Conversation
Codecov Report
@@ Coverage Diff @@
## master #1373 +/- ##
==========================================
- Coverage 74.88% 74.88% -0.01%
==========================================
Files 69 69
Lines 5001 5000 -1
==========================================
- Hits 3745 3744 -1
Misses 1256 1256
Continue to review full report at Codecov.
|
| # build static lambda | ||
| unrollHypo! = if _slack === nothing | ||
| () -> cf( measurement_[smpid]..., (getindex.(varParams, smpid))... ) | ||
| () -> cf( measurement_[smpid], (getindex.(varParams, smpid))... ) |
There was a problem hiding this comment.
Why drop the tuple part of a measurement here? Code all over has residual functions follow
(cf)(z1,z2,x1,x2,x3)
seems like you restricting to only z1?
There was a problem hiding this comment.
I take it the user must decompose the measurement sample data themselves inside the residual function, z1 here can be anything.
There was a problem hiding this comment.
Why drop the tuple part of a measurement here? Code all over has residual functions follow
(cf)(z1,z2,x1,x2,x3)
Only found DERelative that works this way and updated it.
I take it the user must decompose the measurement sample data themselves inside the residual function, z1 here can be anything.
yes, "one factor = one measurement". Measurement can be any type and it's up to the user to decide how to compose it. This way it's more flexible and we get rid of the tuples for simpler cases, this makes maintenance easier in my opinion.
close #1371