From sebastian at sipsolutions.net  Tue Aug  2 13:22:34 2016
From: sebastian at sipsolutions.net (Sebastian Berg)
Date: Tue, 02 Aug 2016 19:22:34 +0200
Subject: [Numpy-discussion] Euroscipy
Message-ID: <1470158553.5372.2.camel@sipsolutions.net>

Hi all,

I am still pondering whether or not (and if which days) to go to
EuroScipy. Who else is there and would like to discuss a bit or
whatever else?

- Sebastian
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160802/fa47fda9/attachment.sig>

From siegfried.gonzi at ed.ac.uk  Tue Aug  2 13:41:21 2016
From: siegfried.gonzi at ed.ac.uk (Siegfried Gonzi)
Date: Tue, 2 Aug 2016 18:41:21 +0100
Subject: [Numpy-discussion] scipy curve_fit variable list of optimisation
	parameters
Message-ID: <F44130F6-80CC-45E6-9AD1-467A769E5FDE@ed.ac.uk>

Hi all

Does anyone know how to invoke curve_fit with a variable number of parameters, e.g. a1 to a10 without writing it out,

e.g.

def func2( x, a1,a2,a3,a4 ):

        # Bessel function
	tmp = scipy.special.j0( x[:,:] )

	return np.dot( tmp[:,:] , np.array( [a1,a2,a3,a4] )


### yi = M measurements (.e.g M=20)
### x = M (=20) rows of N (=4) columns
popt = scipy.optimize.curve_fit( func2, x, yi )

I'd like to get *1 single vector* (in this case of size 4) of optimised A(i) values.

The function I am trying to minimise (.e.g F(r) is a vector of 20 model measurements): F(r) = SUM_i_to_N [ A(i) * bessel_function_J0(i * r) ] 


Thanks,
Siegfried Gonzi




-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.



From evgeny.burovskiy at gmail.com  Tue Aug  2 17:50:42 2016
From: evgeny.burovskiy at gmail.com (Evgeni Burovski)
Date: Tue, 2 Aug 2016 22:50:42 +0100
Subject: [Numpy-discussion] scipy curve_fit variable list of
	optimisation parameters
In-Reply-To: <F44130F6-80CC-45E6-9AD1-467A769E5FDE@ed.ac.uk>
References: <F44130F6-80CC-45E6-9AD1-467A769E5FDE@ed.ac.uk>
Message-ID: <CAMRo0ivCS8HqR6Pyo9-wvnhboPGYzaFwfZNmVwE5Q5NTXna9Dg@mail.gmail.com>

On Tue, Aug 2, 2016 at 6:41 PM, Siegfried Gonzi
<siegfried.gonzi at ed.ac.uk> wrote:
> Hi all
>
> Does anyone know how to invoke curve_fit with a variable number of parameters, e.g. a1 to a10 without writing it out,
>
> e.g.
>
> def func2( x, a1,a2,a3,a4 ):
>
>         # Bessel function
>         tmp = scipy.special.j0( x[:,:] )
>
>         return np.dot( tmp[:,:] , np.array( [a1,a2,a3,a4] )
>
>
> ### yi = M measurements (.e.g M=20)
> ### x = M (=20) rows of N (=4) columns
> popt = scipy.optimize.curve_fit( func2, x, yi )
>
> I'd like to get *1 single vector* (in this case of size 4) of optimised A(i) values.
>
> The function I am trying to minimise (.e.g F(r) is a vector of 20 model measurements): F(r) = SUM_i_to_N [ A(i) * bessel_function_J0(i * r) ]
>
>
> Thanks,
> Siegfried Gonzi
>
>
>
>
> --
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.

You can use `leastsq` or `least_squares` directly: they both accept an
array of parameters.

BTW, since all of these functions are actually in scipy, you might
want to redirect this discussion to the scipy-user mailing list.

Cheers,

Evgeni


From ralf.gommers at gmail.com  Wed Aug  3 03:50:00 2016
From: ralf.gommers at gmail.com (Ralf Gommers)
Date: Wed, 3 Aug 2016 09:50:00 +0200
Subject: [Numpy-discussion] Euroscipy
In-Reply-To: <1470158553.5372.2.camel@sipsolutions.net>
References: <1470158553.5372.2.camel@sipsolutions.net>
Message-ID: <CABL7CQjTgo0yDvasZ5vP6jgs0Sa0-S-SY7-Gr=YwU_Mj44wPKg@mail.gmail.com>

On Tue, Aug 2, 2016 at 7:22 PM, Sebastian Berg <sebastian at sipsolutions.net>
wrote:

> Hi all,
>
> I am still pondering whether or not (and if which days) to go to
> EuroScipy. Who else is there and would like to discuss a bit or
> whatever else?
>

Unfortunately I'm not able to go this year.

