Closed cscherrer closed 3 years ago
Merging #216 (6d6ff54) into master (97e49f2) will increase coverage by
2.45%
. The diff coverage is16.66%
.
@@ Coverage Diff @@
## master #216 +/- ##
==========================================
+ Coverage 28.53% 30.99% +2.45%
==========================================
Files 43 45 +2
Lines 1265 1381 +116
==========================================
+ Hits 361 428 +67
- Misses 904 953 +49
Impacted Files | Coverage Δ | |
---|---|---|
src/core/models/conditional.jl | 0.00% <0.00%> (ø) |
|
src/core/models/jointdistribution.jl | 21.42% <ø> (ø) |
|
src/core/models/model.jl | 77.17% <ø> (ø) |
|
src/core/models/abstractmodel.jl | 25.00% <25.00%> (ø) |
|
src/core/toposort.jl | 100.00% <100.00%> (ø) |
|
src/particles.jl | 14.54% <0.00%> (-1.25%) |
:arrow_down: |
src/core/dependencies.jl | 100.00% <0.00%> (ø) |
|
src/transforms/basictransforms.jl | 100.00% <0.00%> (ø) |
|
... and 16 more |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 97e49f2...6d6ff54. Read the comment docs.
WIP-- DO NOT MERGE
Currently we write, e.g.,
There's an implicit step here of creating the conditional distribution of
m(x=x)
givenobservations
, but we currently have no way to write this. Also, it would be useful to move closer to Turing's representation for sampling, which doesn't require observations in a separate argument.At the same time, we need to keep our generative representation. So this is a start on letting us write
Note that the
|
here is usually a bitwiseor
. But I don't see this being used for models, so the overloading feels safe for this usage.With this change, we now have
Model
,JointDistribution
, and alsoConditional
. Duplicating methods is starting to seem a bit much. So let's addAbstractModel
and define operations on that whenever possible.For now, we have
AbstractModel <: Distribution
, but we'll soon move this to<: MeasureTheory.Measure