Galene provides an HTTP-based API that can be used to create groups and users. For example, in order to create a group, a client may do
PUT /galene-api/v0/.groups/groupname/
Content-Type: application/json
If-None-Match: *
The If-None-Match
header avoids overwriting an existing
group.
In order to edit a group definition, a client first does
GET /galene-api/v0/.groups/groupname/
This yields the group definition and an entity tag (in the ETag header). The client then modifies the group defintion, and does
PUT /galene-api/v0/.groups/groupname/
If-Match: "abcd"
where “abcd” is the entity tag returned by the GET request. If the group definition has changed in the meantime, the entity tag will no longer be valid, and the server will fail the update, which avoids losing an update in the case of a concurrent modification.
The API is located under /galene-api/v0/
. The
/v0/
is a version number, and will be incremented if we
ever find out that the current API cannot be extended in a backwards
compatible manner.
/galene-api/v0/.stats
Provides a number of statistics about the running server, in JSON. The exact format is undocumented, and may change between versions. The only allowed methods are HEAD and GET.
/galene-api/v0/.groups/
Returns a list of groups, as a JSON array. The only allowed methods are HEAD and GET.
/galene-api/v0/.groups/groupname
Contains a “sanitised” group definition in JSON format, analogous to
the on-disk format but without any user definitions or cryptographic
keys. Allowed methods are HEAD, GET, PUT and DELETE. The only accepted
content-type is application/json
.
/galene-api/v0/.groups/groupname/.keys
Contains the keys used for validation of stateless tokens, encoded as
a JSON key set (RFC 7517). Allowed methods are PUT and DELETE. The only
accepted content-type is application/jwk-set+json
.
/galene-api/v0/.groups/groupname/.users/
Returns a list of users, as a JSON array. The only allowed methods are HEAD and GET.
/galene-api/v0/.groups/groupname/.users/username
/galene-api/v0/.groups/groupname/.empty-user
/galene-api/v0/.groups/groupname/.wildcard-user
Contains a “sanitised” user definition (without any passwords), a
JSON object with a single field permissions
. The entries
.empty-user
and .wildcard-user
are for the
user with the empty username and the wildcard user respectively. Allowed
methods are HEAD, GET, PUT and DELETE. The only accepted content-type is
application/json
.
/galene-api/v0/.groups/groupname/.users/username/.password
/galene-api/v0/.groups/groupname/.empty-user/.password
/galene-api/v0/.groups/groupname/.wildcard-user/.password
Contains the password of a given user. The PUT method takes a full
password definition, identical to what can appear in the
"password"
field of the on-disk format, while the POST
method takes a string which will be hashed on the server. Allowed
methods are PUT, POST and DELETE. Accepted content-types are
application/json
for PUT and text/plain
for
POST.
/galene-api/v0/.groups/groupname/.wildcard-user
Contains a dictionary defining the wildcard user, in the same format as the dictionary defining an ordinary user. Allowed methods are HEAD, GET, PUT and DELETE.
/galene-api/v0/.groups/groupname/.wildcard-user/.password
This is analogous to the password of an ordinary user. Allowed methods are PUT, POST and DELETE.
/galene-api/v0/.groups/groupname/.users/username/.tokens/
GET returns the list of stateful tokens, as a JSON array. POST
creates a new token, and returns its name in the Location
header. Allowed methods are HEAD, GET and POST.
/galene-api/v0/.groups/groupname/.users/username/.tokens/token
The full contents of a single token, in JSON. The exact format may change between versions, so a client should first GET a token, update one or more fields, then PUT the resulting token. Allowed methods are HEAD, GET and PUT.