Molecular Hamiltonian #327
Conversation
Benchmark resultJudge resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsA ratio greater than
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoTargetBaselineTarget resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoBaseline resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoRuntime information
Architecture: x86_64
Benchmark resultJudge resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsA ratio greater than
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoTargetBaselineTarget resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoBaseline resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoRuntime information
Architecture: x86_64
Benchmark resultJudge resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsA ratio greater than
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoTargetBaselineTarget resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoBaseline resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoRuntime information
Architecture: x86_64
Benchmark resultJudge resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsA ratio greater than
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoTargetBaselineTarget resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoBaseline resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoRuntime information
Architecture: x86_64
Benchmark resultJudge resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsA ratio greater than
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoTargetBaselineTarget resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoBaseline resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoRuntime information
Architecture: x86_64
Benchmark resultJudge resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsA ratio greater than
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoTargetBaselineTarget resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoBaseline resultBenchmark Report for /home/runner/work/Rimu.jl/Rimu.jlJob Properties
ResultsBelow is a table of this job's results, obtained by running the benchmarks.
Benchmark Group ListHere's a list of all the benchmark groups executed by this job:
Julia versioninfoRuntime information
Architecture: x86_64 |
lch
force-pushed
the
feature/molecular-hamiltonian
branch
from
bb044f3 to
257ae36
Compare
Pull Request Test Coverage Report for Build 17467385451Details
💛 - Coveralls |
|
Had a quick look at the failing tests: This might be the same issue @mtsch saw recently and fixed on the |
lch
force-pushed
the
feature/molecular-hamiltonian
branch
from
508227f to
36c3b09
Compare
lch
force-pushed
the
feature/molecular-hamiltonian
branch
2 times, most recently
from
6e31c0e to
af50074
Compare
lch
force-pushed
the
feature/molecular-hamiltonian
branch
from
981aa3c to
0d76a6c
Compare
lch
force-pushed
the
feature/molecular-hamiltonian
branch
from
0d76a6c to
3bedfec
Compare
lch
force-pushed
the
feature/molecular-hamiltonian
branch
3 times, most recently
from
424a9e1 to
af7ed0e
Compare
lch
force-pushed
the
feature/molecular-hamiltonian
branch
2 times, most recently
from
d3f04a5 to
ee3cb1c
Compare
lch
force-pushed
the
feature/molecular-hamiltonian
branch
from
f11f0d7 to
382866d
Compare
There was a problem hiding this comment.
It is nice to see that you got iteration working without allocations. This is great progress!
I appreciate the comments and docstrings and improved readability of the iteration code. I think that the code can still be simplified, which can also make the code more readable and faster.
I'm not quite sure whether iteration works as it should, as repeatedly 'void' states are being returned (and I thought they shouldn't):
julia> while result != nothing
elem, state = result
println(state)
result = iterate(od, state)
end
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (1, 0), (2, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (1, 0), (3, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (1, 0), (4, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (1, 0), (5, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (1, 0), (6, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (2, 0), (1, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (2, 0), (2, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (2, 0), (3, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (2, 0), (4, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (2, 0), (5, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (2, 0), (6, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 1), (0, 0), (0, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (1, 0), (2, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (1, 0), (3, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (1, 0), (4, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (1, 0), (5, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (1, 0), (6, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (2, 0), (1, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (2, 0), (2, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (2, 0), (3, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (2, 0), (4, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (2, 0), (5, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (2, 0), (6, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((1, 0), (0, 0), (0, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (1, 3))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (1, 4))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (1, 5))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (1, 6))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (2, 3))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (2, 4))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (2, 5))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (2, 6))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (3, 4))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (3, 5))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (3, 6))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (4, 5))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (4, 6))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (1, 2), (5, 6))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((0, 2), (0, 0), (0, 0))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((2, 0), (1, 2), (1, 3))
Rimu.Hamiltonians.MolecularHamiltonianOffDiagonalsIteratorState((2, 0), (1, 2), (1, 4))
...This was for the h2_dual Hamiltonian and offdiagonals from the starting address.
lch
force-pushed
the
feature/molecular-hamiltonian
branch
from
b485cab to
61e1337
Compare
Add initial skeleton implementation for MolecularHamiltonian type
There was a problem hiding this comment.
A few more comments from me.
It would be good to have the unexported docstrings show up in Documenter as well.
