glotzerlab / gsd Goto Github PK

View Code? Open in Web Editor NEW

24.0 24.0 7.0 2.16 MB

Read and write GSD files for use with HOOMD-blue.

Home Page: http://gsd.readthedocs.io

License: BSD 2-Clause "Simplified" License

CMake 2.68% Python 50.13% C 31.95% C++ 1.45% Cython 13.78%

binary file-format gsd hoomd python

gsd's People

Contributors

Stargazers

Watchers

Forkers

jennyfothergill oywz99 mgorny joaander stukowski honeyankit 1vip1786

gsd's Issues

add body field

Original report by Joshua Anderson (Bitbucket: joaander, GitHub: joaander).

with the new forcecomposite design, the body field is still needed for proper restarts. Also, users may find it useful to detect constituent particles during analysis.

File handle not closed after failed call to gsd_open()

Original report by stukowski (Bitbucket: stukowski, GitHub: stukowski).

It looks like the gsd_open() function leaves the file open after returning with an error code. Consider the following code:

#!c++
gsd_handle handle;
if(gsd_open(&handle, "not_a_gsd_file.txt", GSD_OPEN_READONLY) != 0)
    return;
...

Calling gsd_close() after the failed call to gsd_open() doesn't help either.

I am the developer of the OVITO software, which uses the GSD I/O layer to read HOOMD simulation files. As part of the file format auto-detection feature, OVITO routinely calls gsd_open() on files that are not GSD files. I noticed this problem when the maximum number of open files allowed by the OS was reached by the process.

Release v2.1.1

Release checklist:

Run bumpversion.
Update change log.
- git log --format=oneline --first-parent `git log -n 1 --pretty=format:%H -- CHANGELOG.rst`...
- milestone
Check readthedocs build, especially change log formatting.
- Build status
- Output
Tag and push.
Update conda-forge recipe.

Implement MPI_IO API

Original report by Joshua Anderson (Bitbucket: joaander, GitHub: joaander).

We need an MPI_IO version of the API to support scalable writes and reads.

python >= 3.2 required for tests but not mentioned in docs

Original report by Dominik Mierzejewski (Bitbucket: rathann, GitHub: rathann).

README.md doesn't say that in order to run nose tests, python 3.2 or later is required due to the usage of tempfile.TemporaryDirectory, which was introduced in python 3.2.

As a result, tests fail when run under python 2.7:

+ nosetests-2.7
nose.config: INFO: Set working dir to /builddir/build/BUILD/gsd-v1.5.0/tests
nose.config: INFO: Ignoring files matching ['^\\.', '^_', '^setup\\.py$']
test_fl.test_create ... ERROR
test_fl.test_dtypes(<type 'numpy.uint8'>,) ... ERROR
test_fl.test_dtypes(<type 'numpy.uint16'>,) ... ERROR
test_fl.test_dtypes(<type 'numpy.uint32'>,) ... ERROR
test_fl.test_dtypes(<type 'numpy.uint64'>,) ... ERROR
test_fl.test_dtypes(<type 'numpy.int8'>,) ... ERROR
test_fl.test_dtypes(<type 'numpy.int16'>,) ... ERROR
test_fl.test_dtypes(<type 'numpy.int32'>,) ... ERROR
test_fl.test_dtypes(<type 'numpy.int64'>,) ... ERROR
test_fl.test_dtypes(<type 'numpy.float32'>,) ... ERROR
test_fl.test_dtypes(<type 'numpy.float64'>,) ... ERROR
test_fl.test_metadata ... ERROR
test_fl.test_append ... ERROR
test_fl.test_chunk_exists ... ERROR
test_fl.test_readonly_errors ... ERROR
test_fl.test_fileio_errors ... ERROR
test_fl.test_dtype_errors ... ERROR
test_fl.test_truncate ... ERROR
test_fl.test_namelen ... ERROR
test_fl.test_open ... ERROR
test_hoomd.test_create ... ERROR
test_hoomd.test_append ... ERROR
test_hoomd.test_extend ... ERROR
test_hoomd.test_defaults ... ERROR
test_hoomd.test_fallback ... ERROR
test_hoomd.test_fallback2 ... ERROR
test_hoomd.test_iteration ... ERROR
test_hoomd.test_truncate ... ERROR
test_hoomd.test_state ... ERROR

