TagCloud - auto create META keywords for a single page

238 Posts

Can someone workout how to modify the TagCloud to auto create META keywords for a single page?
Wouldn’t it be great to have each page on your website with individual keywords specific to the page?
I know MODx can do it but it has to be done manually and it’s very "clunky"

I don’t know any PHP so I can’t work out how to mod the TagCloud script

Web Development, Web Hosting & Search Engine Marketing by:

Vitalized | UK
w. www.vitalized.co.uk

Website Design | Search Engine Marketing (SEM) | UK MODx web hosting, secure, fast & 100% MODx compatible

Vitalized | Australia
w. www.vitalized-australia.com.au

Website Design | Search Engine Marketing (SEM) | Australian MODx web hosting, secure, fast & 100% MODx compatible

186 Posts

silent Reply #2, 20 years, 2 months ago

How should list of keywords be defined? By simply counting the words of the page’s (doc’s) content?
And setting the most frequent into the keyword META?
I agree that the original keyword-feature is a bit "clunky", but for the individual keywords I would use a TV, not TagCloud.

1,097 Posts

PaulGregory Reply #3, 20 years, 2 months ago

I agree that a TV is easier to manage for keywords than the built-in system.

There may be some merit in using the words of a document to "suggest" keywords in the manager or something. And a routine that added keywords to blog post entries when they were saved would be useful.

No, I don't know what OpenGeek's saying half the time either.
MODx Documentation: The Wiki | My Wiki contributions | Main MODx Documentation
Forum: Where to post threads about add-ons | Forum Rules
Like MODx? donate (and/or share your resources)
Like me? See my Amazon wishlist
MODx "Most Promising CMS" - so appropriate!

186 Posts

silent Reply #4, 20 years, 2 months ago

mmm.
Maybe there’s a combination of the TagCloud(Inc) approach and a Snippet which could do that job:

- create a list of words - in a chunk - which may possibly appear as keywords (you do not want to have "and" or "You" as META, I assume)
- then place a snippet in the head/keyword section of the template, that checks which of the words in the list appear in the current document and returns them as meta keywords. (Maybe they have to appear more than once to make a match.)

With TagCloudInc as a starting point, I think its very doable.
EDIT: because it’s not so much of adding/inventing something to TagCloudInc, but to remove a lot of the already existing stuff. (Collecting the content of the whole site, creating links, etc.)
EDIT2: @PaulGregory: the snippet could also be used in a TV, so that your use-case of keywording blog posts could also be covered.

238 Posts

Si Reply #5, 20 years, 2 months ago

OK this is how I see a Keyword auto generator working... I think I’ll call it MetaKeywordGenerator (MKG) Cleaver name! cool

1. Create a list of words that won’t be used (a, the, in ,at etc)
2. Have the ability to not use words with a certain amount of letters. Example: 1, 2 or 3 letter words.
We should have to option to turn on/off this feature in the snippet and tell it how many letters the exclusion words should be
3. The page should be scanned for phrases as well as single words
4. The snippet should have an option for the user to define how many repetitions of a word or phrase there should be to define it as a keyword/keyphrase
5. The snippet should let the user define keywords/keyphrases that are generic and will appear at the end of the keywords everytime the snippet in called
6. It might help the MKG if it could read alt tags for keywords as well as most people use keyword rich alt tags in their SEO

Will the MKG also read: page title, H1 tags, page name, meta description, title tags, a href=, link text, li tags, in get it’s keywords???

Anyone else care to add anything?.. I’m very tired at the moment so I’m sure I’m missing something!

EDIT: Forgot to say... Thanks for all your replies silent, let me know what you think to my suggestions

Web Development, Web Hosting & Search Engine Marketing by:

Vitalized | UK
w. www.vitalized.co.uk

Website Design | Search Engine Marketing (SEM) | UK MODx web hosting, secure, fast & 100% MODx compatible

Vitalized | Australia
w. www.vitalized-australia.com.au

Website Design | Search Engine Marketing (SEM) | Australian MODx web hosting, secure, fast & 100% MODx compatible

186 Posts

silent Reply #6, 20 years, 2 months ago

OK, web_designer, then let’s get going:

1. Create a list of words that won’t be used (a, the, in ,at etc)
This list might become virtually endless: What about "you" "and" "me" ? laugh

