WikiGraph With FreeLinks

Will have to extend WikiGraph to support WikiEngines with FreeLinks, because that's what almost all spaces have.

2016 work

Basic concept:

  • For pages table:
  • add url field for the (Camel-Case for-me) page-bit of the URL (don't include the path part)
  • make name lower-case, so effectively get case-insensitive matching
  • (do I also want a clean separated version of this value for a title tag that would be nicely word-searchable, to make more general SearchEngine? Nah, can always just do regex.)
  • now when widget calls the function
  • have to convert Referer to SmashedTogetherWords format if it's not already
  • then make sure using the url value in the return, not the name

Adoption plan

  • ask key engine-runners: Ward Cunningham and Mike Caulfield
  • do their engines have an AllPages page to scrape?
  • do they have a directory of spaces? naming convention per space (with uniqueness)?
  • wikity
  • has all-spaces FrontPage which you can crawl through Prev pages
  • smells like it doesn't enforce name, but can generate from hostname+domain (drop the cc)
  • scrape some data
  • build (update) system
  • start using on WebSeitz
  • then let them call TwinPages widget
  • then SearchEngine
  • then copy from SearchEngine to local space
  • see MikeCaulfield BootstrappingLibrary/ColdStart post
  • store copied-from URL?
  • allow copy to different name?
  • hmm does my WikiEngine need Edit button for guests, so they can copy?

Mar07'2016: realize http://wikity.cc/ covers all Wikity sites, goes back 18 pages. So I can grab each, and regex out the links.

Mar08'2016: write parser for (1st page of) links

Mar09: create old db in local postgres, import old records. (Was in SQLite before.)

Mar09: tweak Wikity parser code to make pages.url lower-case

Mar09: scrape all wikity pages, parse (1766 urls, 75 spaces)

Mar09: get old/current wikigraph db/app working

  • WikiGraph/WebPy won't talk to psql, realize none of my old WebPy apps ever used it, always used SQLite
  • Launch WikiGraph using SQLite, tweak WikiWeb footer to call it. It works!
  • change db structure, fill past fields, import couple records
  • adjust code to use new db

Mar10: map Referer to space to support wikitys and not link to self

Mar11: adjust import code for extra field, and for spaces table

Mar11: native search form - WikiSearch

Mar12: deploy

  • ugh bug when spaces.page_pattern includes a path, like for me! Hacked for me, so working. But need to fix.

Mar12: tell MikeCaulfield

Mar12: add custom SearchEngine! https://cse.google.com/cse/all

Mar14: add Phil Jones SFW space after he emails me list. Need to fix bug from Mar12 for him to use the widget too

Later: handle SFW sites

  • scrape Google? or use Ward's scraper code?

Edited:    |       |    Search Twitter for discussion