[seer]

OK.... another day... another step closer to the vision ..... more questions

People are goddam lazy when it comes to tagging. I know I am, and don't doubt all submitters will be too. So, does anyone know a way to auto-insert-tags?

I would like this to happen in two different scenarios which would require distinct processes as far as I can tell:

1. During submission ... take all words in the title and put them in the tags box (perhaps with an ignore list for words like 'the' 'and' 'it' etc). The tag box would remain editable so if the user didn't like the suggestions they could change them.

2. RSS imports - same thing take all words form title, put them in tags. Obviously there is no user involved so these would simply be saved.

However there is a 3rd possibility... a job that I can run occasionally that runs on all untagged stories in the db. It could run once an hour and do:
- find stories where tagfield = null
- parse title for those stories for unique words in story title excluding those words we store in a file somewhere (if, the, and, what, etc)
- write remaining words to tagfield

Now here is the biggest problem.... I am a terrible programmer... I can 'hello world' in a million languages but not much more lol.
So is there a solution out there for any of these scenarios? Anyone have something similar that could be adapted?
Or am I screwed in which case which do you think would be the easiest approach for me to try and tackle?

[seer]

Well I have been playing around with the RSS feed side of things today with reasonable success.

Simply creating a field link between the link_title of the feed and the tag field of pligg seems to work most of the time on the few feeds I tried it on. However occasionally it says there was an error in the SQL but think this is due to a symbol such as ? being in the title confusing the query. Plus I get a lot of the crap words like 'the' and 'and' in tags.... there is obviously no intelligence in this method

So right now I would like to focus on the user side of things if anyone has ideas?

[seer]

OK, not that I think many are paying attention but as the vast majority of my content so far is pulled in from RSS feeds I am fairly happy with the solution of creating a field link between the title of the feed and the pligg tag field. It works and puts all the words from the title into the tags (I only have it enabled on a couple of the imported fields atm but that will change soon).

My concern now is only getting rid of all the words from the tags that are not important.

e.g. the, and, at, to, some, many, a, then, that, etc

My only blunt instrument to do this is for me to run a delete query on the db as a cron job. However my host only allows me one cron job (lol) which is being used for the RSS import itself.

So I was wondering if anyone could be my saviour and suggest the code to run this query from the RSS import itself, i.e. appended to the end of the import.php

e.g. delete link_tags from pligg_links where link_tags contains 'at' or 'or' or 'and' or 'that' or 'they' .... etc etc

[Andtony]

Sorry, can't help you with your problem but want to ask something.

When you're using feed title as tags I think the whole title is one tag because there is not commas. Am I wrong?
Does tags that you import from feeds appear on your tag cloud? At least a few months ago there was somekind of bug with this.

[seer]

Originally Posted by Andtony

Sorry, can't help you with your problem but want to ask something.

When you're using feed title as tags I think the whole title is one tag because there is not commas. Am I wrong?
Does tags that you import from feeds appear on your tag cloud? At least a few months ago there was somekind of bug with this.

Crap you are right

so now the query needs to replace spaces with comma & space perhaps? I'll delve into the db itself and see exactly what is written and whether it is correctly showing in the tag cloud.

[blaze]

Originally Posted by seer

Crap you are right

so now the query needs to replace spaces with comma & space perhaps? I'll delve into the db itself and see exactly what is written and whether it is correctly showing in the tag cloud.

Well, I'm sure this won't exactly be helpful, but try using "category" as your tag from the imported feeds. I know it will be only one keyword usually, but at least it will be a useful tag.

On another note, I have been editing the rss feed I import by creating my own xml file for it. I edit the original file and manually insert the tag <keywords> for each item. So my new xml file looks like this:

<?xml version="1.0" encoding="utf-8"?>
<rss version="2.0">
<channel>
<title>Somesite.com</title>
<link>http://www.somesite.com</link>
<description>Here is my description</description>
<language>en</language>
<copyright>Somesite.com</copyright>
<lastBuildDate>Fri, 16 Nov 2007 08:05:02 +0300</lastBuildDate>
<managingEditor>info@somesite.com</managingEditor>
<item>
 <title>My first title</title>
 <link>http://www.somesite.com/my-first-link.html</link>
 <keywords>keyword 1, keyword 2, keyword 3</keywords>
 <description>Pligg is a great script....</description>
 <category>Category 1</category>
 <pubDate>Fri, 16 Nov 2007 08:06:02 +0300</pubDate>
</item>
<item>
 <title>My second title</title>
 <link>http://www.somesite.com/my-second-link.html</link>
 <keywords>keyword 4, keyword 5, keyword 6</keywords>
 <description>Don't you just love pligg?!!</description>
 <category>Category 2</category>
 <pubDate>Fri, 16 Nov 2007 08:07:02 +0300</pubDate>
</item>
</channel>
</rss>

I then save the file and upload it onto my server, whereby importing that file as an rss feed. I then use the keywords_field as a tags_field. It's a lot of extra work (especially if you have a lot of items), but I find it necessary for me to do. It takes me about an hour to do roughly 100 items. I only use one rss feed though, so I am sure this is not exactly the best way for you to do if you have many rss feeds imported, as it would be a lot of extra work.

Sorry I can't be more helpful.

[clems365]

Maybe an other way is compare the words in the title with the data tags, for example if the title of the story is:
"The Truth about Your Seasonal Allergies"

a function could find in the data tags, "the", "truth", "about", "Your","Seasonal" and "Allergies".

If this function find an existing word it will put it in the field tag, we can imagine that "allegies" exist in the data tags then in the field tags we can see this world "allegies".

If not, nothing, so we need to update the field tag

I think in a frist time there a lot of work, but next with a big data tags all will be ok and just update some news with some new tags.

I'm sure it's possible to make this mod...

[Rob472]

If your trying to get keywords from the title and your title is stored in the variable $theTitle for example

Build an array of keywords you don't want put into tags

$search = array('the', 'and', 'of', 'or');

convert your title to lowercase

$theTitle = strtolower($theTitle);

Remove any special characters

$str = preg_replace('#[^\w ]#', '', $theTitle);

Convert it to an array

$arr = explode(' ', $str);

Do the search and replace using the $search array created above

$keywords = str_replace($search, '', $arr);

Unset the empty values in the array

foreach($keywords as $key => $value) {
	if($value == ''){
		unset($keywords[$key]);
    }
}

Now you should have a clean array of keywords you can use for your tags. If you want to make it a comma delimited list then you can implode it

$keyString = implode($keywords, ",");

Hope that helps someone.

[seer]

Originally Posted by Rob472

If your trying to get keywords from the title and your title is stored in the variable $theTitle for example

Build an array of keywords you don't want put into tags

$search = array('the', 'and', 'of', 'or');

convert your title to lowercase

$theTitle = strtolower($theTitle);

Remove any special characters

$str = preg_replace('#[^\w ]#', '', $theTitle);

Convert it to an array

$arr = explode(' ', $str);

Do the search and replace using the $search array created above

$keywords = str_replace($search, '', $arr);

Unset the empty values in the array

foreach($keywords as $key => $value) {
	if($value == ''){
		unset($keywords[$key]);
    }
}

Now you should have a clean array of keywords you can use for your tags. If you want to make it a comma delimited list then you can implode it

$keyString = implode($keywords, ",");

Hope that helps someone.

Really awesome feedback there, THANKS!

I will see what I can do with it today and post back my results.

[seer]

The linking to category option may actually be the simplest solution for what I want (I just want RSS imported items to have at least one tag).

I will try this with a feed and let you all know how it goes. In the meantime I will try and construct something more complex along the lines that Rob suggested.