Closed aandreev0 closed 6 years ago
That exists and I can give you code that will get this using the standard APIs. Let’s talk next week
Sent from my iPhone
On Aug 17, 2018, at 7:04 PM, aandreev notifications@github.com<mailto:notifications@github.com> wrote:
Right now i don't have API access to type of experiment in cohort. That is, Bill sends me file that looks like:
Learners: Dsy20160116A - 6025 - 3 (stars) Zdu20170516A - 6483 - 3 Dsy20170316B - 6493 - 4 Dsy20161202B - 6791 - 3 Dsy20170316A - 7949 - 4 Dsy20171019B - JK6 - 4 Dsy20170426A - R5Y - 3* Dsy20180404B - ZHT - 4 Zdu20180404A - 11YG - 3 Zdu20180404B - 11YR - 4
Nonlearners: Dsy20170329C - 6479 - 4 Dsy20171019A - JKE - 4 Dsy20170202A - 7965 - 4 Dsy20170412A - 6481 - 4 Zdu20171114B - JJE - 3 Zdu20171020A - JJY - 4 Zdu20171031A - JJP - 3 Zdu20171121A - JJ6 - 4 Zdu20180403A - 11Z0 - 3 Dsy20180404A - 11Z8 - 3
I would like to be able to pull type of experiment from storage somehow using fish ID (not sure correct terminology) like "Dsy20160116A".
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHubhttps://github.com/informatics-isi-edu/synspy/issues/66, or mute the threadhttps://github.com/notifications/unsubscribe-auth/ADbjXrwGOPpMhtd3Ix2K3YBJehF6B8wbks5uR3YRgaJpZM4WCYqW.
Did you guys ever talk about the access APIs used in the jupyter notebooks? I'm closing this with a few extra comments. Please re-open if you still need help...
At the model level, there is a boolean column on the Behavior
table called Learned?
which should be set to true
or false
for these runs. Using filters on this, it is easy to query for a list of just learners or just non-learners. I would expect this to remain null
for the various control groups which experience a variant Behavior experiment (e.g. lacking CS and/or US) and those will also be associated with a different protocol.
In general, we don't directly associate a fish with a protocol. We associate it indirectly by linking individual observations (e.g. Image
or Behavior
records) with protocol steps (e.g. tpt1
, tpt2
, or behave
) which are themselves children of a specific protocol. In general, I think an individual subject will only get associated with one protocol in this manner. But, the catalog model doesn't restrict this, so a messy longitudinal record could mix observations from several protocols...
Also, I don't think Bill's "stars" QA concept is captured in the catalog at present.
Right now i don't have API access to type of experiment in cohort. That is, Bill sends me file that looks like:
I would like to be able to pull type of experiment from storage somehow using fish ID (not sure correct terminology) like "Dsy20160116A".