I can see that INCLUDE path is sysconfig.get_path('include').
But I don't see any similar value for LIB.
NumPy outright hardcodes it as os.path.join(sys.prefix, "libs") in Windows and get_config_var('LIBDIR') (not documented and missing in Windows) otherwise.
Is there a more supported way?
Since it's not a part of any official spec/doc, and, as shown by another answer, there are cases when none of appropriate variables from sysconfig/distutils.sysconfig .get_config_var() are set,
the only way to reliably get it in all cases, exactly as a build would (e.g. even for a Python in the sourcetree) is to delegate to the reference implementation.
In distutils, the logic that sets the library path for a compiler is located in distutils.commands.build_ext.finalize_options(). So, this code would get it with no side effects on the build:
import distutils.command.build_ext #imports distutils.core, too
d = distutils.core.Distribution()
b = distutils.command.build_ext.build_ext(d) #or `d.get_command_class('build_ext')(d)',
# then it's enough to import distutils.core
b.finalize_options()
print b.library_dirs
Note that:
Not all locations in the resulting list necessarily exist.
If your setup.py is setuptools-based, use setuptools.Distribution and setuptools.command.build_ext instead, correspondingly.
If you pass any values to setup() that affect the result, you must pass them to Distribution here, too.
Since there are no guarantees that the set of the additional values you need to pass will stay the same, and the value is only needed when building an extension,
it seems like you aren't really supposed to get this value independently at all:
If you're using another build facility, you should rather subclass build_ext and get the value from the base method during the build.
Below is the (rather long) subroutine in skbuild.cmaker that locates libpythonxx.so/pythonxx.lib for the running Python. In CMake, 350-line Modules/FindPythonLibs.cmake is dedicated to this task.
The part of the former that gets just the directory is much simpler though:
libdir = dustutils.sysconfig.get_config_var('LIBDIR')
if sysconfig.get_config_var('MULTIARCH'):
masd = sysconfig.get_config_var('multiarchsubdir')
if masd:
if masd.startswith(os.sep):
masd = masd[len(os.sep):]
libdir = os.path.join(libdir, masd)
if libdir is None:
libdir = os.path.abspath(os.path.join(
sysconfig.get_config_var('LIBDEST'), "..", "libs"))
def get_python_library(python_version):
"""Get path to the python library associated with the current python
interpreter."""
# determine direct path to libpython
python_library = sysconfig.get_config_var('LIBRARY')
# if static (or nonexistent), try to find a suitable dynamic libpython
if (python_library is None or
os.path.splitext(python_library)[1][-2:] == '.a'):
candidate_lib_prefixes = ['', 'lib']
candidate_extensions = ['.lib', '.so', '.a']
if sysconfig.get_config_var('WITH_DYLD'):
candidate_extensions.insert(0, '.dylib')
candidate_versions = [python_version]
if python_version:
candidate_versions.append('')
candidate_versions.insert(
0, "".join(python_version.split(".")[:2]))
abiflags = getattr(sys, 'abiflags', '')
candidate_abiflags = [abiflags]
if abiflags:
candidate_abiflags.append('')
# Ensure the value injected by virtualenv is
# returned on windows.
# Because calling `sysconfig.get_config_var('multiarchsubdir')`
# returns an empty string on Linux, `du_sysconfig` is only used to
# get the value of `LIBDIR`.
libdir = du_sysconfig.get_config_var('LIBDIR')
if sysconfig.get_config_var('MULTIARCH'):
masd = sysconfig.get_config_var('multiarchsubdir')
if masd:
if masd.startswith(os.sep):
masd = masd[len(os.sep):]
libdir = os.path.join(libdir, masd)
if libdir is None:
libdir = os.path.abspath(os.path.join(
sysconfig.get_config_var('LIBDEST'), "..", "libs"))
candidates = (
os.path.join(
libdir,
''.join((pre, 'python', ver, abi, ext))
)
for (pre, ext, ver, abi) in itertools.product(
candidate_lib_prefixes,
candidate_extensions,
candidate_versions,
candidate_abiflags
)
)
for candidate in candidates:
if os.path.exists(candidate):
# we found a (likely alternate) libpython
python_library = candidate
break
# TODO(opadron): what happens if we don't find a libpython?
return python_library
I'm trying to determine how the gen_io_ops module is generated by bazel when building TensorFlow from source.
