Commit 5fa5773c authored by Pierre Augier's avatar Pierre Augier
Browse files

Reopen dev branch

--HG--
branch : dev
......@@ -2,21 +2,29 @@ syntax: glob
__pycache__
.vscode/
*~
doc/_build/*
doc/**generated/*
doc/**.zip
**/.ipynb_checkpoints/*
build/*
dist/*.tar.gz
dist*
.tox/*
.coverage/*
**/tmp*.py
doc/ipynb/executed/*.rst
doc/ipynb/*.rst
.pytest_cache
*.egg-info/*
**/_pythran/_*
**/__pythran__/_*
fluidpythran/_path_data_tests.py
\ No newline at end of file
e6747f29d9a9e1a00d71837c1cc2c05730d1a60e 0.0.1
944b19191afd5dd4d03619196f7533d8bfa16968 0.0.4
1b81afabb2c802af0fa33c98e8341b33ce19100f 0.0.8
e6bb435e2a4d9b7c8465201507ed9fce3933f902 0.0.9
f365e250af615922db7a362b27e1af40695a3491 0.1.0
conda:
file: doc/environment.yml
formats:
- none
python:
version: 3
python:
setup_py_install: true
\ No newline at end of file
The main contributors are:
- `Pierre Augier <http://www.legi.grenoble-inp.fr/people/Pierre.Augier/>`_
(LEGI, CNRS, UGA)
with ideas from `Serge Sans Paille <https://github.com/serge-sans-paille/>`_,
`Ashwin Vishnu <https://github.com/ashwinvis>`_ and `Wolf Vollprecht
<https://twitter.com/wuoulf>`_.
\ No newline at end of file
0.1.1 (2018-11-28)
------------------
- :code:`wait_for_all_extensions`
- Bug fixes
- :code:`mocked_modules` argument for functions making AOT Pythran files
0.1.0 (2018-11-23)
------------------
- Compatibility PyPy3.5
- Fix bug script importing local script
- Command line option "clear-cache"
0.0.9 (2018-11-20)
------------------
- :code:`PYTHRANIZE_AT_IMPORT` mode for ahead-of-time and just-in-time
compilation (works also in IPython)
- By default, the fluidpythran commandline uses Pythran
0.0.8 (2018-11-16)
------------------
- Fix cachedjit when calling with new types
- :code:`cachedjit` in IPython / Jupyter
0.0.7 (2018-11-15)
------------------
- :code:`cachedjit` decorator (supports also type hints)
0.0.6 (2018-11-05)
------------------
- Type annotations to define Pythran functions and blocks
0.0.5 (2018-10-14)
------------------
......
develop:
python setup.py develop
pip install -e .[test]
black:
black -l 82 fluidpythran
tests:
pytest fluidpythran
pytest fluidpythran data_tests/ipynb
tests_nbval:
pytest --nbval data_tests/ipynb
tests_coverage:
mkdir -p .coverage
......
FluidPythran: use Pythran in non-pythranizable code
===================================================
FluidPythran: easily speedup your Python code with Pythran
==========================================================
|release| |coverage|
|release| |docs| |coverage|
.. |release| image:: https://img.shields.io/pypi/v/fluidpythran.svg
:target: https://pypi.python.org/pypi/fluidpythran/
:alt: Latest version
.. |docs| image:: https://readthedocs.org/projects/fluidpythran/badge/?version=latest
:target: http://fluidpythran.readthedocs.org
:alt: Documentation status
.. |coverage| image:: https://codecov.io/bb/fluiddyn/fluidpythran/branch/default/graph/badge.svg
:target: https://codecov.io/bb/fluiddyn/fluidpythran/branch/default/
:alt: Code coverage
......@@ -14,27 +18,31 @@ FluidPythran: use Pythran in non-pythranizable code
.. warning ::
FluidPythran is still just a prototype. Remarks and suggestions are very
FluidPythran is in an early stage. Remarks and suggestions are very
welcome.
