climada_python/test_notebooks.py at main · Numlet/climada_python · GitHub

History

129 lines (101 loc) · 4.02 KB

Raw

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

#!/usr/bin/env python

# coding: utf-8

import os

import sys

import unittest

import nbformat

from climada.util.constants import SOURCE_DIR

NOTEBOOK_DIR = os.path.abspath('doc/tutorial')

'''The path to the notebook directories.'''

BOUND_TO_FAIL = '# Note: execution of this cell will fail'

'''Cells containing this line will not be executed in the test'''

class NotebookTest(unittest.TestCase):

'''Generic TestCase for testing the executability of notebooks

Attributes

----------

wd : str

Absolute Path to the working directory, i.e., the directory of the notebook.

notebook : str

File name of the notebook.

'''

def __init__(self, methodName, wd=None, notebook=None):

super(NotebookTest, self).__init__(methodName)

self.wd = wd

self.notebook = notebook

def test_notebook(self):

'''Extracts code cells from the notebook and executes them one by one, using `exec`.

Magic lines and help/? calls are eliminated.

Cells containing `BOUND_TO_FAIL` are elided.

Cells doing multiprocessing are elided.'''

# cd to the notebook directory

os.chdir(self.wd)

print(f'start testing {self.notebook}')

# read the notebook into a string

with open(self.notebook, encoding='utf8') as nb:

content = nb.read()

# parse the string with nbformat.reads

cells = nbformat.reads(content, 4)['cells']

namespace = dict()

for i, c in enumerate(cells):

# skip markdown cells

if c['cell_type'] != 'code': continue

# skip deliberately failing cells

if BOUND_TO_FAIL in c['source']: continue

# skip multiprocessing cells

if any([ tabu in c['source'].split() for tabu in [

'pathos.pools',

'mulitprocessing',

]]):

print('\n'.join([

f'\nskip multiprocessing cell {i} in {self.notebook}',

'+'+'-'*68+'+',

c['source']

]))

continue

# remove non python lines and help calls which require user input

python_code = "\n".join([ln for ln in c['source'].split("\n")

if not ln.startswith('%')

and not ln.startswith('help(')

and not ln.startswith('ask_ok(')

and not ln.strip().endswith('?')

])

# execute the python code

try:

exec(python_code, namespace)

# report failures

except Exception as e:

failure = "\n".join([

f"notebook {self.notebook} cell {i} failed with {e.__class__}",

f"{e}",

'+'+'-'*68+'+',

c['source']

])

print(f'failed {self.notebook}')

self.fail(failure)

print(f'succeeded {self.notebook}')

def main():

# list notebooks in the NOTEBOOK_DIR

notebooks = [(NOTEBOOK_DIR, f)

for f in sorted(os.listdir(NOTEBOOK_DIR))

if os.path.splitext(f)[1] == ('.ipynb')]

# build a test suite with a test for each notebook

suite = unittest.TestSuite()

for (jd,nb) in notebooks:

suite.addTest(NotebookTest('test_notebook', jd, nb))

# run the tests depending on the first input argument: None or 'report'.

# write xml reports for 'report'

if sys.argv[1:]:

arg = sys.argv[1]

if arg == 'report':

import xmlrunner

output = os.path.join(SOURCE_DIR, '../tests_xml')

xmlrunner.XMLTestRunner(output=output).run(suite)

else:

jd, nb = os.path.split(arg)

unittest.TextTestRunner(verbosity=2).run(NotebookTest('test_notebook', jd, nb))

# with no argument just run the test

else:

unittest.TextTestRunner(verbosity=2).run(suite)

if __name__ == '__main__':

sys.path.append(os.getcwd())

main()

climada_python/test_notebooks.py at main · Numlet/climada_python · GitHub

Expand file tree

Breadcrumbs

test_notebooks.py

Latest commit

History

Breadcrumbs

test_notebooks.py

File metadata and controls