Gossamer Forum
Home : Products : Gossamer Links : Development, Plugins and Globals :

[NEW] pre-alpha version of Nexus Links Keyword import

Quote Reply
[NEW] pre-alpha version of Nexus Links Keyword import
I've posted about the Nexus Links concept for awhile. Do a search on it if you want more, probably as far as 2 years back. But, as we build more directories, this concept has gotten closer and closer to fruition, and it's starting to bear delicious fruit!!

For the past 4-5 months, I've been working on a project for a customer, that just went south, for bizarre reasons. They deep sixed our agreement for no apparant reason, and left me holding a lot of code, and development work that was unpaid for, and about half finished. It was a "time" based agreement for the past 2 months, which, well, meant I was doing a lot of work based on future payment.... Not a bad thing for everyone, if everyone remained honest. I was cleaning up some plugins, rewriting code, and working on new code for just for them. The "just for them" stuff I'm fuming over, but the other stuff was time well spent.

The bottom line here, is that what they needed for the core of their business, was pretty much exactly the Nexus Links project I've been tossing around for over 2 years.

In short, a means to update and maintain niche directories, with imports from DMOZ (and perhaps elsewhere) on a regular basis.

Nexus Links as a concept was to start with a DMOZ import, then clean it up, add to it, improve the descriptions, etc. Every month or so when they release a new update, import in _NEW_ links, flag changed links, and check deleted links (links get deleted from DMOZ for politics, so no automatic deletions).

Sites would subscribe to the Nexus Project, and share their link database and descriptions, and pull from the "Nexus". It is a partial community project. The difference being, it is subscription based, and not "Open" to scrapping, and such. It requires contributions, as much as it offers.

Sites could subscribe to updates based on keywords, and other criteria.

Imports from this "Nexus" could be made to your site, via update announcements or even automated.

I'm not going to rehash all the details here.

Bottom Line: I have the keyword export routines working, so that if you wanted to find _ALL_SITES_ with "Blue Widgets" in their name, and import them into a category of "Blue Widget Sellers" you could do that. You don't have to search for categories, import by category, or deal with the dmoz dumps.

Next month, when the new update comes out, you can find all sites _ADDED_ or _CHANGED_ since your last import, check to make sure they weren't already in your database, then add them.

Coupled to that, is the Thumbnailer and other features we offer -- comments, advanced ratings, etc. Lots of stuff coming really soon!

I'm now working on the importer, and that is a bear and a half, mated with a T-Rex.

What I want the importer to do, is pretty much everything <G>.

The first version of the importer, will be some what "messy" designed for new sites mainly.

It will take the keyword dump, and put the links into the categories they were in, rooted where you want. (Meaning, you can cut off long parts of category names, or make simple changes to the root paths). Or, it will take all the links it finds, and puts them into ONE category, with the old category name as a entry in the Links table you can use later on. (this would work nicely for existing sites, as only _new_ links will be imported, and could be moved elsewhere easily).

Eventually, the import plug in for your site, will allow you to customize relationships on path names, so that import links can be mapped to your own directories, and new links found in /cat1/cat2/cat3/ can be imported into your directory at /new1/new2 or anywhere else, automatically.

The *main* program, the "nexus" is *not* going to attempt to allow you to make any mappings. By disallowing that, I've cut 90% or more of the work, frustration, and pointless time wasting from the project. I had been hung up for months working on how to map the exports to your imports, then realized, *I* didn't have to! <G> If I have a standard import system, to get the links into the Nexus, then all I need is a standard export to get them out! *where* they go after that, is up to you !

So, the Nexus is just a big cloud of links, that sucks in what it can find, and pretties it up.

On YOUR site, you develop your own mappings for incoming links.

This makes the import plugin more complex, *but* it also allows YOU to modify it, work with it, and be assured that the data coming from the Nexus, is going to slide through your filters the way you want.

Make sense????

No two sites are alike, so the Nexus shouldn't have to deal with sites at all. Just with the importer's back end.

I should have a basic importer soon, and at that point, it will be ready for public consumption.

Nexus Links is not a mass-market solution. I've thought about that. This is a value-added solution, and by keeping the spiders, search engines, and casual observers out, we can decrease costs (no scraping sites, no leeches, etc) and offer higher quality content. It's an extension of something for people seriously building higher-end niche/targeted databases, not simply massing links.

BTW: The keyword exporter/import will be available on a basic DMOZ dump for seeding your sites. It will be a standard commercial plugin. You can either run it against your own DMOZ install, or subscribe to our DMOZ install. You need a DMOZ database running in GLinks to do the searches/export, and then you need the importer to pull it into your own public install. The dump could *probably* be imported to other directories, since it contains enough information, but we only write plugins for GLinks/Gossamer.

The NexusLinks is like the "Pro" version of the keyword plugin.

The keyword export/import plugin should be available soon. Right now, the export is available, and we can do keyword dumps for you. The interface is not pretty, but it's being cleaned up.

Using similar code to our bookmarks, we'll try to keep your searches, and counts, and use them to update you when new links meeting your criteria have arrived. It will also let you re-run searches later on.

Post questions/comments on the plugin's development or features here, but requests for dumps, or software should be made at http://ultranerds.com/forum.

PUGDOG´┐Ż Enterprises, Inc.

The best way to contact me is to NOT use Email.
Please leave a PM here.