FluidPythran just starts to be used in `FluidSim
FluidPythran starts to be used in `FluidSim
<https://bitbucket.org/fluiddyn/fluidsim>`_ (for example in `this file
<https://bitbucket.org/fluiddyn/fluidsim/src/c0e170ea7c68f2abc4b0f7749b1c89df79db6573/fluidsim/base/time_stepping/pseudo_spect.py>`_).
See also `this blog post
<http://www.legi.grenoble-inp.fr/people/Pierre.Augier/broadcasting-numpy-abstraction-cython-pythran-fluidpythran.html>`_
for an explanation of my motivations.
<https://bitbucket.org/fluiddyn/fluidsim/src/default/fluidsim/base/time_stepping/pseudo_spect.py>`_).
FluidPythran is a pure Python package (requiring Python >= 3.6 or Pypy3) to
help to write Python code that can use `Pythran
<https://github.com/serge-sans-paille/pythran>`_.
help to write Python code that *can* use `Pythran
<https://github.com/serge-sans-paille/pythran>`_ if it is available.
Let's recall that "Pythran is an ahead-of-time (AOT) compiler for a subset of
the Python language, with a focus on scientific computing. It takes a Python
module annotated with a few interface description and turns it into a native
Python module with the same interface, but (hopefully) faster."
Pythran is able to produce **very efficient C++ code and binaries from high
level Numpy code**. If the algorithm is easier to express without loops, don't
write loops!
Let's recall that "Pythran is an ahead of time compiler for a subset of the
Python language, with a focus on scientific computing. It takes a Python module
annotated with a few interface description and turns it into a native Python
module with the same interface, but (hopefully) faster."
Pythran always releases the GIL and can use SIMD instructions and OpenMP!
**FluidPythran does not depend on Pythran.**
**Pythran is not a hard dependency of FluidPythran:** Python code using
FluidPythran run fine without Pythran and without compilation (and of course
without speedup)!
Overview
--------
......@@ -44,29 +52,51 @@ scientific programs and libraries.
To use Pythran, one needs to isolate the numerical kernels functions in modules
that are compiled by Pythran. The C++ code produced by Pythran never uses the
Python interpretor. It means that only a subset of what is doable in Python can
Python interpreter. It means that only a subset of what is doable in Python can
be done in Pythran files. Some `language features
<https://pythran.readthedocs.io/en/latest/MANUAL.html#disclaimer>`_ are not
supported by Pythran (for example no classes) and most of the extension
packages cannot be used in Pythran files (basically `only Numpy and some Scipy
functions <https://pythran.readthedocs.io/en/latest/SUPPORT.html>`_).
Another cause of frustration for Python developers when using Pythran is
related to manual writting of Pythran function signatures in comments, which
can not be automated. Pythran uses C++ templates but Pythran users can not
think with this concept. We would like to be able to **express the templated
nature of Pythran with modern Python syntax** (in particular **type
annotations**). Finally, another limitation is that it is not possible to use
Pythran for **just-in-time** (JIT) compilation so one needs to manually write
all argument types.
With FluidPythran, we try to overcome these limitations. FluidPythran provides
few supplementary Pythran commands and a tiny Python API to define Pythran
few supplementary Pythran commands and a small Python API to define Pythran
functions without writing the Pythran modules. The code of the numerical
kernels can stay in the modules and in the classes where they were written. The
Pythran files (i.e. the files compiled by Pythran), which are usually written
by the user, are produced automatically by FluidPythran.
**Implementation detail:** For each Python file using FluidPythran, an
associated Pythran file is created in a directory :code:`_pythran`. For
example, for a Python file :code:`foo.py`, the associated file would be
:code:`_pythran/_foo.py`.
