The initial fix can be found in the branch.

`cobra_target` (previously `cobra_movement`) includes cobra targeting values such as the target positions and step size of each cobra in each iteration. `cobra_matching` (previously `cobra_status`) covers the matched position of spots on the MCS image. I just revert onto `pfi_nominal_[xy]_mm` as the final cobra destination, which is consistent with the datamodel.

Anyway, any feedback regarding structure, naming, additional columns is all welcome. chyan karr cloomis hassan ?

Attached below is the zoom-in diagram for the related tables:

opdb_cobra_tables_rev1.pdf

Some thoughts:

• having cobra_target key on mcs_frame_id bothers me: the cobra_target row will be inserted before the mcs exposure is taken. And the cobra_target table is filled by a completely separate actor (fps) from the mcs_exposure table (mcs).
• should cobra_matching link to cobra_target and drop the "nominal" fields?
• the cobra_matching table needs a flags field.
• the cobra_target table should have target angles as well as (x, y) and (step, ontime).
• should there be two motor_calib_ids and should they actually be cobra_motor_model_ids?

• I think I would rename "cobra_matching" to "cobra_match", maybe?

This is the second revision, which is a hasty fix so may be something wrong. Anyway, let's talk about this at the phonecon tomorrow.

opdb_cobra_tables_rev2.pdf