======================================================================
ERROR: test_fl.test_create
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 12, in test_create
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_dtypes(<type 'numpy.uint8'>,)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 24, in check_dtype
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_dtypes(<type 'numpy.uint16'>,)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 24, in check_dtype
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_dtypes(<type 'numpy.uint32'>,)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 24, in check_dtype
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_dtypes(<type 'numpy.uint64'>,)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 24, in check_dtype
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_dtypes(<type 'numpy.int8'>,)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 24, in check_dtype
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_dtypes(<type 'numpy.int16'>,)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 24, in check_dtype
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_dtypes(<type 'numpy.int32'>,)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 24, in check_dtype
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_dtypes(<type 'numpy.int64'>,)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 24, in check_dtype
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_dtypes(<type 'numpy.float32'>,)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 24, in check_dtype
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_dtypes(<type 'numpy.float64'>,)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 24, in check_dtype
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_metadata
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 52, in test_metadata
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_append
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 83, in test_append
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_chunk_exists
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 116, in test_chunk_exists
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_readonly_errors
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 186, in test_readonly_errors
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_fileio_errors
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 218, in test_fileio_errors
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_dtype_errors
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 226, in test_dtype_errors
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_truncate
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 258, in test_truncate
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_namelen
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 288, in test_namelen
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_fl.test_open
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_fl.py", line 312, in test_open
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_hoomd.test_create
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_hoomd.py", line 11, in test_create
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_hoomd.test_append
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_hoomd.py", line 20, in test_append
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_hoomd.test_extend
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_hoomd.py", line 41, in test_extend
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_hoomd.test_defaults
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_hoomd.py", line 55, in test_defaults
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_hoomd.test_fallback
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_hoomd.py", line 119, in test_fallback
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_hoomd.test_fallback2
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_hoomd.py", line 335, in test_fallback2
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_hoomd.test_iteration
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_hoomd.py", line 361, in test_iteration
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_hoomd.test_truncate
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_hoomd.py", line 392, in test_truncate
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

======================================================================
ERROR: test_hoomd.test_state
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/builddir/build/BUILD/gsd-v1.5.0/tests/test_hoomd.py", line 408, in test_state
    with tempfile.TemporaryDirectory() as d:
AttributeError: 'module' object has no attribute 'TemporaryDirectory'

----------------------------------------------------------------------
Ran 29 tests in 0.064s

FAILED (errors=29)

publish on pypi

Original report by Max Linke (Bitbucket: max_linke, ).

I just published my first package on pypi following this guide.

I was also able to generate a source distribution and wheel package for gsd following the guide (wheels aren't publishable due to libc version issues though). I only made a few changes to the MANIFEST.in from the pypi branch (adding changelog and readme into the includes). To generate the source distribution I ran

#!bash

python setup.py sdist

This will produce a file dist/gsd-version.tar.gz. You can check this archive if all necessary files are included to build gsd.

publishing a source distribution should be straight forward with this.

I reduced the guide cited above into a short list of commands for myself that I can follow for a release.

https://github.com/bio-phys/MDBenchmark/blob/master/DEVELOPER.rst

note: I'm a core-dev of MDAnalysis

note: If Hoomd writes performance data into log files I'm happy to try include it in my little benchmark utility.

Release v2.3.0

Release checklist:

Run bumpversion.
Update change log.
- git log --format=oneline --first-parent `git log -n 1 --pretty=format:%H -- CHANGELOG.rst`...
- milestone
Check readthedocs build, especially change log formatting.
- Build status
- Output
Tag and push.
Update conda-forge recipe.

gsd.hoomd.open can't load the film

Hello,

I'm trying to use the gsd.hoomd in Windows 10 but have been experiencing the following problem. I think I installed the gsd module correctly but can't use gsd.hoomd.open. The error message was "file layer module is not available". Can anyone please give me some advice if you have had this problem before?

Thanks,
Insung Han

Installation fails for Python 3 with 1.5.2

Original report by Richard Gowers (Bitbucket: rgowers, ).

Hello

I'm trying to build gsd and I get the following error message:

AttributeError: 'dict' object has no attribute '__NUMPY_SETUP__'

This is python 3.6, numpy 1.14.2

From what I can tell this issue is similar to this:

healpy/healpy#391

It's a little hard to reproduce as it only happens when I'm building MDAnalysis (which uses gsd as a prerequisite)

Test on windows python 3.8

TODO: update test.yml to test on python 3.8.

Prerequisite: Python 3.8 exists in the windows image

Use Azure Pipelines

Description

Perform CI tests on Azure Pipelines.

Proposed solution

Combine existing CircleCI and Appveyor jobs into Azure Pipelines.
Add Mac testing.

Additional context

Azure Pipelines provides 10 parallel jobs on Linux, Mac, and Windows. CircleCI provides only 4, and Appveyor only one. Combining all testing and deployment on the same platform will save developer time and effort in the future.

Developer

Seeking developer. This would be a good opportunity for a developer to learn how to configure Azure Pipelines. See recent changes to hoomd-blue and fresnel for a template.

Please create a release tarball

Original report by yurivict (Bitbucket: yurivict, GitHub: yurivict).

Notes in commits say that this is version 1.5.2, but the Downloads page doesn't have the tarball:
https://bitbucket.org/glotzer/gsd/downloads/

Release v2.4.1

Release checklist:

Run bumpversion.
Update change log.
- git log --format=oneline --first-parent `git log -n 1 --pretty=format:%H -- CHANGELOG.rst`...
- milestone
Check readthedocs build, especially change log formatting.
- Build status
- Output
Tag and push.
Update conda-forge recipe.

Release v2.1.0

Release checklist:

Run bumpversion.
Update change log.
- git log --format=oneline --first-parent `git log -n 1 --pretty=format:%H -- CHANGELOG.rst`...
- milestone
Check readthedocs build, especially change log formatting.
- Build status
- Output
Tag and push.
Update conda-forge recipe.

Typos in gsd.hoomd documentation

Description

There are numerous typos in the gsd.hoomd documentation, such as BondData referring to the number of particles. We also need to switch to the preferred notation list[str] instead of using Typing.

