.gitignore and dummy numba

Author: jhauschild

.gitignore

@@ -0,0 +1,3 @@
+__pycache__
+*.txt
+*.pkl

numba.py

@@ -0,0 +1,43 @@
+"""Provide dummy alternative for `numba.jit`.
+
+The python package `numba <http://numba.pydata.org/>` provides just-in-time compilation,
+which can often give a huge speed-up. However, it is not included into the python standard library.
+Although e.g. the python distributions provided by anaconda already come with numba, it might not
+be available on your laptops, so we provide a "dummy" alternative for the "@jit" decorator used
+in some of the python scripts.
+Moreover, this dummy is also usefull for debugging, since the "@jit" often leads to confusing,
+lengthy tracebacks and error messages.
+
+If you get an "ImportError: No module named 'numba'" or want to disable the @jit for debugging,
+simply copy this file in the folder where you run the scripts.
+Note that you need might need to restart the kernel to force a new import, simply calling the cell
+with the "import" is not enough.
+
+
+Example
+-------
+The usage is as with the jit decorator of the "real" numba package,
+but the @jit simply doesn't do anything::
+
+ from numba import jit
+
+ @jit
+ def example_function(a, b):
+ return a + b
+
+ @jit(nopython=True)
+ def another_example_function(a, b):
+ return a * b
+
+"""
+warnings.warn("did not import numba, but dummy jit. Code will work but run slowly!")
+
+def jit(*args, **kwargs):
+ """Dummy decorator, doing nothing, for replacing numba.jit if not available"""
+ if len(args) > 0:
+ return args[0]
+ else:
+ def dummy_decorator(func):
+ """Dummy decorator"""
+ return func
+ return dummy_decorator