Skip to content

Commit

Permalink
Fixes in tuning, Pkg updates
Browse files Browse the repository at this point in the history
  • Loading branch information
@graidl
graidl committed Nov 19, 2024
1 parent 7cf2a81 commit 5b55b7d
Show file tree
Hide file tree
Showing 13 changed files with 42 additions and 27 deletions.
3 changes: 3 additions & 0 deletions CHANGELOG.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,9 @@

Major changes in releases:

## Version 0.1.15
- `Tuning`renamed to `tuning`and small fixes in it, updates for Julia 1.11.

## Version 0.1.14
- `Results` from `Scheduler` is extended with field `is_local_optimum` to be able to
indicate if the solution is a local optimum in respect to the current method and it
Expand Down
10 changes: 5 additions & 5 deletions MHLibDemos/Manifest.toml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -99,9 +99,9 @@ version = "0.2.2"

[[deps.JuliaInterpreter]]
deps = ["CodeTracking", "InteractiveUtils", "Random", "UUIDs"]
git-tree-sha1 = "2984284a8abcfcc4784d95a9e2ea4e352dd8ede7"
git-tree-sha1 = "fc8504eca188aaae4345649ca6105806bc584b70"
uuid = "aa1ae85d-cabe-5617-a682-6adf51b2e16a"
version = "0.9.36"
version = "0.9.37"

[[deps.LibGit2]]
deps = ["Base64", "LibGit2_jll", "NetworkOptions", "Printf", "SHA"]
Expand Down Expand Up @@ -157,7 +157,7 @@ version = "3.0.5"
deps = ["ArgParse", "DataStructures", "Logging", "Printf", "Random", "Reexport", "StatsBase"]
path = ".."
uuid = "0f3a0e6e-e0e7-4894-bd64-856f0ae3b635"
version = "0.1.13"
version = "0.1.15"

[[deps.MacroTools]]
deps = ["Markdown", "Random"]
Expand Down Expand Up @@ -239,9 +239,9 @@ version = "1.3.0"

[[deps.Revise]]
deps = ["CodeTracking", "Distributed", "FileWatching", "JuliaInterpreter", "LibGit2", "LoweredCodeUtils", "OrderedCollections", "REPL", "Requires", "UUIDs", "Unicode"]
git-tree-sha1 = "7f4228017b83c66bd6aa4fddeb170ce487e53bc7"
git-tree-sha1 = "834aedb1369919a7b2026d7e04c2d49a311d26f4"
uuid = "295af30f-e4ad-537b-8983-00126c2a3abe"
version = "3.6.2"
version = "3.6.3"

[[deps.SHA]]
uuid = "ea8e919c-243c-51af-8825-aaa63cd721ce"
Expand Down
2 changes: 1 addition & 1 deletion Manifest.toml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2,7 +2,7 @@

julia_version = "1.11.1"
manifest_format = "2.0"
project_hash = "b8b3564b02bd77ab3d1613a78103290e33d669ad"
project_hash = "8a01778624db31048bc446d5aac140b2a907499f"

[[deps.ArgParse]]
deps = ["Logging", "TextWrap"]
Expand Down
4 changes: 2 additions & 2 deletions Project.toml
View file
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
name = "MHLib"
uuid = "0f3a0e6e-e0e7-4894-bd64-856f0ae3b635"
authors = ["Guenther Raidl <[email protected]> and others"]
version = "0.1.14"
version = "0.1.15"

[deps]
ArgParse = "c7e460c6-2fb9-53a9-8c5b-16f535851c63"
Expand All @@ -20,4 +20,4 @@ Reexport = "1.2"
Logging = "1"
Printf = "1"
Random = "1"
julia = "1.10"
julia = "1.11"
8 changes: 6 additions & 2 deletions README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -44,7 +44,8 @@ installed via

## Major Components

Note that `MHLib.jl` is still behind the capabilities of the Python `pymhlib`, however, much more performant.
Note that `MHLib.jl` is still behind the capabilities of the Python `pymhlib`,
however, much more performant.