Developer

I will work on this at some point.

Windows support

Original report by Joshua Anderson (Bitbucket: joaander, GitHub: joaander).

GSD needs windows support. The primary use-case is for use in file read plugins for VMD and OVITO on windows. I plan to also set up the python module to build on windows as well, mainly to enable testing of the windows back end. For now, I will not plan on providing conda builds.

To do:

Configure MSVC build environment
ifdef around POSIX I/O calls and replace with windows I/O
ifdef around POSIX mmap and replace with windows equivalent
Test and validate on local system
configure appveyor for CI testing

Reading GSD file contents from in-memory buffer

Description

I would like to propose an extension to the GSD C API, which would allow an application to read data from GSD files stored as blob in memory (and not in the local file system).

The GSD C API function gsd_open() allows opening only physical files, but in certain application scenarios, the GSD data may not be stored in the local file system in the first place. Consider for example a program that downloads a GSD file from a web server, and which would like to read the contained data directly without having to write the received data stream to disk first.

Proposed solution

Additional context

Another scenario where the proposed extension is essential is an application running in a sandbox environment without direct access to the file system and which receives the GSD file data as a blob from some other party. A concrete use case I am working on is a port of the OVITO code running in a web browser. See this demo: https://ovito.org/webdemo/

Here, the original OVITO C++ source code has been compiled to platform-independent WebAssembly code running in a browser tab. When the user picks a local file for import into the application, the browser hands over just the file content as a blob and not the actual path to the file. WebAssembly applications have no access to the file system. Thus, OVITO would need to parse the GSD file content from an in-memory buffer.

Developer

test_large_N[268435456] and test_large_N[536870913] FAIL on 32-bit architectures

Description

When running the tests on 32 bit architectures (e.g. ARMv7 or x86 32-bit), the following two tests fail: test_large_N[268435456] and test_large_N[536870913].

Script

pytest-3 -v tests/test_largefile.py

Output

...
tests/test_largefile.py::test_large_N[134217728] PASSED                  [ 96%]
tests/test_largefile.py::test_large_N[268435456] FAILED                  [ 98%]
tests/test_largefile.py::test_large_N[536870913] FAILED                  [100%]
=================================== FAILURES ===================================
___________________________ test_large_N[268435456] ____________________________
tmp_path = PosixPath('/tmp/pytest-of-mockbuild/pytest-0/test_large_N_268435456_0')
N = 268435456
    @pytest.mark.parametrize("N", [2**27, 2**28, 2**29+1])
    def test_large_N(tmp_path, N):
        gc.collect()
    
>       data = numpy.linspace(0, N, num=N, endpoint=False, dtype=numpy.uint32)
tests/test_largefile.py:13: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
<__array_function__ internals>:5: in linspace
    ???
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
start = 0.0, stop = 268435456.0, num = 268435456, endpoint = False
retstep = False, dtype = <class 'numpy.uint32'>, axis = 0
    @array_function_dispatch(_linspace_dispatcher)
    def linspace(start, stop, num=50, endpoint=True, retstep=False, dtype=None,
                 axis=0):
        """
        Return evenly spaced numbers over a specified interval.
    
        Returns `num` evenly spaced samples, calculated over the
        interval [`start`, `stop`].
    
        The endpoint of the interval can optionally be excluded.
    
        .. versionchanged:: 1.16.0
            Non-scalar `start` and `stop` are now supported.
    
        Parameters
        ----------
        start : array_like
            The starting value of the sequence.
        stop : array_like
            The end value of the sequence, unless `endpoint` is set to False.
            In that case, the sequence consists of all but the last of ``num + 1``
            evenly spaced samples, so that `stop` is excluded.  Note that the step
            size changes when `endpoint` is False.
        num : int, optional
            Number of samples to generate. Default is 50. Must be non-negative.
        endpoint : bool, optional
            If True, `stop` is the last sample. Otherwise, it is not included.
            Default is True.
        retstep : bool, optional
            If True, return (`samples`, `step`), where `step` is the spacing
            between samples.
        dtype : dtype, optional
            The type of the output array.  If `dtype` is not given, infer the data
            type from the other input arguments.
    
            .. versionadded:: 1.9.0
    
        axis : int, optional
            The axis in the result to store the samples.  Relevant only if start
            or stop are array-like.  By default (0), the samples will be along a
            new axis inserted at the beginning. Use -1 to get an axis at the end.
    
            .. versionadded:: 1.16.0
    
        Returns
        -------
        samples : ndarray
            There are `num` equally spaced samples in the closed interval
            ``[start, stop]`` or the half-open interval ``[start, stop)``
            (depending on whether `endpoint` is True or False).
        step : float, optional
            Only returned if `retstep` is True
    
            Size of spacing between samples.
    
    
        See Also
        --------
        arange : Similar to `linspace`, but uses a step size (instead of the
                 number of samples).
        geomspace : Similar to `linspace`, but with numbers spaced evenly on a log
                    scale (a geometric progression).
        logspace : Similar to `geomspace`, but with the end points specified as
                   logarithms.
    
