Use julia_project to manage Julia dependency

.gitignore

@@ -56,3 +56,6 @@ docs/_build/
 
 # PyBuilder
 target/
+
+**/Manifest.toml
+**/*.so

LICENSE.md

@@ -1,7 +1,7 @@
 
 The MIT License (MIT)
 
-Copyright (c) 2018: Chris Rackauckas.
+Copyright (c) 2018, 2022: Chris Rackauckas.
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

MANIFEST.in

@@ -1,4 +1,7 @@
 include README.md
 include LICENSE.md
 include docs/*.txt
+include diffeqpy/sys_image/*.jl
+include diffeqpy/sys_image/Project.toml
+include diffeqpy/Project.toml
 include diffeqpy/*.jl

README.md

@@ -31,16 +31,14 @@ To install diffeqpy, use pip:
 pip install diffeqpy
 ```
 
-Using diffeqpy requires that Julia is installed and in the path, along
-with DifferentialEquations.jl and PyCall.jl. To install Julia,
-download a generic binary from
+You can either install Julia yourself, or allow diffeqpy to do this for you.
+To install Julia yourself, download a generic binary from
 [the JuliaLang site](https://julialang.org/downloads/) and add it to your path.
 To install Julia packages required for diffeqpy, open up Python
 interpreter then run:
 
