters and pages" part of the code if you haven't done so already. In addition, in the "Iterate through the entries in batches" part of the code, you don't need batches since you only want the last N entries. Just remove the outer $batches for loop and start directly at the read_entries call:

Code: Select all

        $entries = $PIVOTX['db']->read_entries(array('show' => N, 'order' => 'desc', 'full' => false, 'status'=>'publish'));
        foreach ($entries as $entry) {
            ...

The parameters for read_entries is documented.

Finally, you might want to make N, the number of news entries, a config setting, right? Look at the function xml_sitemapAdmin and modify.

Before submitting the code to us, make sure to rename everything from xml_sitemapwhatever to xml_news_sitemapwhatever (or someting similar).

Sorry about the late reply.

[cablock]

Thank you very much. Yes I am close now. If I get stuck I'll post back. I also need to test once more that Google will accept the map without errors. A preliminary check a few days ago seemed to indicate that the output was spot on. - it was just the ordering and amount of links that needed resolving before I can say its done.

[cablock]

Finally got around to setting this up. I only have two parameters I'm struggling with now - how to get the entry title and keywords. Whats the call for these?

http://www.cabnuneaton.org.uk/?xml_sitenewsmap

source is:

$output_item = <<<EOM
<url>
<loc>${host}%loc%</loc>
<n:news>
<n:publication>
<n:name>Brancab News</n:name>
<n:language>en</n:language>
</n:publication>
<n:genres>UserGenerated</n:genres>
<n:title>Need Title Here</n:title>
<n:publication_date>%lastmod%</n:publication_date>
<n:keywords>Need keywords/Tags here</n:keywords>
</n:news>
</url>
EOM;

Can I assume that to pull in this data, I actually need to work on these lines or those that follow?

$PIVOTX['db']->read_entry($entry['uid']);
$item = $output_item;
$item = str_replace('%loc%', $link, $item);

Presumably its something like $item = str_replace('%titel', ______, ______);

[hansfn]

You should quote code carefully so you don't introduce errors:

Code: Select all

            $entry = $PIVOTX['db']->read_entry($entry['uid']);
            $item = $output_item;
            $item = str_replace('%loc%', $link, $item);
            $item = str_replace('%lastmod%', formatDate($entry['edit_date'], '%year%-%month%-%day%'), $item);

The lines you need are something like:

Code: Select all

            $item = str_replace('%title%', $entry['title'], $item);
            $item = str_replace('%keywords%', $entry['keywords'], $item);


[cablock]

Perfect! (see link above). Right, I'll tidy everything up and remove the code not needed and package it all up. I can't be the only one that could really use this extension. I was looking in the right place then. I'm not familiar with PHP but did figure out the logic

[hansfn]

I tried your URL above. Looks OK, but the keywords should be comma-separated (according to the docs).

Replace

Code: Select all

            $item = str_replace('%keywords%', $entry['keywords'], $item);

with

Code: Select all

            $item = str_replace('%keywords%', preg_replace('/[ ]+/', ', ', $entry['keywords']), $item);

PS! Consider submitting the extension to our site - see http://extensions.pivotx.net/page/extensions-guidelines