        Examples
        --------
        >>> np.linspace(2.0, 3.0, num=5)
        array([2.  , 2.25, 2.5 , 2.75, 3.  ])
        >>> np.linspace(2.0, 3.0, num=5, endpoint=False)
        array([2. ,  2.2,  2.4,  2.6,  2.8])
        >>> np.linspace(2.0, 3.0, num=5, retstep=True)
        (array([2.  ,  2.25,  2.5 ,  2.75,  3.  ]), 0.25)
    
        Graphical illustration:
    
        >>> import matplotlib.pyplot as plt
        >>> N = 8
        >>> y = np.zeros(N)
        >>> x1 = np.linspace(0, 10, N, endpoint=True)
        >>> x2 = np.linspace(0, 10, N, endpoint=False)
        >>> plt.plot(x1, y, 'o')
        [<matplotlib.lines.Line2D object at 0x...>]
        >>> plt.plot(x2, y + 0.5, 'o')
        [<matplotlib.lines.Line2D object at 0x...>]
        >>> plt.ylim([-0.5, 1])
        (-0.5, 1)
        >>> plt.show()
    
        """
        # 2016-02-25, 1.12
        num = _index_deprecate(num)
        if num < 0:
            raise ValueError("Number of samples, %s, must be non-negative." % num)
        div = (num - 1) if endpoint else num
    
        # Convert float/complex array scalars to float, gh-3504
        # and make sure one can use variables that have an __array_interface__, gh-6634
        start = asanyarray(start) * 1.0
        stop  = asanyarray(stop)  * 1.0
    
        dt = result_type(start, stop, float(num))
        if dtype is None:
            dtype = dt
    
        delta = stop - start
>       y = _nx.arange(0, num, dtype=dt).reshape((-1,) + (1,) * ndim(delta))
E       ValueError: array is too big; `arr.size * arr.dtype.itemsize` is larger than the maximum possible size.
/usr/lib/python3.8/site-packages/numpy/core/function_base.py:143: ValueError
___________________________ test_large_N[536870913] ____________________________
tmp_path = PosixPath('/tmp/pytest-of-mockbuild/pytest-0/test_large_N_536870913_0')
N = 536870913
    @pytest.mark.parametrize("N", [2**27, 2**28, 2**29+1])
    def test_large_N(tmp_path, N):
        gc.collect()
    
>       data = numpy.linspace(0, N, num=N, endpoint=False, dtype=numpy.uint32)
tests/test_largefile.py:13: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
<__array_function__ internals>:5: in linspace
    ???
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
start = 0.0, stop = 536870913.0, num = 536870913, endpoint = False
retstep = False, dtype = <class 'numpy.uint32'>, axis = 0
    @array_function_dispatch(_linspace_dispatcher)
    def linspace(start, stop, num=50, endpoint=True, retstep=False, dtype=None,
                 axis=0):
        """
        Return evenly spaced numbers over a specified interval.
    
        Returns `num` evenly spaced samples, calculated over the
        interval [`start`, `stop`].
    
        The endpoint of the interval can optionally be excluded.
    
        .. versionchanged:: 1.16.0
            Non-scalar `start` and `stop` are now supported.
    
        Parameters
        ----------
        start : array_like
            The starting value of the sequence.
        stop : array_like
            The end value of the sequence, unless `endpoint` is set to False.
            In that case, the sequence consists of all but the last of ``num + 1``
            evenly spaced samples, so that `stop` is excluded.  Note that the step
            size changes when `endpoint` is False.
        num : int, optional
            Number of samples to generate. Default is 50. Must be non-negative.
        endpoint : bool, optional
            If True, `stop` is the last sample. Otherwise, it is not included.
            Default is True.
        retstep : bool, optional
            If True, return (`samples`, `step`), where `step` is the spacing
            between samples.
        dtype : dtype, optional
            The type of the output array.  If `dtype` is not given, infer the data
            type from the other input arguments.
    
            .. versionadded:: 1.9.0
    
        axis : int, optional
            The axis in the result to store the samples.  Relevant only if start
            or stop are array-like.  By default (0), the samples will be along a
            new axis inserted at the beginning. Use -1 to get an axis at the end.
    
            .. versionadded:: 1.16.0
    
        Returns
        -------
        samples : ndarray
            There are `num` equally spaced samples in the closed interval
            ``[start, stop]`` or the half-open interval ``[start, stop)``
            (depending on whether `endpoint` is True or False).
        step : float, optional
            Only returned if `retstep` is True
    
            Size of spacing between samples.
    
    
        See Also
        --------
        arange : Similar to `linspace`, but uses a step size (instead of the
                 number of samples).
        geomspace : Similar to `linspace`, but with numbers spaced evenly on a log
                    scale (a geometric progression).
        logspace : Similar to `geomspace`, but with the end points specified as
                   logarithms.
    