So I prefer the approach of creating a list of words which might be of relevance for the site, an include-list, not an exclude-list.
2. Have the ability to not use words with a certain amount of letters. Example: 1, 2 or 3 letter words.
Well, if, for instance, you sell a product with an incredibly cool acronym, like "MOD"? You’d lock it out. Length is quantitative, whereas the semantics of keywords should be qualitative.
3. The page should be scanned for phrases as well as single words
No problem: a keyword, in the technical sense of the term, is just a sequence of characters (including the space character) delimited by a delimiter (a comma, for instance). Feature already present in the existing snippet.
4. The snippet should have an option for the user to define how many repetitions of a word or phrase there should be to define it as a keyword/keyphrase
Feature already present in the existing snippet.
5. The snippet should let the user define keywords/keyphrases that are generic and will appear at the end of the keywords everytime the snippet in called
I guess this should be done in the template or in a template-variable. No problem either.
6. It might help the MKG if it could read alt tags for keywords as well as most people use keyword rich alt tags in their SEO
Hm, not sure what you’re talking about here. Alt-tags for links? (see next topic)
Will the MKG also read: page title, H1 tags, page name, meta description, title tags, a href=, link text, li tags, in get it’s keywords???
It will read everything which is outputted as the content of the page, normally everything between <body> and </body>

BTW: Thanks for your PM, if we come closer to your specific settings we’d better continue the discussion on this channel. But for now, I think, it is still relevant for other "keyworders", so let us stay here.

Regards, and come back - after a good amount of hours of sleep! wink

238 Posts

Si Reply #7, 20 years, 2 months ago

1. Create a list of words that won’t be used (a, the, in ,at etc)
This list might become virtually endless: What about "you" "and" "me" ? Cheesy
So I prefer the approach of creating a list of words which might be of relevance for the site, an include-list, not an exclude-list.

I think it will be very effective, people can also add their own site spacific words to the list. Links below, I think the first link is best:
http://rdsweb2.rdsinc.com/help/stopword_list.html
http://www.dcs.gla.ac.uk/idom/ir_resources/linguistic_utils/stop_words
http://www.ranks.nl/tools/stopwords.html
The snippet should allow users to turn on/off this feature because they may use some of the stop words in their targeted keywords.
It would be a good idea to have an include list as well, as an option. The problem with only using an include list is their is still work to do building the list, and for a blog site that publishes an article(s) on everything and anything everyday it will be quite annoying to remember to update the list everytime... Just my thoughts... I know I’ll find it annoying anyway wink

2. Have the ability to not use words with a certain amount of letters. Example: 1, 2 or 3 letter words.
Well, if, for instance, you sell a product with an incredibly cool acronym, like "MOD"? You’d lock it out. Length is quantitative, whereas the semantics of keywords should be qualitative.
It was just a thought. Possibly a bad one... Possibly not!

3. The page should be scanned for phrases as well as single words
No problem: a keyword, in the technical sense of the term, is just a sequence of characters (including the space character) delimited by a delimiter (a comma, for instance). Feature already present in the existing snippet.
Good, makes life easy

4. The snippet should have an option for the user to define how many repetitions of a word or phrase there should be to define it as a keyword/keyphrase
Feature already present in the existing snippet.
Good, makes life easy

5. The snippet should let the user define keywords/keyphrases that are generic and will appear at the end of the keywords everytime the snippet in called
I guess this should be done in the template or in a template-variable. No problem either.
I know the snippet isn’t technically hard to create, just with I knew some PHP

6. It might help the MKG if it could read alt tags for keywords as well as most people use keyword rich alt tags in their SEO
Hm, not sure what you’re talking about here. Alt-tags for links? (see next topic)
Did I say Alt-tags for links?... oops!... you covered this in the next bit

Will the MKG also read: page title, H1 tags, page name, meta description, title tags, a href=, link text, li tags, in get it’s keywords???
It will read everything which is outputted as the content of the page, normally everything between <body> and </body>
Can you make it scan the page from <html> to </html> or rather from html to </html> because I use <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en"> as the first html tag... or even <head> to </body> ... I only say this because the <title> and "meta name="description" content" tags are always full of keywords

Web Development, Web Hosting & Search Engine Marketing by:

Vitalized | UK
w. www.vitalized.co.uk

Website Design | Search Engine Marketing (SEM) | UK MODx web hosting, secure, fast & 100% MODx compatible

Vitalized | Australia
w. www.vitalized-australia.com.au

Website Design | Search Engine Marketing (SEM) | Australian MODx web hosting, secure, fast & 100% MODx compatible

186 Posts

silent Reply #8, 20 years, 2 months ago

Ups! Things are getting quite complicated, just because of the quoting-mechanics.
I’ll try to simplify it:

Quote from: web_designer at Aug 15, 2006, 04:09 PM

for a blog site that publishes an article(s) on everything and anything everyday it will be quite annoying to remember to update the list everytime... Just my thoughts... I know I’ll find it annoying anyway

True. I assumed that the site has some kind of focus, not just everything and anything.

