Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

More verbose naming #95

Merged
merged 7 commits into from
Jan 5, 2021
Merged

More verbose naming #95

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
af9d589
Renamed src,test,docs. Not yet in examples
odunbar Dec 23, 2020
c44ddf5
Merge branch 'master' into orad/verbose-names
odunbar Dec 23, 2020
36d880b
extract -> get
odunbar Dec 23, 2020
0fef9ea
examples
odunbar Dec 23, 2020
7ffdd9b
Update EKP test including find_ekp_stepsize test.
ilopezgp Jan 5, 2021
06b99a4
removed def of plot flag - done in test/runtests
odunbar Jan 5, 2021
0b6b0e1
add Utility tests
odunbar Jan 5, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions src/EnsembleKalmanProcesses.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -112,11 +112,11 @@ end


"""
find_ekp_stepsize(ekp::EnsembleKalmanProcess{FT, IT, Inversion}, g::Array{FT, 2}; cov_threshold::FT=0.01) where {FT}
find_ekp_stepsize(ekp::EnsembleKalmanProcess{FT, IT, Inversion}, g::AbstractArray{FT, 2}; cov_threshold::FT=0.01) where {FT}
Find largest stepsize for the EK solver that leads to a reduction of the determinant of the sample
covariance matrix no greater than cov_threshold.
"""
function find_ekp_stepsize(ekp::EnsembleKalmanProcess{FT, IT, Inversion}, g::Array{FT, 2}; cov_threshold::FT=0.01) where {FT, IT}
function find_ekp_stepsize(ekp::EnsembleKalmanProcess{FT, IT, Inversion}, g::AbstractArray{FT, 2}; cov_threshold::FT=0.01) where {FT, IT}
accept_stepsize = false
if !isempty(ekp.Δt)
Δt = deepcopy(ekp.Δt[end])
Expand Down
6 changes: 6 additions & 0 deletions test/EnsembleKalmanProcesses/runtests.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,6 +62,11 @@ using CalibrateEmulateSample.ParameterDistributionStorage
ekiobj = EnsembleKalmanProcesses.EnsembleKalmanProcess(initial_ensemble,
y_obs, Γy, Inversion())

# Find EKI timestep
g_ens = G(ekiobj.u[end]')'
Δ = find_ekp_stepsize(ekiobj, g_ens)
@test Δ ≈ 0.0625

# EKI iterations
for i in 1:N_iter
params_i = ekiobj.u[end]
Expand All @@ -77,6 +82,7 @@ using CalibrateEmulateSample.ParameterDistributionStorage
# Plot evolution of the EKI particles
eki_final_result = vec(mean(ekiobj.u[end], dims=1))

TEST_PLOT_OUTPUT = false
Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

When running test/EnsembleKalmanProcesses/runtests.jl, it seems like TEST_PLOT_OUTPUT is not defined. Can you verify this behavior @odunbar ?

In theory we should run the tests as a block i.e from test/runtests.jl and so this is where TEST_PLOT_OUTPUT is defined (via a get method with 0 default if the CES_TEST_PLOT_OUTPUT variable is not defined). If you directly run the EKP block from test/EnsembleKalmanProcesses/runtests.jl you would need to define it.

So i would remove the line you've added here. Otherwise these run for me! Thanks!

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Edit: I've removed this in latest commit.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Great, thanks!

if TEST_PLOT_OUTPUT
gr()
p = plot(ekiobj.u[1][:,1], ekiobj.u[1][:,2], seriestype=:scatter)
Expand Down