Finally, please check the Codecoverage report, as parts of the Molecular.jl code are not tested currently. Please either add tests, or delete the code if it is obsolete!
| # ```julia | ||
| # s = solve(p) | ||
| # ``` |
There was a problem hiding this comment.
Maybe we should warn here that this may require larger computational resources.
I'm actually not sure what exactly happens and why this seems to stall. solve(p; full_basis=true) seems a good idea, as initialising the LinearOperator is done much faster (12s on my laptop), but the matrix-vector multiplications seem to stall. We can deal with this later, though, as I don't think it is an issue related to this PR.
| df = DataFrame(result); | ||
|
|
||
| # To analyse the energy shift, we can use [`shift_estimator`](@ref). | ||
| se = shift_estimator(df; skip=steps_equilibrate) No newline at end of file |
There was a problem hiding this comment.
Add a bit of interpretation here, i.e. that the mean is an estimator for the energy. Also explain that and why you used the initiator and what this means for the energy estimate.
1. Improve the document. 2. Remove obsolete code.
lch
force-pushed
the
feature/molecular-hamiltonian
branch
from
7a9206f to
072e1f6
Compare
lch
force-pushed
the
feature/molecular-hamiltonian
branch
from
072e1f6 to
477499e
Compare
Make ElemCo an optional dependency through a package extension.
| function MolecularHamiltonian( | ||
| fcidump_path::String, | ||
| starting_address::Union{Nothing,FermiFS2C}=nothing, | ||
| specifier::String="", | ||
| ) |
There was a problem hiding this comment.
The string that is printed at the REPL currently does not parse as a valid constructor:
julia> h = MolecularHamiltonian(fcidump)
MolecularHamiltonian("/Users/brand/git/code/Rimu/test/examples/h2o.FCIDUMP", starting_addresss=fs"|⇅⇅⇅⇅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⟩")
julia> h == MolecularHamiltonian("/Users/brand/git/code/Rimu/test/examples/h2o.FCIDUMP", starting_addresss=fs"|⇅⇅⇅⇅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⋅⟩")
ERROR: MethodError: no method matching MolecularHamiltonian(::String; starting_addresss::FermiFS2C{4, 4, 23, 8, FermiFS{…}, FermiFS{…}})
This method may not support any kwargs.
Closest candidates are:
MolecularHamiltonian(::String) got unsupported keyword argument "starting_addresss"
@ Rimu ~/git/code/Rimu/src/Hamiltonians/Molecular.jl:46
MolecularHamiltonian(::String, ::Union{Nothing, FermiFS2C{N1, N2, M} where {N1, N2, M}}) got unsupported keyword argument "starting_addresss"
@ Rimu ~/git/code/Rimu/src/Hamiltonians/Molecular.jl:46
MolecularHamiltonian(::String, ::Union{Nothing, FermiFS2C{N1, N2, M} where {N1, N2, M}}, ::String) got unsupported keyword argument "starting_addresss"
@ Rimu ~/git/code/Rimu/src/Hamiltonians/Molecular.jl:46
...This can be fixed by making the optional arguments keyword arguments:
| function MolecularHamiltonian( | |
| fcidump_path::String, | |
| starting_address::Union{Nothing,FermiFS2C}=nothing, | |
| specifier::String="", | |
| ) | |
| function MolecularHamiltonian( | |
| fcidump_path::String; | |
| starting_address::Union{Nothing,FermiFS2C}=nothing, | |
| specifier::String="", | |
| ) |
If you make this change, please update the docstring such that it is accurate!
I'd also like to see this tested, e.g. by adding something like this to the tests:
@test eval(Meta.parse(repr(H))) == HObviously, this only works when constructing from a path, and when the path is still available and accurate.
Update: I notice that for this to work you may need to define a method for isequal for the TFDump struct. This can be done in the package extension (or ideally as a PR to ElemCo.jl, as otherwise it would be type piracy 😄) .
There was a problem hiding this comment.
1. Improve `MolecularHamiltonian` constructor. 2. Rename functions calculating diagonal term. 3. Fix document display error. 4. Improve imath equation for clarity under REPL. 5. Improve method `unrank_combination` document.
1. Fix the test of `MolecularHamiltonian`. 2. Fix the typo in documents.
Improve the docstring of `MolecularHamiltonian`
New features
This PR addresses issue #321 by introducing the
MolecularHamiltoniantype along with its related types and methods. These additions enableRimu.jlto be used for post–Hartree-Fock calculations in quantum chemistry research accompanied withElemCo.jl.A detailed description to the PR could be find at #348 .