brendanheywood / climbing-api Goto Github PK

Make a perl script that takes raw text from legacy guides and use 
heuristics to separate and mark up as much data as possible.

eg:

1) Some title    22m 22 **
Start: up the massive nose. Thin crack to ledge.
Joe Blogs, Fred Smith 2/78

+ some config:
Map 1 star to 1, 2 stars to 3 and 3 stars to 5
Default grade is Ewbank and V for boulder.

turns into

<entry>
  <title>Some climb</title>
  <dc:type>Ewbank</dc:type>
  <gd:rating min="1" max="5" rating>
  <dc:format>22m 22</dc:format>
  <summary>Start: up the massive nose. Thin crack to ledge.</summary>
  <creator>Joe Blogs</creator>
  <creator>Fred Smith</creator>
  <dc:date>1978 Feb</dc:date>
</entry>

It will handle all the atom calls for initializing a local db, updating it, 
creating a new local route, syncing it with the server, deleting local 
entries as needed etc. Will work with or without a local in memory DB.

It will then be used by both the website and the mobile web version of the 
client reference implementation.

The area's in the db could be represented using the 'envelope' in GML

http://www.georss.org/gml

Add tags to flickr photos:

grade:V2
height:11m

the usual lat long
alt:900m

Put all your photos into a set, and then make a script that converts that 
set into a climb feed.

Issues: how to model more than one route per photo?

It will expect columns to be named certain things.