In tensorflow/python/ops/io_ops.py, there is this piece of code:
from tensorflow.python.ops.gen_io_ops
[...]
# used in the TextLineReader initialization
rr = gen_io_ops._text_line_reader_v2(...)
referring to the bazel-genfiles/tensorflow/python/ops/gen_io_ops.py module (and generated by bazel when building TensorFlow).
The _text_line_reader_v2 is a wrapper of the TextLineReaderV2 defined in tensorflow/tensorflow/core/kernels/text_line_reader_op.cc.
As far as I understand, the build step are the followings:
1) The kernel library for the text_line_reader_op is built in tensorflow/tensorflow/core/kernels/BUILD
tf_kernel_library(
name = "text_line_reader_op",
prefix = "text_line_reader_op",
deps = IO_DEPS,)
where tf_kernel_library basically looks for text_line_reader_op.c file and build it.
2) The :text_line_reader_op kernel library is then used as a dependency by the io library defined in the same file:
cc_library(
name = "io",
deps = [
":text_line_reader_op", ...
],
)
I suppose the io library now contains the definition of the TextLineReaderV2kernel.
From what I get from this answer, there should be a third step where the io library is used to generate the python wrappers that are in the bazel-genfiles/tensorflow/python/ops/gen_io_ops.py module. This file generation can be done by the tf_op_gen_wrapper_py rule in Basel or by thetf.load_op_library() method, but none of them seem involved.
Does someone know where this third step is defined in the build process?
I finally got it.
There is indeed a call to tf_op_gen_wrapper_py but it's hidden in a call to tf_gen_op_wrapper_private_py:
def tf_gen_op_wrapper_private_py(name, out=None, deps=[],
require_shape_functions=True,
visibility=[]):
if not name.endswith("_gen"):
fail("name must end in _gen")
[...]
bare_op_name = name[:-4]
tf_gen_op_wrapper_py(name=bare_op_name, ...
So the steps are the following.
In tensorflow/tensorflow/python/BUILD, there is this rule
tf_gen_op_wrapper_private_py(
name = "io_ops_gen",
[...]
)
And so, in this rule the _gen suffix will be removed (in tf_gen_op_wrapper_private_py) and a gen_ prefix will be added in tf_gen_op_wrapper_py and therefore the gen_io_ops.py module will be generated by this rule.
I have a target, that requires some time-expensive actions and depends on similar time-expensive targets. It's also a phony target, so it doesn't depends on a source files nor it doesn't produces a files. And I want to make SCons decide when to build this target or not depending on some abstract value.
To me it would be ideal to use the capability of SCons, described in the manual as:
Value(value, [built_value]) , env.Value(value, [built_value])
... Value Nodes can be used as dependencies of targets. If the result of calling str(value) changes between SCons runs, any targets depending on Value(value) will be rebuilt. ...
So I've build this SConstruct to test it:
import random as rnd
env = Environment()
def action(**_):
print('Do crazy things')
input = env.Value('a')
env['BUILDERS']['Custom'] = Builder(action=action)
env.Custom(target='target', source=input)
env.Pseudo('target')
And I expected SCons to stop rebuilding the target after the first build, since the value of the input Value remains constant between builds. But SCons rebuilds it every time.
Is there any way to achieve what I'm talking about?