Bonus: There are FluidPythran syntaxes for both **ahead-of-time** and
**just-in-time** compilations!
At run time, FluidPythran uses when possible the pythranized functions, but
let's stress again that codes using FluidPythran work fine without Pythran (of
course without speedup)!
To summarize, a **strategy to quickly develop a very efficient scientific
application/library** with Python could be:
At run time, FluidPythran replaces the Python functions (and blocks) by their
versions in the Pythran files.
- Use modern Python coding, standard Numpy/Scipy for the computations and all
the cool libraries you want.
Let's stress again that codes using FluidPythran work fine without Pythran!
- Profile your applications on real cases, detect the bottlenecks and apply
standard optimizations with Numpy.
- Add few lines of FluidPythran to compile the hot spots.
**Implementation details:** Under the hood, FluidPythran creates Pythran files
(one per module for AOT compilation and one per function for JIT compilation)
that can be compiled at build, import or run times depending of the cases. Note
that the developers can still read the Pythran files if needed.
Installation
------------
......@@ -75,8 +105,8 @@ Installation
pip install fluidpythran
Using Pythran in Python files
-----------------------------
A short tour of FluidPythran syntaxes
-------------------------------------
Command :code:`# pythran def`
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
......@@ -108,11 +138,78 @@ Most of this code looks familiar to Pythran users. The differences:
Python function by the pythranized function if FluidPythran has been used to
produced the associated Pythran file.
Pythran using type annotations
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The previous example can be rewritten without Pythran commands:
.. code :: python
import h5py
import mpi4py
from fluidpythran import pythran_def
@pythran_def
def myfunc(a: int, b: float):
return a * b
...
Nice but very limited... So it is possible to mix type hints and :code:`#
pythran def` commands. Moreover, one can also elegantly define many Pythran
signatures with type variables (see `these examples in the documentation
<https://fluidpythran.readthedocs.io/en/latest/examples/type_hints.html>`_).
Cached Just-In-Time compilation
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
With FluidPythran, one can use the Ahead-Of-Time compiler Pythran in a
Just-In-Time mode. It is really the **easiest way to speedup a function with
Pythran**, just by adding a decorator! And it also works `in notebooks
<https://fluidpythran.readthedocs.io/en/latest/ipynb/executed/demo_cachedjit.html>`_!
It is a "work in progress" so (i) it could be buggy and (ii) the API is not
great, but it is a good start!
.. code :: python
import numpy as np
# pythran import numpy as numpy
from fluidpythran import cachedjit, used_by_cachedjit
@used_by_cachedjit("func1")
def func0(a, b):
return a + b
@cachedjit
def func1(a, b):
return np.exp(a) * b * func0(a, b)
Note that the :code:`@cachedjit` decorator takes into account type hints (see
`the example in the documentation
<https://fluidpythran.readthedocs.io/en/latest/examples/using_cachedjit.html>`_).
If the environment variable :code:`PYTHRANIZE_AT_IMPORT` is set, fluidpythran
compiles at import time the functions with type hints.
**Implementation details for just-in-time compilation:** A Pythran file is
produced for each "cachedjited" function (function decorated with
:code:`@cachedjit`). The file is compiled at the first call of the function and
the compiled version is used as soon as it is ready. The warmup can be quite
long but the compiled version is saved and can be reused (without warmup!) by
another process.
Command :code:`# pythran block`
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
One of the most evident application of :code:`# pythran block` is code in
classes:
FluidPythran blocks can be used with classes and more generally in functions
with lines that cannot be compiled by Pythran.
.. code :: python
......@@ -127,7 +224,7 @@ classes:
def func(self, n):
a, b = self.something_that_cannot_be_pythranized()
if fp.is_pythranized:
if fp.is_transpiled:
result = fp.use_pythranized_block("name_block")
else:
# pythran block (
......@@ -152,71 +249,50 @@ For blocks, we need a little bit more of Python.
