Content-type: text/html Downes.ca ~ Stephen's Web ~ gRSShopper Graph API

Stephen Downes

Knowledge, Learning, Community

Jun 25, 2011

Just setting up the gRSShopper graph system


URL: http://www.downes.ca/cgi-bin/page.cgi?api=graph&apikey=[key]&cutoff=[unixdate]

  
    unixdate: cutoff is the unix date; the API will deliver results only greater than the cutoff value. Default is 0 (though I may change that :) )
    key: your API key. You have to get one from me.

Output is in OPML format.

Each graph edge is as follows:


  
  


The lines:

  
       title: Graph [edge_type] [edge_id]
           right now the only type is 'link' but that may change over time
       text = date the edge was created, un unix date

  
       Each node connected by the edge is indicated with a single ouline line
       text = [table][id] 
           table = gRSShopper internal table; values include link,post,media,feed,author, etc. this list is extensible
           id = gRSShopper ID number for that entity
       title = [link]
           link = gRSShopper internal table of the entity
       htmlURL =[url]
            url = the URL of the entity

     Note that while all entities will have a URL value, some entities may not have a table or ID, if they are not actually stored in the gRSShopper database

That's it. Let me know what you think.


Tony Hirst Jun 27, 2011 06:01 AM

Hi Stephen
Could you just clarify what the graph describes? As a first guess, I'd say the graph describes:
- nodes: pages you link to from OLDaily and pages those pages link to (with node annotations where you try to add semantics to the node; do you have a list of the annotations you currently support and intend to support?);
- edges: connect pages you link to from OLDaily with the pages those pages link to;

Or maybe you're doing everything you subscribe to and not just OLDaily linked to items?

Quick observation:
outline title="Graph link 409" text="1308944421" suggests it's a feed but only gives htmlURL - the XML/RSS url would also make sense here?

DownesJul 5, 2011 06:59 AM

Not sure what happened to my response, but, following up...

The graph describes relations between resources. There are different types of resources, eg., 'link', 'author', 'feed', etc.

Where possible, each resource is identified by a URL, and thus the graph in the first instance identifies links between URLs.

Where known, the type of link between URLs is identified. Right now the only type of link identified is 'link' but I will expand this.

As indicated, the URL's may represent different types of resources. Eg:

'link' - is a blog post or web site on the web
'post' - is one of my own blog posts or comments on my own site
'author' - is a person linked to a resource
'media' - is an image, presentation, audio or video resource

Each resource linked also has a unique identifier on my own system. This identifier is two-part, defining the type of resource and an ID number. Eg. 'Media 145' or 'Author 43'. Hence, the graph defines relations between entities in my system as well.


Stephen Downes Stephen Downes, Casselman, Canada
stephen@downes.ca

Copyright 2024
Last Updated: Dec 15, 2024 4:53 p.m.

Canadian Flag Creative Commons License.

Force:yes