Hall of Mirrors
WARNING: THIS IS NOT READY FOR REAL LIFE USAGE UNLESS YOU DON'T MIND USING
HALF-WRITTEN IN-PROGRESS THINGS THAT MAY POSSIBLY CREATE BREAKING CHANGES IN THE
FUTURE.
Build your own corner of Versailles; with your data. You remember data, right?
That stuff you keep putting on other company's sites. Assuming that they won't
change. They won't break.
These tools will (eventually) allow you to pull all the relevant data you have
put on a site, and then serve a version of that data as a nice website. These do
not attempt to fully recreate the site, but the data has been pulled and if you
play with the templates enough then you could.
Each folder contains three tools,
Before doing anything you need to run [service]/init. This will get you to
authenticate the tool (usually by copying a code from a url in), then any
preparation that needs to occur. This only needs to be run once.
Now you're logged in you can run [service]/update. This pulls all the new
data down. It has a pretty decent default, but you may want to run it with
--help to see what options are available.
Once you have you're data you can see it with [service]/serve. This spins up a
local version of the site. It isn't a one-one mirror, but should show enough to
be useful.
State of the hall
Note: percentages are fairly arbitrary.
facebook [80%]
Pulls your photos, photos you've been tagged in and the rest of the album those
photos belong too. Needs to get friends' profile picture(s?) to display
also. And some things should be better about the served site, such as next/prev
for photo pages, and maybe add maps.
If you really want to play with this:
- Create an app on Facebook
Export your App ID and App secret in your .bashrc (or similar) like:
export FACEBOOK_APP_ID='xxxxxx...'
export FACEBOOK_APP_SECRET='xxxxxxx...'
- Run
facebook/init and follow the instructions
- Run
facebook/update and wait for it to pull all the photos
- Run
facebook/serve and open the address (usually http://localhost:4567)
flickr [75%]
Pulls your photos. Need to work on getting faves also. Needs to have a page to
select user. And if the URL could remember whether you entered a path alias,
that would be nice. Speed-ups would be good. Oh, and pretty up the tag list
page, and add similar tags to each tag page. Special navigation for machine tags
maybe? in a tree? Oh yes and allow wildcard machine tag searches. And pretty up
set list pages.
If you really want to play with this:
- (Create an app)[http://www.flickr.com/services/apps/create/] on flickr to get
a key/secret combo
- Export them as environment variables in your
.bashrc (or similar) like:
bash
export FLICKR_KEY='xxxxxxxxxxxxxxxx...'
export FLICKR_SECRET='xxxxxxxxxxxxxxxxxx...'
Run flickr/init, and follow the instructions
Run flickr/update and do something else while it pulls your photos
Rune flickr/serve and open the address (usually http://localhost:4567)
github [0%]
last.fm [0%]
pinboard [0%]
tumblr [80%]
Pulls your data down, but need to work on getting faves and drafts as well.
Apart for that it works pretty well. It pulls images, audio and videos (from
tumblr, youtube and vimeo) and serves them with html5 elements. Warning:
this can make the backups quite large, especially if you have a habit of posting
long youtube videos (for reference my rather modest tumblr weighs in at 2.18 GB).
To use:
- (Register an application)[http://www.tumblr.com/oauth/apps]
- Export the OAuth Consumer Key and Secret Key, like:
bash
export TUMBLR_CONSUMER_KEY='xxxxxxx...'
export TUMBLR_CONSUMER_SECRET='xxxxxxxxxx...'
- Run flickr/init and follow the instructions, it hacks into the auth flow of mwunsch/tumblr so isn't as smooth as the others
- Run flickr/update and wait as it will download LOTS!
- Run flickr/serve and open the address (usually http://localhost:4567)
### twitter [0%]