- In the function, we define a block with three lines of Python and special
Pythran annotations (:code:`# pythran block`). The 3 lines of Python are used
(i) at run time to choose between the two branches (:code:`is_pythranized` or
(i) at run time to choose between the two branches (:code:`is_transpiled` or
not) and (ii) at compile time to detect the blocks.
Note that the annotations in the command :code:`# pythran block` are different
(and somehow easier to write) than in the standard command :code:`# pythran
export`.
.. note ::
`Blocks can now also be defined with type hints!
<https://fluidpythran.readthedocs.io/en/latest/examples/blocks.html>`_
Moreover, for the time being, one needs to explicitly write the "returned"
variables (after :code:`->`). However, it is a redundant information so we
could avoid this in future (see `issue #1
<https://bitbucket.org/fluiddyn/fluidpythran/issues/1/no-need-for-explicit-return-values-in>`_).
.. warning ::
The two branches of the :code:`if fp.is_pythranized` are not equivalent! The
user has to be careful because it is not difficult to write such buggy code:
.. code :: python
c = 0
if fp.is_pythranized:
a, b = fp.use_pythranized_block("buggy_block")
else:
# pythran block () -> (a, b)
a = b = c = 1
assert c == 1
.. note ::
The Pythran keyword :code:`or` cannot be used in block annotations (not yet
implemented, see `issue #2
<https://bitbucket.org/fluiddyn/fluidpythran/issues/2/implement-keyword-or-in-block-annotation>`_).
Python classes: :code:`@pythran_def` for methods
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Command :code:`# pythran class`
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Just a NotImplemented idea! See https://bitbucket.org/fluiddyn/fluidpythran/issues/3/pythran-class
Just a NotImplemented idea! See https://bitbucket.org/fluiddyn/fluidpythran/issues/3
For simple methods only using simple attributes, if could be simple and useful
to support this:
For simple methods only using simple attributes, if could be simple and *very*
useful to support this:
.. code :: python
from fluidpythran import pythran_class
from fluidpythran import Type, NDim, Array, pythran_def
import numpy as np
@pythran_class
T = Type(int, np.float64)
N = NDim(1)
A1 = Array[T, N]
A2 = Array[float, N+1]
class MyClass:
# pythran class (
# int[] or float[]: arr0, arr1;
# float[][]: arr2
# )
arr0: A1
arr1: A1
arr2: A2
def __init__(self, n, dtype=int):
self.arr0 = np.zeros(n, dtype=dtype)
self.arr1 = np.zeros(n, dtype=dtype)
self.arr2 = np.zeros(n)
# pythran def compute(object, float)
def compute(self, alpha):
@pythran_def
def compute(self, alpha: int):
tmp = (self.arr0 + self.arr1).mean()
return tmp ** alpha * self.arr2
......@@ -224,7 +300,8 @@ Make the Pythran files
----------------------
There is a command-line tool :code:`fluidpythran` which makes the associated
Pythran files from Python files with annotations and fluidpythran code.
Pythran files from Python files with annotations and fluidpythran code. By
default and if Pythran is available, the Pythran files are compiled.
There is also a function :code:`make_pythran_files` that can be used in a
setup.py like this:
......@@ -240,13 +317,17 @@ setup.py like this:
paths = ["fluidsim/base/time_stepping/pseudo_spect.py"]
make_pythran_files([here / path for path in paths])
Note that FluidPythran never uses Pythran. Compiling the associated Pythran
file can be done if wanted (see for example how it is done in the example
package `example_package_fluidpythran
Note that the function :code:`make_pythran_files` does not use Pythran.
Compiling the associated Pythran file can be done if wanted (see for example
how it is done in the example package `example_package_fluidpythran
<https://bitbucket.org/fluiddyn/example_package_fluidpythran>`_ or in
`fluidsim's setup.py
<https://bitbucket.org/fluiddyn/fluidsim/src/default/setup.py>`_).
