erli

NOTE: erli is being finalized, but still far from feature-complete. ETA: soon ;)

erli is an Erlang based URL shortener.

It started as a more involved experiment with REST and webmachine as an abstraction of it. erli comes with custom URLs, simple statistics for each shortened URL and optional landing pages (that include screenshots of the target website).

Status

erli is currently still undergoing a rewrite. Remaining TODOs:

• Review and optimize test quality and coverage
• Test suite(s) for thumbnail generation related components
• A proper release

To run the test suite, call ct_run -logdir $PWD/ct/logs -dir $PWD/ct/ -include $PWD/include/ -pa $PWD/ebin -pa $PWD/deps/*/ebin -erl_args -config $PWD/erli.config from the erli root dir.

The backend

The backend is a CRUD API written in webmachine using mnesia as storage backend. Its root URL is at /api/. The API is designed to be self-explanatory and -discoverable. However, here is a quick overview:

API Endpoints and Applicable Methods

/api/paths                          GET, HEAD, OPTIONS, POST
/api/paths/{PATH_ID}                GET, HEAD, OPTIONS, DELETE
/api/paths/{PATH_ID}/targets        GET, HEAD, OPTIONS
/api/paths/{PATH_ID}/visits         GET, HEAD, OPTIONS
/api/targets                        GET, HEAD, OPTIONS, POST
/api/targets/{TARGET_ID}            GET, HEAD, OPTIONS, DELETE
/api/targets/{TARGET_ID}/paths      GET, HEAD, OPTIONS
/api/targets/{TARGET_ID}/visits     GET, HEAD, OPTIONS
/api/visits                         GET, HEAD, OPTIONS
/api/visits/{VISIT_ID}              GET, HEAD, OPTIONS

Interna

For collection types pagination is handled via the Range header. The data returned by the server is intended to suffice for client driven exploration.

Purging of unwanted or infringing content is community driven in the form of infringement flagging. Every DELETE request towards a path or target resource counts towards a configurable limit (15 by default). The limit is per target, since multiple paths may point to the same target. Once the limit is reached all paths pointing to said target are marked as banned, as is the target itself. All requests to banned resources are answered with a 410 GONE. Note that the current implementation is "dumb". To avoid storing user-related data on the server there is no way to prevent users from abusing the flagging process.