        Examples
        --------
        >>> np.linspace(2.0, 3.0, num=5)
        array([2.  , 2.25, 2.5 , 2.75, 3.  ])
        >>> np.linspace(2.0, 3.0, num=5, endpoint=False)
        array([2. ,  2.2,  2.4,  2.6,  2.8])
        >>> np.linspace(2.0, 3.0, num=5, retstep=True)
        (array([2.  ,  2.25,  2.5 ,  2.75,  3.  ]), 0.25)
    
        Graphical illustration:
    
        >>> import matplotlib.pyplot as plt
        >>> N = 8
        >>> y = np.zeros(N)
        >>> x1 = np.linspace(0, 10, N, endpoint=True)
        >>> x2 = np.linspace(0, 10, N, endpoint=False)
        >>> plt.plot(x1, y, 'o')
        [<matplotlib.lines.Line2D object at 0x...>]
        >>> plt.plot(x2, y + 0.5, 'o')
        [<matplotlib.lines.Line2D object at 0x...>]
        >>> plt.ylim([-0.5, 1])
        (-0.5, 1)
        >>> plt.show()
    
        """
        # 2016-02-25, 1.12
        num = _index_deprecate(num)
        if num < 0:
            raise ValueError("Number of samples, %s, must be non-negative." % num)
        div = (num - 1) if endpoint else num
    
        # Convert float/complex array scalars to float, gh-3504
        # and make sure one can use variables that have an __array_interface__, gh-6634
        start = asanyarray(start) * 1.0
        stop  = asanyarray(stop)  * 1.0
    
        dt = result_type(start, stop, float(num))
        if dtype is None:
            dtype = dt
    
        delta = stop - start
>       y = _nx.arange(0, num, dtype=dt).reshape((-1,) + (1,) * ndim(delta))
E       ValueError: array is too big; `arr.size * arr.dtype.itemsize` is larger than the maximum possible size.
/usr/lib/python3.8/site-packages/numpy/core/function_base.py:143: ValueError
===================== 2 failed, 56 passed in 19.44 seconds =====================

Expected output

All tests PASS.

Configuration

Platform:

Linux (Fedora rawhide/32)

Installation method:

Compiled from source

Versions:

Python version: 3.8.0
GSD version: 1.10.0

Developer

CONTRIBUTING.md missing

Original report by Carl Simon Adorf (Bitbucket: csadorf, GitHub: csadorf).

The documentation states that contributors should follow instructions in a CONTRIBUTING.MD file, however that file does not exist.

https://gsd.readthedocs.io/en/stable/community.html#contribute

gsd fl module docs are out of date

Description

The examples in the docs show the write_chunk method has 3 arguments, where it should only take two arguments.

Developer

Anyone who contributes do gsd could do the fix.

Support arbitrarily long names

Description

GSD will be enabled as a log file backed in HOOMD v3. This requires long namespaces for quantities. GSD's current 64-byte key name significantly limits this. Additionally, name lookups should operate in log(N) time.

Proposed solution

No changes to the GSD API are necessary. Internally, we need to bump the file format specification to v2.0 and implement a new namelist storage format. Files written in the v1.0 format will need to be detected and read in a backwards compatible way.

I'm thinking of using a simple memory buffer with all strings in order one after another with null terminators in between them. To make key search in log(N) time, we can make a list of pointers into this buffer and sort those pointers to allow for binary searches. Alternately, we could store the keys sorted in the file, but that would complicate updating the namelist index.

Additional context

We can either a) transparently upgrade files to v2 when writing or b) fail when attempting to write a long key to a v1 file.

This implementation should also solve #43.

Developer

I will implement this.

Windows builds with setup.py fail

Description

When building for Windows with setup.py, the package builds but tests fail. CMake builds on Windows work fine (only CMake builds are tested on Appveyor).

pip install gsd also gives failing Windows builds since it uses the source distribution on PyPI.

Failed Outputs

pytest.log
pytest-15.zip

Configuration

Platform:

Windows (Visual Studio 15 2017 Win64)

Installation method:

Compiled from source

Versions:

Python version: 3.6.6
GSD version: 1.9.2

Developer

I plan to work with @joaander to debug this.

Release v0.10.0

Release checklist:

Run bumpversion.
Update change log.
- git log --format=oneline --first-parent `git log -n 1 --pretty=format:%H -- CHANGELOG.rst`...
- milestone
Check readthedocs build, especially change log formatting.
- Build status
- Output
Tag and push.
Build tarball.
Update conda-forge recipe.

Writing of large GSD files fails

Description

Writing a GSD file with 2^28 particles crashes GSD consistently in the writing process.

Script

import sys
import gsd
import gsd.hoomd
import numpy as np

def main(argv):
    if len(argv) != 1:
        print("Usage: ./test.py")
        return
    test(2**26)
    print("2^26 passed")
    test(2**27)
    print("2^27 passed")
    test(2**28)
    print("2^28 passed")

def test(size):

    s = gsd.hoomd.Snapshot()
    s.particles.N = int(size)
    s.particles.position = np.zeros( (s.particles.N,3) )
    s.particles.position += 0.5
    s.particles.velocity = np.zeros( (s.particles.N,3) )
    s.particles.velocity += 0.3
    s.particles.image = np.zeros( (s.particles.N,3), dtype=np.int )
    s.particles.mass = np.ones(s.particles.N)


    with gsd.hoomd.open(name="temp.gsd",mode="wb") as f:
        f.append(s)

if __name__ == "__main__":
    main(sys.argv)