The main module provides the following types for candidate solutions and various
functions for them:
Expand Down Expand Up @@ -91,7 +92,10 @@ Further modules:

## Demos

For demonstration purposes subdirectory [`MHLibDemos`](MHLibDemos/README.md) provides a package (not officially registered at JuliaHub), with basic implementations for the following classical combinatorial optimization problems, to which some of MHLib's metaheuristics are applied:
For demonstration purposes subdirectory [`MHLibDemos`](MHLibDemos/README.md) provides
a package (not officially registered at JuliaHub), with basic implementations for the
following classical combinatorial optimization problems, to which some of
MHLib's metaheuristics are applied:

- `OneMax`: basic test problem in which the goal is to set all digits in a binary
string to `true`
Expand Down
8 changes: 4 additions & 4 deletions test/Manifest.toml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -108,9 +108,9 @@ version = "0.2.2"

[[deps.JuliaInterpreter]]
deps = ["CodeTracking", "InteractiveUtils", "Random", "UUIDs"]
git-tree-sha1 = "2984284a8abcfcc4784d95a9e2ea4e352dd8ede7"
git-tree-sha1 = "fc8504eca188aaae4345649ca6105806bc584b70"
uuid = "aa1ae85d-cabe-5617-a682-6adf51b2e16a"
version = "0.9.36"
version = "0.9.37"

[[deps.LibCURL]]
deps = ["LibCURL_jll", "MozillaCACerts_jll"]
Expand Down Expand Up @@ -260,9 +260,9 @@ version = "1.3.0"

[[deps.Revise]]
deps = ["CodeTracking", "Distributed", "FileWatching", "JuliaInterpreter", "LibGit2", "LoweredCodeUtils", "OrderedCollections", "REPL", "Requires", "UUIDs", "Unicode"]
git-tree-sha1 = "7f4228017b83c66bd6aa4fddeb170ce487e53bc7"
git-tree-sha1 = "834aedb1369919a7b2026d7e04c2d49a311d26f4"
uuid = "295af30f-e4ad-537b-8983-00126c2a3abe"
version = "3.6.2"
version = "3.6.3"

[[deps.SHA]]
uuid = "ea8e919c-243c-51af-8825-aaa63cd721ce"
Expand Down
0 Tuning/Manifest.toml → tuning/Manifest.toml
View file
File renamed without changes.
0 Tuning/Project.toml → tuning/Project.toml
View file
File renamed without changes.
7 changes: 4 additions & 3 deletions Tuning/README.md → tuning/README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -18,16 +18,17 @@ command line argument:
2) Here, for each call of the function to tune, a new Julia process is started. While this
approach allows for parallelization, also using a compute cluster, it can be quite
inefficient when the function to tune is comparably fast w.r.t. Julia's startup and
pre-compile time.
pre-compile times.

3) Here, the main Python script spawns a fixed number of Julia "server" subprocesses
realized by `julia_server.jl`, which accept function calls, perform them,
and return the result. The communication is done via pipes, respectively stdin/stdout.
In this way multiprocessing is utilized, but the overhead of starting
a new Julia process for each function call is avoided.

Note that the demo needs to be started from the `Tuning` subdirectory.
Note that the demo needs to be started from the `tuning` subdirectory.

Ensure that Python is installed with the packages `juliacall`, `smac` and `ConfigSpace`.
Ensure that Python is installed with current versions of packages `juliacall`, `smac`,
and `ConfigSpace`.


2 changes: 1 addition & 1 deletion Tuning/call-julia-function-f.jl → tuning/call-julia-function-f.jl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@ include("julia-function-to-tune.jl")
# println(f, d)
# end

c = f(d["--instance"], parse(Int, d["--seed"]),
c = f(d["--instance"], parse(Int, d["--seed"]),
parse(Float64, d["--x"]),
parse(Int, d["--y"]),
d["--z"],
Expand Down
2 changes: 1 addition & 1 deletion Tuning/julia-function-to-tune.jl → tuning/julia-function-to-tune.jl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -5,7 +5,7 @@
Demo function to tune with SMAC3 in different ways.
"""
function f(instance::AbstractString, seed::Int, x::Float64, y::Int, z::String)::Float64
function f(instance::AbstractString, seed::Int, x::Float64, y::Int, z::AbstractString)::Float64
# just some busy waiting:
xx=3
for i in 1:10000000
Expand Down
0 Tuning/julia_server.py → tuning/julia_server.py
View file
File renamed without changes.
23 changes: 15 additions & 8 deletions Tuning/tuning.py → tuning/tuning.py
View file
Original file line number Diff line number Diff line change
@@ -1,15 +1,20 @@
#!/usr/bin/env python3
# Demo for using SMAC3 to optimize the configuration of a Julia algorithm.