 ```pycon
 >>> import diffeqpy
->>> diffeqpy.install()
 ```
 
 and you're good! In addition, to improve the performance of your code it is

diffeqpy/Project.toml

@@ -0,0 +1,5 @@
+[deps]
+DiffEqBase = "2b5f629d-d688-5b77-993f-72d75c75574e"
+DifferentialEquations = "0c46a032-eb83-5123-abaf-570d42b7fbaa"
+OrdinaryDiffEq = "1dea7af3-3e70-54e6-95c3-0bf5283fa5ed"
+PyCall = "438e738f-606a-5dbb-bf0a-cddfbfd45ab0"

diffeqpy/__init__.py

@@ -1,37 +1,3 @@
-import os
-import shutil
-import subprocess
-import sys
+from ._julia_project import julia, compile_diffeqpy, update_diffeqpy
 
-from jill.install import install_julia
-
-script_dir = os.path.dirname(os.path.realpath(__file__))
-
-
-def _find_julia():
-    # TODO: this should probably fallback to query jill
-    return shutil.which("julia")
-
-
-def install(*, confirm=False):
-    """
-    Install Julia (if required) and Julia packages required for diffeqpy.
-    """
-    julia = _find_julia()
-    if not julia:
-        print("No Julia version found. Installing Julia.")
-        install_julia(confirm=confirm)
-        julia = _find_julia()
-        if not julia:
-            raise RuntimeError(
-                "Julia installed with jill but `julia` binary cannot be found in the path"
-            )
-    env = os.environ.copy()
-    env["PYTHON"] = sys.executable
-    subprocess.check_call([julia, os.path.join(script_dir, "install.jl")], env=env)
-
-
-def _ensure_installed(*kwargs):
-    if not _find_julia():
-        # TODO: this should probably ensure that packages are installed too
-        install(*kwargs)
+from ._version import __version__

diffeqpy/_julia_project.py

@@ -0,0 +1,38 @@
+import julia
+import logging
+
+from julia_project import JuliaProject
+
+import os
+diffeqpy_path = os.path.dirname(os.path.abspath(__file__))
+
+julia_project = JuliaProject(
+    name="diffeqpy",
+    package_path=diffeqpy_path,
+    preferred_julia_versions = ['1.7', '1.6', 'latest'],
+    env_prefix = 'DIFFEQPY_',
+    logging_level = logging.INFO, # or logging.WARN,
+    console_logging=False
+)
+
+julia_project.run()
+
+# logger = julia_project.logger
+
+def compile_diffeqpy():
+    """
+    Compile a system image for `diffeqpy` in the subdirectory `./sys_image/`. This
+    system image will be loaded the next time you import `diffeqpy`.
+    """
+    julia_project.compile_julia_project()
+
+
+def update_diffeqpy():
+    """
+    Remove possible stale Manifest.toml files and compiled system image.
+    Update Julia packages and rebuild Manifest.toml file.
+    Before compiling, it's probably a good idea to call this method, then restart Python.
+    """
+    julia_project.update()
+
+

diffeqpy/_version.py

@@ -0,0 +1,2 @@
+"""Define version number here and read it from setup.py automatically"""
+__version__ = "1.2.0"

diffeqpy/de.py

@@ -1,12 +1,6 @@
 import os
 import sys
 
-from . import _ensure_installed
-
-# This is terrifying to many people. However, it seems SciML takes pragmatic approach.
-_ensure_installed()
-
-# PyJulia have to be loaded after `_ensure_installed()`
 from julia import Main
 
 script_dir = os.path.dirname(os.path.realpath(__file__))

diffeqpy/sys_image/Project.toml

@@ -0,0 +1,6 @@
+[deps]
+DiffEqBase = "2b5f629d-d688-5b77-993f-72d75c75574e"
+DifferentialEquations = "0c46a032-eb83-5123-abaf-570d42b7fbaa"
+OrdinaryDiffEq = "1dea7af3-3e70-54e6-95c3-0bf5283fa5ed"
+PackageCompiler = "9b87118b-4619-50d2-8e1e-99f35a4d4d9d"
+PyCall = "438e738f-606a-5dbb-bf0a-cddfbfd45ab0"

diffeqpy/sys_image/compile_exercise_script.jl

@@ -0,0 +1,51 @@
+using DifferentialEquations
+const DE = DifferentialEquations
+
+# This is curiously slow
+function sde_exercise()
+    f = (u,p,t) -> 1.01*u
+    g = (u,p,t) -> 0.87*u
+    u0 = 0.5
+    tspan = (0.0,1.0)
+    prob = DE.SDEProblem(f,g,u0,tspan)
+    sol = DE.solve(prob,reltol=1e-3,abstol=1e-3)
+    return nothing
+end
+
+function ode_exercise()
+    f = (u,p,t) -> -u
+    u0 = 0.5
+    tspan = (0., 1.)
+    prob = DE.ODEProblem(f, u0, tspan)
+    sol = DE.solve(prob)
+    return nothing
+end
+
+function ode_exercise2()
+    f = function(u,p,t)
+        x, y, z = u
+        sigma, rho, beta = p
+        return [sigma * (y - x), x * (rho - z) - y, x * y - beta * z]
+    end
+    u0 = [1.0,0.0,0.0]
+    tspan = (0., 100.)
+    p = [10.0,28.0,8/3]
+    prob = DE.ODEProblem(f, u0, tspan, p)
+    sol = DE.solve(prob,saveat=0.01)
+    return nothing
+end
+
+# From ODE docs
+function ode_exercise3()
+    f(u,p,t) = 1.01*u
+    u0 = 1/2
+    tspan = (0.0,1.0)
+    prob = ODEProblem(f,u0,tspan)
+    sol = solve(prob, Tsit5(), reltol=1e-8, abstol=1e-8)
+    return nothing
+end
+
+ode_exercise()
+ode_exercise2()
+ode_exercise3()
+sde_exercise()

diffeqpy/sys_image/compile_julia_project.jl

@@ -0,0 +1,11 @@
+using PackageCompiler
+using Libdl: Libdl
+
+packages = [:PyCall, :DiffEqBase, :DifferentialEquations, :OrdinaryDiffEq]
+
+sysimage_path = joinpath(@__DIR__, "sys_julia_project." * Libdl.dlext)
+
+#create_sysimage(packages; sysimage_path=sysimage_path)
+
+create_sysimage(packages; sysimage_path=sysimage_path,
+                precompile_execution_file=joinpath(@__DIR__, "compile_exercise_script.jl"))

setup.py

@@ -4,8 +4,14 @@ def readme():
     with open('README.md') as f:
         return f.read()
 
+
+version = {}
+with open("./diffeqpy/_version.py") as fp:
+    exec(fp.read(), version)
+
+
 setup(name='diffeqpy',
-      version='1.2.0',
+      version=version['__version__'],
       description='Solving Differential Equations in Python',
       long_description=readme(),
       long_description_content_type="text/markdown",
@@ -24,6 +30,7 @@ def readme():
       author_email='[email protected]',
       license='MIT',
       packages=['diffeqpy','diffeqpy.tests'],
-      install_requires=['julia>=0.2', 'jill'],
+      install_requires=['julia>=0.2',
+                        'julia_project>=0.0.23'],
       include_package_data=True,
       zip_safe=False)

tox.ini

@@ -3,7 +3,7 @@ deps =
     pytest
     pytest-cov
 commands =
-    python -c 'import diffeqpy; diffeqpy.install()'
+    python -c 'import diffeqpy'
     py.test \
         --pyargs diffeqpy \
         {posargs}
@@ -12,6 +12,9 @@ whitelist_externals =
 setenv =
     # Do not use matplotlib GUI backend during tests.
     MPLBACKEND = agg
+    DIFFEQPY_INSTALL_JULIA = y
+    DIFFEQPY_COMPILE = y
+    DIFFEQPY_DEPOT = y
 passenv =
     # Allow a workaround for "error initializing LibGit2 module":
     # https://github.com/JuliaLang/julia/issues/18693