Output

2^26 passed
2^27 passed
Traceback (most recent call last):
File "test.py", line 36, in
main(sys.argv)
File "test.py", line 17, in main
test(2**28)
File "test.py", line 33, in test
f.append(s)
File "/usr/local/lib/python3.5/dist-packages/gsd/hoomd.py", line 567, in append
self.file.write_chunk(path + '/' + name, data)
File "fl.pyx", line 519, in gsd.fl.GSDFile.write_chunk
OSError: [Errno 0] Success: 'particles/position'

Expected output

2^26 passed
2^27 passed
2^28 passed

Configuration

CPU: Intel(R) Xeon(R) Gold 6126 CPU @ 2.60GHz
Memory: 1TB
disk: network storage with quota of more than 1TB left

Platform:

Linux

Installation method:

Compiled from source
glotzerlab-software container

Versions:

Python version: 3.5.2
GSD version: 1.81

Developer

Name list only holds 128 entries

Description

GSD does not expand the namelist. write_chunk silently writes data with an invalid type id.

Script

# Include a minimal script that reproduces the problem
import gsd.fl
import numpy

with gsd.fl.open(name='file.gsd', mode='wb', application="test", schema="test", schema_version=[1,0]) as f:
    for i in range(256):
        f.write_chunk(name=str(i), data=numpy.array([i], dtype=numpy.int32))
    f.end_frame()

with gsd.fl.open(name='file.gsd', mode='rb', application="test", schema="test", schema_version=[1,0]) as f:
    for i in range(256):
        v = f.read_chunk(frame=0, name=str(i))
        print(v[0])

Output

0
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
Traceback (most recent call last):
  File "gsd-namelist-limit.py", line 11, in <module>
    v = f.read_chunk(frame=0, name=str(i))
  File "fl.pyx", line 623, in gsd.fl.GSDFile.read_chunk
KeyError: 'frame 0 / chunk 128 not found in: file.gsd'
j

Expected output

The values 0 through 255, or an error when attempting to write chunk 128.

Configuration

Platform:

Mac
Linux

Installation method:

Compiled from source

Versions:

Python version: 3.7
GSD version: 1.8.0

Developer

I will add an error condition to write_chunk in v1.9.0.

HOOMD files only use ~50 names. A user would need to enable a large number of custom logs before we need to fix the root cause of the bug. Users with that many chunk names may also wish to implement a hash map for the name list to improve lookup speed.

PyPI still has gsd 1.5.2

Original report by yurivict (Bitbucket: yurivict, GitHub: yurivict).

https://pypi.org/project/gsd/

Did you forget to upload the tarball for 1.5.3?

GSD read-only mode should only memory map the index

Description

Currently, in read-only mode the entire GSD file is memory mapped and then the index is read. However, this can prevent the reading of very large files on operating systems where virtual memory is limited because even allocating a sufficient number of memory pages exceeds the system limitations. To solve this problem, we should only memory map the index itself, and then access the rest of the file on demand.

Script

# This requires working with a GSD file and an operating system where the filesize exceeds the output of `ulimit -v`.
import gsd.hoomd
with gsd.hoomd.open('sample.gsd', 'rb') as f:
    pass

Note that using mode 'rb+' fixes this issue. I have not attached a usable file due to the required file sizes.

Developer

@vyasr

Release v2.1.2

Release checklist:

Run bumpversion.
Update change log.
- git log --format=oneline --first-parent `git log -n 1 --pretty=format:%H -- CHANGELOG.rst`...
- milestone
Check readthedocs build, especially change log formatting.
- Build status
- Output
Tag and push.
Update conda-forge recipe.

pip install picks up GSD 1.8.0 on py2.7

Description

I think this is partially related to #30

We're doing a legacy support release for MDAnalysis and we're noticing that pip install gsd will pick up v1.8.0 on python 2.7. As far as I'm aware 1.8 dropped support for py2.7, causing gsd calls to fail.

This causes issues with pip-only installs. Technically we could pin the release to <1.8.0 but we'd prefer not limit users from using the newer GSD versions where possible.

Assuming you can reproduce this behaviour, I was therefore wondering if you'd be willing to yank the 1.8 release from pypi so that py2.7 didn't default to it?

Configuration

Platform:

Linux: ubuntu 20.04 LTS

Installation method:

pip (pypi)

Versions:

Python version: 2.7
GSD version: 1.8.0

Developer

Unfortunately any fix here would have to be done by the pypi maintainer.

build 1.8.0 showing up for python 2.7 on conda

Description

Hi,

Going from #28 it looks like python 2.7 support got dropped in GSD 1.8.0. However, there are still python2.7 conda builds being created. I couldn't work out who maintains the conda-forge build, so apologies if I am reporting in the wrong place.

Script

conda search gsd -c conda-forge
...
gsd 1.8.0 py27hc1659b7_0 conda-forge
gsd 1.8.0 py36hc1659b7_0 conda-forge
gsd 1.8.0 py36hc1659b7_1 conda-forge
gsd 1.8.0 py37hc1659b7_0 conda-forge
gsd 1.8.0 py37hc1659b7_1 conda-forge

Platform:

Linux

Installation method:

Conda package

Versions:

Python version: 2.7
GSD version: 1.8.0

Developer

Augment schema to address 2D vs 3D spheres and ellipsoids

Description

Currently the dimensionality of spheres and ellipsoids is not part of the GSD shape schema and is only available by inference (by considering the box dimensionality). This ambiguity is undesirable because it makes it impossible for code that only deals with shape (and not simulations) to know what to generate without additional information. I would like to resolve this ambiguity directly within the shape schema.