# Ensure that Python is installed with the packages `pyjulia`, `smac` and `ConfigSpace`
# Ensure that Python is installed with the packages `juliacall`, `smac` and `ConfigSpace`.
# You may provide 1, 2, or 3 as command line argument to select the variant to use.


import ConfigSpace
from ConfigSpace import Configuration, ConfigurationSpace
from smac import AlgorithmConfigurationFacade, Scenario
from smac import AlgorithmConfigurationFacade, HyperparameterOptimizationFacade, Scenario
from smac.intensifier.intensifier import Intensifier
import os
import sys

# print(os.getcwd()); os.chdir("tuning") #!!!


# simple way to specify a configuration space
config_space = ConfigurationSpace({
"x": (0.1, 4.0),
Expand All @@ -19,7 +24,7 @@

# alternative long form, advanced configuration aspects possible:
config_space2 = ConfigurationSpace()
config_space2.add_hyperparameters([
config_space2.add([
ConfigSpace.UniformFloatHyperparameter("x", 0.1, 4.0),
ConfigSpace.UniformIntegerHyperparameter("y", 1, 3),
ConfigSpace.CategoricalHyperparameter("z", ["opt1", "opt2"]),
Expand All @@ -29,11 +34,11 @@
instance_dir = "../test/data"
# names of problem instances to be used for tuning
instances = [fn for fn in os.listdir(instance_dir) if fn.startswith("maxsat")]
instances = instances[:1] # limit to first two instances for testing

# features of the problem instances
# a mapping of the problem instances to their features,
# in the simplest case just the index, or otherwise some more relevant features
features = {fn: [i] for (i, fn) in enumerate(instances)}
# names of problem instances to be used for tuning

# Scenario object specifying the optimization environment
scenario = Scenario(config_space2, deterministic=False,
Expand All @@ -54,13 +59,15 @@

# exemplary wrapper for Julia function to tune
def f(config: Configuration, instance: str, seed: int) -> float:
print(f'f({instance}, {seed}, {config["x"]}, {config["y"]}, {config["z"]})',
end=" -> ")
res = jl.f(instance, seed, config["x"], config["y"], config["z"])
x = float(config["x"]); y = int(config["y"]); z = str(config["z"])
print(f'f({instance}, {seed}, {x}, {y}, {y})', end=" -> ")
res = jl.f(instance, seed, x, y, z)
print(res)
return res

smac = AlgorithmConfigurationFacade(scenario, f, overwrite=True)
# smac = HyperparameterOptimizationFacade(scenario, f, overwrite=True)



elif variant_to_use == "2":
Expand Down

2 comments on commit 5b55b7d

@graidl
Copy link
Collaborator Author

@graidl graidl commented on 5b55b7d Nov 19, 2024

Choose a reason for hiding this comment

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

@JuliaRegistrator register

@JuliaRegistrator
Copy link

Choose a reason for hiding this comment

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

Registration pull request created: JuliaRegistries/General/119735

Tip: Release Notes

Did you know you can add release notes too? Just add markdown formatted text underneath the comment after the text
"Release notes:" and it will be added to the registry PR, and if TagBot is installed it will also be added to the
release that TagBot creates. i.e.

@JuliaRegistrator register

Release notes:

## Breaking changes

- blah

To add them here just re-invoke and the PR will be updated.

Tagging

After the above pull request is merged, it is recommended that a tag is created on this repository for the registered package version.

This will be done automatically if the Julia TagBot GitHub Action is installed, or can be done manually through the github interface, or via:

git tag -a v0.1.15 -m "<description of version>" 5b55b7d2a51cf8b146db817a36b564cdd464c67c
git push origin v0.1.15

Please sign in to comment.