If the environment variable :code:`PYTHRANIZE_AT_IMPORT` is set, FluidPythran
compiles at import time (i.e. only when needed) the Pythran file associated
with the imported module.
License
-------
......
......@@ -16,6 +16,13 @@ pipelines:
script:
- pip install -U tox --user
- tox -e py36,codecov
- step:
image: fluiddyn/python3-stable
caches:
- pip
script:
- pip install -U tox --user
- tox -vv -e py36-pythran,codecov
dev:
- parallel:
- step:
......@@ -25,3 +32,10 @@ pipelines:
script:
- pip install -U tox --user
- tox -e py36,codecov
- step:
image: fluiddyn/python3-stable
caches:
- pip
script:
- pip install -U tox --user
- tox -vv -e py36-pythran,codecov
......@@ -17,7 +17,7 @@ class MyClass:
state_spect_n12 = self._state_spect_tmp
if fp.is_pythranized:
if fp.is_transpiled:
fp.use_pythranized_block("rk2_step0")
else:
# pythran block (
......
import numpy as np
# pythran import numpy as np
from fluidpythran import FluidPythran, Type, NDim, Array
T = Type(float, complex)
N = NDim(1, 2)
A = Array[T, N]
A1 = Array[T, N + 1]
fp = FluidPythran()
class MyClass:
def __init__(self, a, b):
self.a = a
self.b = b
def compute(self, n):
a = self.a
b = self.b
if fp.is_transpiled:
result = fp.use_pythranized_block("block0")
else:
# pythran block (
# A a; A1 b;
# int n
# ) -> result
# pythran block (
# int[:] a, b;
# float n
# ) -> result
result = a ** 2 + b.mean() ** 3 + n
return result
if __name__ == "__main__":
shape = 100, 100
a = np.random.rand(*shape)
b = np.random.rand(*shape)
obj = MyClass(a, b)
obj.compute(10)
if fp.is_transpiled:
ret = obj.compute(10)
fp.is_transpiled = False
ret1 = obj.compute(10)
fp.is_transpiled = True
assert np.allclose(ret, ret1)
print("allclose OK")
......@@ -19,7 +19,7 @@ class MyClass:
a = self.a
b = self.b
if fp.is_pythranized:
if fp.is_transpiled:
result = fp.use_pythranized_block("block0")
else:
# pythran block (
......@@ -40,13 +40,13 @@ class MyClass:
a = result
if fp.is_pythranized:
if fp.is_transpiled:
result = fp.use_pythranized_block("block1")
else:
# pythran block (
# float[][] a, b;
# int n
# ) -> (result, a)
# ) -> (result)
# blabla
# blibli
......@@ -54,7 +54,7 @@ class MyClass:
# pythran block (
# float[][][] a, b;
# int n
# ) -> (result, a)
# ) -> (result)
result = np.zeros_like(a)
for _ in range(n):
result += a ** 2 + b ** 3
......@@ -64,18 +64,21 @@ class MyClass:
if __name__ == "__main__":
shape = 100, 100
shape = 2, 2
a = np.random.rand(*shape)
b = np.random.rand(*shape)
obj = MyClass(a, b)
obj.compute(10)
ret0 = obj.compute(10)
if fp.is_pythranized:
print("(is_transpiled, is_compiling, is_compiled)", (fp.is_transpiled, fp.is_compiling, fp.is_compiled))
if fp.is_transpiled:
ret = obj.compute(10)
fp.is_pythranized = False
assert np.allclose(ret, ret0), (ret - ret0)
fp.is_transpiled = False
ret1 = obj.compute(10)
fp.is_pythranized = True
assert np.allclose(ret, ret1)
fp.is_transpiled = True
assert np.allclose(ret, ret1), (ret - ret1)
print("allclose OK")
......@@ -3,6 +3,8 @@ import numpy as np
from fluidpythran import pythran_def
# FLUIDPYTHRAN_NO_IMPORT
# pythran import numpy as np
# pythran def func(
......