/
utils.py
270 lines (219 loc) · 9.86 KB
/
utils.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
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
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
#!/usr/bin/env python
# encoding: utf-8
# Andre Anjos <andre.dos.anjos@gmail.com>
# Fri 21 Mar 2014 10:37:40 CET
"""General utilities for building extensions"""
import os
import re
import sys
import pkg_resources
def load_requirements(f=None):
"""Loads the contents of requirements.txt on the given path.
Defaults to "./requirements.txt"
"""
def readlines(f):
retval = [str(k.strip()) for k in f]
return [k for k in retval if k and k[0] not in ("#", "-")]
# if f is None, use the default ('requirements.txt')
if f is None:
f = "requirements.txt"
if isinstance(f, str):
f = open(f, "rt")
# read the contents
return readlines(f)
def find_packages(directories=["bob"]):
"""This function replaces the ``find_packages`` command from ``setuptools`` to search for packages only in the given directories.
Using this function will increase the building speed, especially when you have (links to) deep non-code-related directory structures inside your package directory.
The given ``directories`` should be a list of top-level sub-directories of your package, where package code can be found.
By default, it uses ``'bob'`` as the only directory to search.
"""
from setuptools import find_packages as _original
if isinstance(directories, str):
directories = [directories]
packages = []
for d in directories:
packages += [d]
packages += ["%s.%s" % (d, p) for p in _original(d)]
return packages
def link_documentation(
additional_packages=["python", "numpy"],
requirements_file="../requirements.txt",
server=None,
):
"""Generates a list of documented packages on our documentation server for the packages read from the "requirements.txt" file and the given list of additional packages.
Parameters:
additional_packages : [str]
A list of additional bob packages for which the documentation urls are added.
By default, 'numpy' is added
requirements_file : str or file-like
The file (relative to the documentation directory), where to read the requirements from.
If ``None``, it will be skipped.
server : str or None
The url to the server which provides the documentation.
If ``None`` (the default), the ``BOB_DOCUMENTATION_SERVER`` environment variable is taken if existent.
If neither ``server`` is specified, nor a ``BOB_DOCUMENTATION_SERVER`` environment variable is set, the default ``"http://www.idiap.ch/software/bob/docs/bob/%(name)s/%(version)s/"`` is used.
"""
def smaller_than(v1, v2):
"""Compares scipy/numpy version numbers"""
c1 = v1.split(".")
c2 = v2.split(".")[: len(c1)] # clip to the compared version
for i in range(len(c2)):
n1 = c1[i]
n2 = c2[i]
try:
n1 = int(n1)
n2 = int(n2)
except ValueError:
n1 = str(n1)
n2 = str(n2)
if n1 < n2:
return True
if n1 > n2:
return False
return False
import urllib.error as error
import urllib.request as urllib
HTTPError = error.HTTPError
URLError = error.URLError
# collect packages are automatically included in the list of indexes
packages = []
version_re = re.compile(r"\s*[\<\>=]+\s*")
if requirements_file is not None:
if not isinstance(requirements_file, str) or os.path.exists(
requirements_file
):
requirements = load_requirements(requirements_file)
packages += [version_re.split(k)[0] for k in requirements]
packages += additional_packages
def _add_index(name, addr, packages=packages):
"""Helper to add a new doc index to the intersphinx catalog
Parameters:
name (str): Name of the package that will be added to the catalog
addr (str): The URL (except the ``objects.inv`` file), that will be added
"""
if name in packages:
print("Adding intersphinx source for `%s': %s" % (name, addr))
mapping[name] = (addr, None)
packages = [k for k in packages if k != name]
def _add_numpy_index():
"""Helper to add the numpy manual"""
try:
import numpy
ver = numpy.version.version
ver = ".".join(ver.split(".")[:-1])
_add_index("numpy", "https://numpy.org/doc/%s/" % ver)
except ImportError:
_add_index("numpy", "https://numpy.org/devdocs/")
def _add_scipy_index():
"""Helper to add the scipy manual"""
try:
import scipy