Proposed solution

We should add additional Disk and Ellipse shape classes to the schema to explicitly represent the 2D cases.

Additional context

When making this change, we will need to make sure we don't break any of HOOMD's behavior. Specifically, we need to see how the ellipse overlap checks in HPMC work in 2D and whether setting c to 0 or removing it entirely causes incorrect overlap detection.

Developer

I can try to implement this, but I would like some discussion and am unlikely to do this in the short term.

Release v2.4.0

Release checklist:

Run bumpversion.
Update change log.
- git log --format=oneline --first-parent `git log -n 1 --pretty=format:%H -- CHANGELOG.rst`...
- milestone
Check readthedocs build, especially change log formatting.
- Build status
- Output
Tag and push.
Update conda-forge recipe.

Rewrite to a chunk

Description

Rewrite into an existing chunk at a specific frame.

Proposed solution

Now users can read data of any frames, while write data can only be implemented by adding extra frames.
Since I am trying to prepress my data, which means read the data at a frame then processed and overwrite that chunk at that frame.

Additional context

N/A

Developer

N/A

gsd/gsd.c is missing #include <sys/stat.h>

Original report by yurivict (Bitbucket: yurivict, GitHub: yurivict).

Without it the build breaks on FreeBSD 11.1:

cc -DNDEBUG -O2 -pipe -fno-omit-frame-pointer -fstack-protector -fno-strict-aliasing -fno-omit-frame-pointer -fPIC -I/usr/local/include/python2.7 -I/usr/local/lib/python2.7/site-packages/numpy/core/include -c gsd/gsd.c -o build/temp.freebsd-11.1-STABLE-amd64-2.7/gsd/gsd.o
gsd/gsd.c:461:69: error: use of undeclared identifier 'S_IRUSR'
    int fd = open(fname, O_RDWR | O_CREAT | O_TRUNC | extra_flags,  S_IRUSR | S_IWUSR | S_IRGRP | S_IWGRP);
                                                                    ^

Add install targets to cmake

Original report by Joshua Anderson (Bitbucket: joaander, GitHub: joaander).

python setup.py fails in strange ways on some systems, we need more control over the install process to allow it to be installed on more systems.

Requires numpy>=1.16.0

I'm not certain this is an issue with your code, but I got an issue in my build while importing gsd. It seems to be because the build version is different than my version of NumPy. It would be nice if you listed a specific version of NumPy if it's required to make it work, though I'm not sure at this point that this is the case.

I couldn't find your build information on circleci to see what version of NumPy it was built with.

Thanks!

../../../virtualenv/python3.6.3/lib/python3.6/site-packages/gsd/hoomd.py:25: in <module>
    from gsd import fl
__init__.pxd:918: in init gsd.fl
    ???
E   ValueError: numpy.ufunc size changed, may indicate binary incompatibility. Expected 216 from C header, got 192 from PyObject

Numpy is requried for installation by pip

Original report by Malcolm Ramsay (Bitbucket: malramsay64, GitHub: malramsay64).

When running

#!bash

pip install gsd

in a new environment the install fails with a numpy import exception. Installing numpy then installing gsd is a workaround for this.

Design hoomd schema

Original report by Joshua Anderson (Bitbucket: joaander, GitHub: joaander).

We need a schema to describe hoomd data.

First pass: Just particles (no rigid bodies or class specific restart data).

Remove deprecated functionality

TODO: Remove deprecated methods in the v2.0 release.

Remove generated code from git repository

Description

Remove the generated fl.c file from the git repository. The file is large, makes pull requests difficult to parse, and makes merge conflicts likely.

Proposed solution

In #37 I take a few steps toward this, enabling the use of Cython on circleci and appveyor. We need to update readthedocs so that it can cythonize code in development for documentation preview builds.

Options for release builds:

Cythonize at the time of tarball creation and use packaged fl.c in setup.py
Always cythonize in setup.py and ignore cython's recommendation to bundle generated source

Developer

Seeking developer with python setuptools/pip experience.

Deprecate `read_frame`

Description

Deprecate the read_frame method in the next feature release.

Additional context

Users should use __getitem__ indexing (e.g. trajectory[i]) instead. read_frame is the private implementation backing this API.

Developer

I will make this change.

Add CLI entrypoint for file reading

Description

I frequently open and inspect GSD files in the Python interpreter for testing. It would be convenient to add some simple entry points to the Python interface to be able to do something like read_gsd FILENAME on the command line. This would be extremely simple, the equivalent of executing python -i read.py where read.py is a Python script like

# read.py
import gsd.hoomd
traj = gsd.hoomd.open(FILENAME)

but possibly with additional flags that could be added for e.g. different schema or opening directly with the file layer.

Proposed solution

This could be implemented using Python's built-in entrypoints function for installation when using setup.py. I'm not sure if the CMake build supports this, that might require some additional changes.

