Technical Updates
Presenter: Brian Duggan
US Global Change Research Program
ESIP Summer 2015
$ git rev-list --since=@{"one year ago"} HEAD --count
1216
https://github.com/USGCRP/gcis
https://travis-ci.org/USGCRP/gcis
https://coveralls.io/github/USGCRP/gcis
http://data-stage.globalchange.gov/export/gcis.txtSELECT sum(n_live_tup) FROM pg_stat_user_tables where schemaname='gcis_metadata'
39358
94209
select action, count(1) from audit.logged_actions where action_tstamp_tx > now() - interval '1 year' group by 1;
action | count --------+------- D | 7767 I | 62584 U | 67332 (3 rows)
http://data.globalchange.gov/sparqlSELECT (COUNT(*) AS ?triples) from <http://data.globalchange.gov> { ?s ?p ?o }
311407
131560
$ curl http://data.globalchange.gov/metrics.yaml
--
counts:
person: 8867 # 1141
dataset: 2377 # 23
article: 2150 # 2086
organization: 1204 # 845
report: 821 # 704
figure: 575 # 490
image: 579 # 274
instrument: 530 # 0
journal: 544 # 536
book: 226 # 166
platform: 282 # 0
model: 33 # 0
scenario: 3 # 0
$
http://data.globalchange.gov/metrics.yaml
“Putting datasets on the web ... is actually pretty easy. ...
Linking from one data set to another, so that both people and machines
understand that a resource in one data set is the same thing as a resource
in a second data set is far harder. Those explicit links can be made,
but it requires agreement, standards, and an awful lot of hard work.”
-- timbl, 2015-07-10
$ curl -vL http://data.globalchange.gov/person/4990.json
...
< HTTP/1.1 302 Found
< Location: /person/1024.json
> GET /person/1024.json HTTP/1.1
< HTTP/1.1 200 OK
{
"uri" : "/person/1024",
...
}