A Python function can be a SCons action used in a Builder, as described in the SCons user manual. This function will be called by SCons whenever ANY of the sources changes. Sources are passed as a list of SCons Files objects
Inside this function I want to know if a specific source has been changed (and thus which source files forced SCons to rebuild the target of the builder).
so if 'file.target' is built using 'file1.src' and 'file2.src' using a custom builder that calls to python function 'custom_build':
env.CUSTOM_BUILDER('file.target', ['file1.src', 'file2.src'])
def custom_build(env, source, target):
# Checks if file1.src has been modified
if source[0].has_been_modified: # Does something like this exist??
I have inspected this File objects while debugging, but without success. Is there any property to know that a specific file changed inside a SCons builder?
I found a possible solution while searching on Node.__init__ package (SCons option --debug=explain provides here information about sources changed for each target). It is not very nice, but seems to do the job. At least, it fits me well. The key is to use changed_since_last_build on the target.
def build(env, target, source):
old = target[0].get_stored_info()
old = old.binfo
old.prepare_dependencies()
old_bkids = old.bsources + old.bdepends + old.bimplicit
old_bkidsigs = old.bsourcesigs + old.bdependsigs + old.bimplicitsigs
osig = dict(zip(old_bkids, old_bkidsigs))
for src_file, info in osig.iteritems():
print str(src_file) + ' change status: ' + str(src_file.changed_since_last_build(target[0], info))
I'm trying to add cross-references to external API into my documentation but I'm facing three different behaviors.
I am using sphinx(1.3.1) with Python(2.7.3) and my intersphinx mapping is configured as:
{
'python': ('https://docs.python.org/2.7', None),
'numpy': ('http://docs.scipy.org/doc/numpy/', None),
'cv2' : ('http://docs.opencv.org/2.4/', None),
'h5py' : ('http://docs.h5py.org/en/latest/', None)
}
I have no trouble writing a cross-reference to numpy API with :class:`numpy.ndarray` or :func:`numpy.array` which gives me, as expected, something like numpy.ndarray.
However, with h5py, the only way I can have a link generated is if I omit the module name. For example, :class:`Group` (or :class:`h5py:Group`) gives me Group but :class:`h5py.Group` fails to generate a link.
Finally, I cannot find a way to write a working cross-reference to OpenCV API, none of these seems to work:
:func:`cv2.convertScaleAbs`
:func:`cv2:cv2.convertScaleAbs`
:func:`cv2:convertScaleAbs`
:func:`convertScaleAbs`
How to properly write cross-references to external API, or configure intersphinx, to have a generated link as in the numpy case?
In addition to the detailed answer from #gall, I've discovered that intersphinx can also be run as a module:
python -m sphinx.ext.intersphinx 'http://python-eve.org/objects.inv'
This outputs nicely formatted info. For reference: https://github.com/sphinx-doc/sphinx/blob/master/sphinx/ext/intersphinx.py#L390
I gave another try on trying to understand the content of an objects.inv file and hopefully this time I inspected numpy and h5py instead of only OpenCV's one.
How to read an intersphinx inventory file
Despite the fact that I couldn't find anything useful about reading the content of an object.inv file, it is actually very simple with the intersphinx module.
from sphinx.ext import intersphinx
import warnings
def fetch_inventory(uri):
"""Read a Sphinx inventory file into a dictionary."""