Quote from: web_designer at Aug 15, 2006, 04:09 PM

3. The page should be scanned for phrases as well as single words
No problem: a keyword, in the technical sense of the term, is just a sequence of characters (including the space character) delimited by a delimiter (a comma, for instance). Feature already present in the existing snippet.
Good, makes life easy

The phrases will only work with an include list. Otherwise it would be hard - I mean: impossible - to figure out what to treat as a word and what as a phrase. And, in the end: I do not think you want to end up with a mess of phrases in keyWORDS...

Quote from: web_designer at Aug 15, 2006, 04:09 PM

5. The snippet should let the user define keywords/keyphrases that are generic and will appear at the end of the keywords everytime the snippet in called
I guess this should be done in the template or in a template-variable. No problem either.
I know the snippet isn’t technically hard to create, just with I knew some PHP

I guess with "user" you mean a "manager" or admin. Webusers (like guys who just post to the blog) are not concerned with keyword-management - nor should they.

Quote from: web_designer at Aug 15, 2006, 04:09 PM

Will the MKG also read: page title, H1 tags, page name, meta description, title tags, a href=, link text, li tags, in get it’s keywords???
It will read everything which is outputted as the content of the page, normally everything between <body> and </body>
Can you make it scan the page from <html> to </html> or rather from html to </html> because I use <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en"> as the first html tag... or even <head> to </body> ... I only say this because the <title> and "meta name="description" content" tags are always full of keywords

There is a type of concurrence coming up here, since the "MKG" might not be the only snippet in action. So: who comes last and scans the output of the rest? Since there is no option right now to configure the sequence of snippet-execution this will be a matter of testing and trying...

Your concerns are getting more clear for me - I hope they do so for you too wink

. So it might take just another two posts to get things done as well as they can be done at the moment in MODx 0.9.2 ... grin

EDIT: I think we should get a bit more specific here: What will your site be all about? What is the purpose of the keywords used?
2ND EDIT: And if we do so, we should definitely change to PM/Email, since this will be of no interest for the rest of the community!

238 Posts

Si Reply #9, 20 years, 2 months ago

Right, this is getting fun! grin

... Here goes

I assumed that the site has some kind of focus, not just everything and anything.

"Assumption is the mother of all f**k ups"... Had to through that in as it’s one of my fav quotes hehehe

I do see where your coming from but it feels like it’s missing the point a little... It would be great to not to have to worry about adding keywords anywhere!
That’s the fundermentle point of my idea. If TagCloud can do it then i’m sure we/you can crack it.

The phrases will only work with an include list. Otherwise it would be hard - I mean: impossible - to figure out what to treat as a word and what as a phrase. And, in the end: I do not think you want to end up with a mess of phrases in keyWORDS...

Oh I think we most definatly want phrases in keywords! (as well as keywords)... Do you not use keyphrases in Meta keywords?
Example meta keywords from some of the largest/best websites out there:

BBC
<meta name="keywords" content="BBC, bbc.co.uk, Search, British Broadcasting Corporation, BBCi" />

MODx
<meta http-equiv="keywords" content="web 2.0, tattoo, php, open source, mysql, modx, gpl, free cms, content management system, content management, cms, application framework, ajax" />

LOVE FILM
<meta name="keywords" content="rental, dvd rental, rent dvd, rent movie, rent movies, rental dvd uk, rent dvd uk, film, love film, films, love films, lovefilm, LOVEFiLM" />

Example page on my website:
http://www.vitalized.co.uk/blog/text-link-ads.htm
The MKG would find:
<meta name="Keywords" content="text link advertising, text link ads, advertising, campaign, website, search engine optimisation" />
I would like it to find single keywords as well, just an example above. all the key phrases appear at least three times.

I guess with "user" you mean a "manager" or admin

Sorry yes that what I meant

2ND EDIT: And if we do so, we should definitely change to PM/Email, since this will be of no interest for the rest of the community!

No way, this will be a very popular snippet.. in fact I think it will make it into the next release of MODx lol... well maybe wink

Web Development, Web Hosting & Search Engine Marketing by:

Vitalized | UK
w. www.vitalized.co.uk

Website Design | Search Engine Marketing (SEM) | UK MODx web hosting, secure, fast & 100% MODx compatible

Vitalized | Australia
w. www.vitalized-australia.com.au

Website Design | Search Engine Marketing (SEM) | Australian MODx web hosting, secure, fast & 100% MODx compatible

186 Posts

silent Reply #10, 20 years, 2 months ago

So here comes my snippet "PageKeywords". It’s nothing than a simplified version of "TagCloudInc".
Looks for a chunk "cloudwords", which should contain a comma-separated list of words and phrases.
Then it gets content, pagetitle and description of the current page and returns a cs-list of those words/phrases which match an entry in "cloudwords". Params: min = the number of occurrences a word/phrase must have in the current doc to be returned by the snippet. (defaults to 1).
Hope you find it useful.
EDIT: intended to be used in the <head> section of a template like this:
<meta name="Keywords" content="[[PageKeywords]]" />