Closed mhauru closed 4 years ago
Looks good.
Indeed, the individual tensors don't need to simple symbols or calls to conj
. E.g., if you have 4 tensors in a list tensors
, its perfectly fine to do things like. As long as @ncon
just knowns the number of objects, and which need to be conjugated, that's fine. But I agree my implementation is not very fail-proof!
@ncon([tensors[1], conj(tensors[2]), tensors[3][:,1,:], view(tensors[4], :, 1:2:end)], ...)
Merging #80 into master will increase coverage by
1.1%
. The diff coverage is50%
.
@@ Coverage Diff @@
## master #80 +/- ##
=========================================
+ Coverage 71.91% 73.01% +1.1%
=========================================
Files 22 22
Lines 1766 1827 +61
=========================================
+ Hits 1270 1334 +64
+ Misses 496 493 -3
Impacted Files | Coverage Δ | |
---|---|---|
src/indexnotation/tensormacros.jl | 58.33% <50%> (-0.29%) |
:arrow_down: |
src/indexnotation/preprocessors.jl | 57.37% <0%> (-0.96%) |
:arrow_down: |
src/implementation/diagonal.jl | 69.29% <0%> (-0.9%) |
:arrow_down: |
src/functions/ncon.jl | 100% <0%> (ø) |
:arrow_up: |
src/functions/simple.jl | 100% <0%> (ø) |
:arrow_up: |
src/indexnotation/verifiers.jl | 82.79% <0%> (+0.37%) |
:arrow_up: |
src/indexnotation/instantiators.jl | 86.92% <0%> (+0.76%) |
:arrow_up: |
src/indexnotation/parser.jl | 86.25% <0%> (+1.25%) |
:arrow_up: |
src/implementation/stridedarray.jl | 83.84% <0%> (+2.11%) |
:arrow_up: |
src/indexnotation/optimaltree.jl | 93.65% <0%> (+3.04%) |
:arrow_up: |
... and 1 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 4cc4a06...0ff4365. Read the comment docs.
Resolves https://github.com/Jutho/TensorOperations.jl/issues/79
I considered, instead of just checking that there aren't splats in
tensorlist
, enforcing that all elements should be symbols or calls toconj
. I didn't dare do that for fear of excluding some use case I hadn't thought of.