class MockConfig(object):
intersphinx_timeout = None # type: int
tls_verify = False
class MockApp(object):
srcdir = ''
config = MockConfig()
def warn(self, msg):
warnings.warn(msg)
return intersphinx.fetch_inventory(MockApp(), '', uri)
uri = 'http://docs.python.org/2.7/objects.inv'
# Read inventory into a dictionary
inv = fetch_inventory(uri)
# Or just print it
intersphinx.debug(['', uri])
File structure (numpy)
After inspecting numpy's one, you can see that keys are domains:
[u'np-c:function',
u'std:label',
u'c:member',
u'np:classmethod',
u'np:data',
u'py:class',
u'np-c:member',
u'c:var',
u'np:class',
u'np:function',
u'py:module',
u'np-c:macro',
u'np:exception',
u'py:method',
u'np:method',
u'np-c:var',
u'py:exception',
u'np:staticmethod',
u'py:staticmethod',
u'c:type',
u'np-c:type',
u'c:macro',
u'c:function',
u'np:module',
u'py:data',
u'np:attribute',
u'std:term',
u'py:function',
u'py:classmethod',
u'py:attribute']
You can see how you can write your cross-reference when you look at the content of a specific domain. For example, py:class:
{u'numpy.DataSource': (u'NumPy',
u'1.9',
u'http://docs.scipy.org/doc/numpy/reference/generated/numpy.DataSource.html#numpy.DataSource',
u'-'),
u'numpy.MachAr': (u'NumPy',
u'1.9',
u'http://docs.scipy.org/doc/numpy/reference/generated/numpy.MachAr.html#numpy.MachAr',
u'-'),
u'numpy.broadcast': (u'NumPy',
u'1.9',
u'http://docs.scipy.org/doc/numpy/reference/generated/numpy.broadcast.html#numpy.broadcast',
u'-'),
...}
So here, :class:`numpy.DataSource` will work as expected.
h5py
In the case of h5py, the domains are:
[u'py:attribute', u'std:label', u'py:method', u'py:function', u'py:class']
and if you look at the py:class domain:
{u'AttributeManager': (u'h5py',
u'2.5',
u'http://docs.h5py.org/en/latest/high/attr.html#AttributeManager',
u'-'),
u'Dataset': (u'h5py',
u'2.5',
u'http://docs.h5py.org/en/latest/high/dataset.html#Dataset',
u'-'),
u'ExternalLink': (u'h5py',
u'2.5',
u'http://docs.h5py.org/en/latest/high/group.html#ExternalLink',
u'-'),
...}
That's why I couldn't make it work as numpy references. So a good way to format them would be :class:`h5py:Dataset`.
OpenCV
OpenCV's inventory object seems malformed. Where I would expect to find domains there is actually 902 function signatures:
[u':',
u'AdjusterAdapter::create(const',
u'AdjusterAdapter::good()',
u'AdjusterAdapter::tooFew(int',
u'AdjusterAdapter::tooMany(int',
u'Algorithm::create(const',
u'Algorithm::getList(vector<string>&',
u'Algorithm::name()',
u'Algorithm::read(const',
u'Algorithm::set(const'
...]
and if we take the first one's value:
{u'Ptr<AdjusterAdapter>': (u'OpenCV',
u'2.4',
u'http://docs.opencv.org/2.4/detectorType)',
u'ocv:function 1 modules/features2d/doc/common_interfaces_of_feature_detectors.html#$ -')}
I'm pretty sure it is then impossible to write OpenCV cross-references with this file...
Conclusion
I thought intersphinx generated the objects.inv based on the content of the documentation project in an standard way, which seems not to be the case.
As a result, it seems that the proper way to write cross-references is API dependent and one should inspect a specific inventory object to actually see what's available.
An additional way to inspect the objects.inv file is with the sphobjinv module.
You can search local or even remote inventory files (with fuzzy matching). For instance with scipy:
$ sphobjinv suggest -t 90 -u https://docs.scipy.org/doc/scipy/reference/objects.inv "signal.convolve2d"
Remote inventory found.
:py:function:`scipy.signal.convolve2d`
:std:doc:`generated/scipy.signal.convolve2d`
Note that you may need to use :py:func: and not :py:function: (I'd be happy to know why).
How to use OpenCV 2.4 (cv2) intersphinx
Inspired by #Gall's answer, I wanted to compare the contents of the OpenCV & numpy inventory files. I couldn't get sphinx.ext.intersphinx.fetch_inventory to work from ipython, but the following does work:
curl http://docs.opencv.org/2.4/objects.inv | tail -n +5 | zlib-flate -uncompress > cv2.inv
curl https://docs.scipy.org/doc/numpy/objects.inv | tail -n +5 | zlib-flate -uncompress > numpy.inv
numpy.inv has lines like this:
numpy.ndarray py:class 1 reference/generated/numpy.ndarray.html#$ -
whereas cv2.inv has lines like this:
cv2.imread ocv:pyfunction 1 modules/highgui/doc/reading_and_writing_images_and_video.html#$ -
So presumably you'd link to the OpenCV docs with :ocv:pyfunction:`cv2.imread` instead of :py:function:`cv2.imread`. Sphinx doesn't like it though:
WARNING: Unknown interpreted text role "ocv:pyfunction".