ver = scipy.version.version
if smaller_than(ver, "0.9.0"):
ver = ".".join(ver.split(".")[:-1]) + ".x"
else:
ver = ".".join(ver.split(".")[:-1]) + ".0"
_add_index(
"scipy", "https://docs.scipy.org/doc/scipy-%s/reference/" % ver
)
except ImportError:
_add_index("scipy", "https://docs.scipy.org/doc/scipy/reference/")
def _add_click_index():
"""Helper to add the click manual"""
import click
major, minor = [int(x) for x in click.__version__.split(".")[0:2]]
if major < 8:
ver = f"{major}.x"
else:
ver = f"{major}.{minor}.x"
_add_index("click", "https://click.palletsprojects.com/en/%s/" % ver)
mapping = {}
# add indexes for common packages used in Bob
_add_index(
"python", "https://docs.python.org/%d.%d/" % sys.version_info[:2]
)
_add_numpy_index()
_add_index("scipy", "https://docs.scipy.org/doc/scipy/")
_add_index("matplotlib", "https://matplotlib.org/stable/")
_add_index("setuptools", "https://setuptools.readthedocs.io/en/latest/")
_add_index("six", "https://six.readthedocs.io")
_add_index("sqlalchemy", "https://docs.sqlalchemy.org/en/latest/")
_add_index("docopt", "https://docopt.readthedocs.io/en/latest/")
_add_index("scikit-learn", "https://scikit-learn.org/stable/")
_add_index("scikit-image", "https://scikit-image.org/docs/dev/")
_add_index("pillow", "https://pillow.readthedocs.io/en/latest/")
_add_index("PIL", "https://pillow.readthedocs.io/en/latest/")
_add_index("pandas", "https://pandas.pydata.org/pandas-docs/stable/")
_add_index("dask", "https://docs.dask.org/en/latest/")
_add_index("dask-jobqueue", "https://jobqueue.dask.org/en/latest/")
_add_index("distributed", "https://distributed.dask.org/en/latest/")
_add_click_index()
_add_index("torch", "https://pytorch.org/docs/stable/")
_add_index("xarray", "https://xarray.pydata.org/en/stable/")
_add_index("h5py", "https://docs.h5py.org/en/stable/")
# get the server for the other packages
if server is None:
if "BOB_DOCUMENTATION_SERVER" in os.environ:
server = os.environ["BOB_DOCUMENTATION_SERVER"]
else:
server = (
"http://www.idiap.ch/software/bob/docs/bob/%(name)s/%(version)s/"
"|http://www.idiap.ch/software/bob/docs/bob/%(name)s/%(version)s/sphinx"
"|http://www.idiap.ch/software/bob/docs/bob/%(name)s/main/"
"|http://www.idiap.ch/software/bob/docs/bob/%(name)s/main/sphinx"
"|http://www.idiap.ch/software/bob/docs/bob/%(name)s/master/"
"|http://www.idiap.ch/software/bob/docs/bob/%(name)s/master/sphinx"
)
# array support for BOB_DOCUMENTATION_SERVER
# transforms "(file:///path/to/dir https://example.com/dir| http://bla )"
# into ["file:///path/to/dir", "https://example.com/dir", "http://bla"]
# so, trim eventual parenthesis/white-spaces and splits by white space or |
if server.strip():
server = re.split(r"[|\s]+", server.strip("() "))
else:
server = []
# check if the packages have documentation on the server
for p in packages:
if p in mapping:
continue # do not add twice...
for s in server:
# generate URL
package_name = p.split()[0]
if s.count("%s") == 1: # old style
url = s % package_name
else: # use new style, with mapping, try to link against specific version
try:
version = (
"v" + pkg_resources.require(package_name)[0].version
)
except pkg_resources.DistributionNotFound:
version = "stable" # package is not a runtime dep, only referenced
url = s % {"name": package_name, "version": version}
try:
# otherwise, urlopen will fail
if url.startswith("file://"):
urllib.urlopen(urllib.Request(url + "objects.inv"))
url = url[7:] # intersphinx does not like file://
else:
urllib.urlopen(urllib.Request(url))
# request url
print(
"Found documentation for %s on %s; adding intersphinx source"
% (p, url)
)
mapping[p] = (url, None)
break # inner loop, for server, as we found a candidate!
except HTTPError as exc:
if exc.code != 404:
# url request failed with a something else than 404 Error
print("Requesting URL %s returned error: %s" % (url, exc))
# notice mapping is not updated here, as the URL does not exist
print(f"Nothing found at {url}.")
except URLError as exc:
print(
"Requesting URL %s did not succeed (maybe offline?). "
"The error was: %s" % (url, exc)
)
except IOError as exc:
print("Path %s does not exist. The error was: %s" % (url, exc))
return mapping