Developer

I'm happy to take this on if there's interest.

GSD files not recognized when running on a big-endian arch

Original report by Dominik Mierzejewski (Bitbucket: rathann, GitHub: rathann).

gsd.hoomd.open() fails with RuntimeError: Not a GSD file: ... when running on a big-endian arch like ppc64 or s390x. This is reproducible when running MDAnalysis 0.17.0 or newer test suite.

I'm trying to investigate on my ppc64 VM and come up with a patch.

Release v2.4.2

Release checklist:

Run bumpversion.
Update change log.
- git log --format=oneline --first-parent `git log -n 1 --pretty=format:%H -- CHANGELOG.rst`...
- milestone
Check readthedocs build, especially change log formatting.
- Build status
- Output
Tag and push.
Update conda-forge recipe.

unable to build docs without network access

Original report by Dominik Mierzejewski (Bitbucket: rathann, GitHub: rathann).

When trying to build documentation offline, I'm getting this error:

$ sphinx-build -b html -d _build/doctrees -a -n -W . _build/html
Running Sphinx v1.6.6
loading pickled environment... not yet created
loading intersphinx inventory from https://docs.python.org/3/objects.inv...

Warning, treated as error:
intersphinx inventory 'https://docs.python.org/3/objects.inv' not fetchable due to <class 'requests.exceptions.ConnectionError'>: HTTPSConnectionPool(host='docs.python.org', port=443): Max retries exceeded with url: /3/objects.inv (Caused by NewConnectionError('<urllib3.connection.VerifiedHTTPSConnection object at 0x7fca0a967198>: Failed to establish a new connection: [Errno -2] Name or service not known',))

I'm trying to create a Fedora package and one of the requirements is that it must be possible to build without network access (Fedora builders are configured like that).

Please either provide HTML docs in a (separate) tarball or a way to build them without network access. I can fetch the necessary files beforehand and put them where they're expected, but the build process must not try to download anything.

Add memory efficient append only mode

Original report by Joshua Anderson (Bitbucket: joaander, GitHub: joaander).

For simulation runs that save a massive number of frames, read/write mode is not desirable or necessary. It stores the large index in memory in its entirety. This is not something we can afford to do on nodes with limited memory.

A hoomd gsd dump writer does not need to store the index in memory to append files. A simple append only mode that only cached unwritten index entries (a few K) would be sufficient. It would be unable to read any data chunks from the file, but the writer should not need to do that, except possibly to verify if time steps are in the file (and maybe that is not even necessary).

The only tricky part is that on load it will need to do the binary search through the index doing just one index entry read() at a time to avoid the need to read in the entire index - or perhaps mmap it temporarily and then unmap.

Release v2.0.0

Release checklist:

Run bumpversion.
Update change log.
- git log --format=oneline --first-parent `git log -n 1 --pretty=format:%H -- CHANGELOG.rst`...
- milestone
Check readthedocs build, especially change log formatting.
- Build status
- Output
Tag and push.
Build tarball.
Update conda-forge recipe.

Compilation of 1.5.2 fails on OpenBSD

Original report by tolejema (Bitbucket: tolejema, ).

Compilation fails on OpenBSD.
It seems that

#include <sys/stat.h>

is required under this platform.
You could try something like:

#include <sys/param.h>
#if defined(BSD)
#include <sys/stat.h>
#endif

Windows Compatibility

Original report by Tyler Reddy (Bitbucket: treddy1, GitHub: treddy1).

This is just to confirm that gsd is not supported on 64-bit Windows?

As we move toward adding Windows support for the MDAnalysis library (which now has a gsd dependency), my observation is that there are some issues using gsd on that platform.

I am not sure if there's any interest on the gsd end for supporting Windows, but see i.e., the discussion in MDAnalysis/mdanalysis#1914 related to the way we will likely have to handle gsd on Windows for the time being.

I'm pretty sure I noticed some low-hanging fruit for compatibility improvement with i.e., explicit '/' being used in path handling in gsd instead of os.path.join, but I'm not sure if it would be a much bigger undertaking to add full support for Windows.

LICENSE is missing in pypi release

Original report by Max Linke (Bitbucket: max_linke, ).

The license file isn't included in the pypi release currently. To add it the MANIFEST.in should contain the following line.

include LICENSE

glotzerlab / gsd Goto Github PK

gsd's People

Contributors

Stargazers

Watchers

Forkers

gsd's Issues

Description

Proposed solution

Additional context

Developer

Description

Developer

Description

Proposed solution

Additional context

Developer

Description

Script

Output

Expected output

Configuration

Developer

Description

Developer

Description

Proposed solution

Additional context

Developer

Description

Failed Outputs

Configuration

Developer

Description

Script

Output

Expected output

Configuration

Developer

Description

Script

Output

Expected output

Configuration

Developer

Description

Script

Developer

Description

Configuration

Developer

Description

Script

Developer

Description

Proposed solution

Additional context

Developer

Description

Proposed solution

Additional context

Developer

Description

Proposed solution

Developer

Description

Additional context

Developer

Description

Proposed solution

Developer

Recommend Projects

Recommend Topics

Recommend Org