A bit of Googling revealed that the OpenCV project has its own "ocv" sphinx domain: https://github.com/opencv/opencv/blob/2.4/doc/ocv.py -- presumably because they need to document C, C++ and Python APIs all at the same time.
To use it, save ocv.py next to your Sphinx conf.py, and modify your conf.py:
sys.path.insert(0, os.path.abspath('.'))
import ocv
extensions = [
'ocv',
]
intersphinx_mapping = {
'cv2': ('http://docs.opencv.org/2.4/', None),
}
In your rst files you need to say :ocv:pyfunc:`cv2.imread` (not :ocv:pyfunction:).
Sphinx prints some warnings (unparseable C++ definition: u'cv2.imread') but the generated html documentation actually looks ok with a link to http://docs.opencv.org/2.4/modules/highgui/doc/reading_and_writing_images_and_video.html#cv2.imread. You can edit ocv.py and remove the line that prints that warning.
The accepted answer no longer works in the new version (1.5.x) ...
import requests
import posixpath
from sphinx.ext.intersphinx import read_inventory
uri = 'http://docs.python.org/2.7/'
r = requests.get(uri + 'objects.inv', stream=True)
r.raise_for_status()
inv = read_inventory(r.raw, uri, posixpath.join)
Stubborn fool that I am, I used 2to3 and the Sphinx deprecated APIs chart to revive #david-röthlisberger's ocv.py-based answer so it'll work with Sphinx 2.3 on Python 3.5.
The fixed-up version is here:
https://gist.github.com/ssokolow/a230b27b7ea4a31f7fb40621e6461f9a
...and the quick version of what I did was:
Run 2to3 -w ocv.py && rm ocv.py.bak
Cycle back and forth between running Sphinx and renaming functions to their replacements in the chart. I believe these were the only changes I had to make on this step:
Directive now has to be imported from docutils.parsers.rst
Replace calls to l_(...) with calls to _(...) and remove the l_ import.
Replace calls to env.warn with calls to log.warn where log = sphinx.util.logging.getLogger(__name__).
Then, you just pair it with this intersphinx definition and you get something still new enough to be relevant for most use cases:
'cv2': ('https://docs.opencv.org/3.0-last-rst/', None)
For convenience, I made a small extension for aliasing intersphinx cross references. This is useful as sometimes the object inventory gets confused when an object from a submodule is imported from a package's __init__.py.
See also https://github.com/sphinx-doc/sphinx/issues/5603
###
# Workaround of
# Intersphinx references to objects imported at package level can"t be mapped.
#
# See https://github.com/sphinx-doc/sphinx/issues/5603
intersphinx_aliases = {
("py:class", "click.core.Group"):
("py:class", "click.Group"),
("py:class", "click.core.Command"):
("py:class", "click.Command"),
}
def add_intersphinx_aliases_to_inv(app):
from sphinx.ext.intersphinx import InventoryAdapter
inventories = InventoryAdapter(app.builder.env)
for alias, target in app.config.intersphinx_aliases.items():
alias_domain, alias_name = alias
target_domain, target_name = target
try:
found = inventories.main_inventory[target_domain][target_name]
try:
inventories.main_inventory[alias_domain][alias_name] = found
except KeyError:
print("could not add to inv")
continue
except KeyError:
print("missed :(")
continue
def setup(app):
app.add_config_value("intersphinx_aliases", {}, "env")
app.connect("builder-inited", add_intersphinx_aliases_to_inv)
To use this, I paste the above code in my conf.py and add aliases to the intersphinx_aliases dictionary.