Ralf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160803/6cff0ce2/attachment.html>

From matthew.gidden at gmail.com  Wed Aug  3 04:11:24 2016
From: matthew.gidden at gmail.com (Matthew Gidden)
Date: Wed, 3 Aug 2016 10:11:24 +0200
Subject: [Numpy-discussion] Euroscipy
In-Reply-To: <CABL7CQjTgo0yDvasZ5vP6jgs0Sa0-S-SY7-Gr=YwU_Mj44wPKg@mail.gmail.com>
References: <1470158553.5372.2.camel@sipsolutions.net>
 <CABL7CQjTgo0yDvasZ5vP6jgs0Sa0-S-SY7-Gr=YwU_Mj44wPKg@mail.gmail.com>
Message-ID: <CAHRdPFzEEHFZWUhevSdtJV2QwSX6S4KvEJSJQ0NG9rtda+KiyA@mail.gmail.com>

Hi all,

I've been lurking on the list, but I'll be at EuroScipy and would very much
enjoy meeting any others and getting more involved with the community.

Cheers,
Matt

On Wed, Aug 3, 2016 at 9:50 AM, Ralf Gommers <ralf.gommers at gmail.com> wrote:

>
>
> On Tue, Aug 2, 2016 at 7:22 PM, Sebastian Berg <sebastian at sipsolutions.net
> > wrote:
>
>> Hi all,
>>
>> I am still pondering whether or not (and if which days) to go to
>> EuroScipy. Who else is there and would like to discuss a bit or
>> whatever else?
>>
>
> Unfortunately I'm not able to go this year.
>
> Ralf
>
>
>
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160803/8c214d0f/attachment.html>

From siegfried.gonzi at ed.ac.uk  Wed Aug  3 08:53:09 2016
From: siegfried.gonzi at ed.ac.uk (Siegfried Gonzi)
Date: Wed, 3 Aug 2016 13:53:09 +0100
Subject: [Numpy-discussion] scipy curve_fit variable list
	of	optimisation parameters
In-Reply-To: <mailman.5.1470225601.5005.numpy-discussion@scipy.org>
References: <mailman.5.1470225601.5005.numpy-discussion@scipy.org>
Message-ID: <3088B153-9E2A-4ABD-9A6B-3A46B8CBB441@ed.ac.uk>


On 3 Aug 2016, at 13:00, numpy-discussion-request at scipy.org wrote:

> Message: 3
> Date: Tue, 2 Aug 2016 22:50:42 +0100
> From: Evgeni Burovski <evgeny.burovskiy at gmail.com>
> To: Discussion of Numerical Python <numpy-discussion at scipy.org>
> Subject: Re: [Numpy-discussion] scipy curve_fit variable list of
> 	optimisation parameters
> Message-ID:
> 	<CAMRo0ivCS8HqR6Pyo9-wvnhboPGYzaFwfZNmVwE5Q5NTXna9Dg at mail.gmail.com>
> Content-Type: text/plain; charset=UTF-8
> 
> 
> You can use `leastsq` or `least_squares` directly: they both accept an
> array of parameters.
> 
> BTW, since all of these functions are actually in scipy, you might
> want to redirect this discussion to the scipy-user mailing list.


Hi all

I found the solution in the following thread:

http://stackoverflow.com/questions/28969611/multiple-arguments-in-python

One has to call curve_fit with 'p0' (giving curve_fit a clue about the unknown number of variables)

I changed func2 to (note the *):

def func2( x, *a ):

       # Bessel function
	tmp = scipy.special.j0( x[:,:] )

	return np.dot( tmp[:,:] , a[:] )


and call it:

N = number of optimisation parameters

popt = scipy.optimize.curve_fit( func2, x, yi , p0=[1.0]*N)



Regards,
Siegfried Gonzi
Met Office, Exeter, UK
-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.



From charlesr.harris at gmail.com  Wed Aug  3 16:09:32 2016
From: charlesr.harris at gmail.com (Charles R Harris)
Date: Wed, 3 Aug 2016 14:09:32 -0600
Subject: [Numpy-discussion] Numpy 1.11.2
Message-ID: <CAB6mnxKz0EEGMgDSVqRtT8V+9VMP7yAU1pHo1K36OgkXS4PX0g@mail.gmail.com>

Hi All,

I would like to release Numpy 1.11.2rc1 this weekend. It will contain a few
small fixes and enhancements for windows and the last Scipy release. If
there are any pending PRs that you think should go in or be backported for
this release, please speak up.

Chuck
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160803/e913d870/attachment.html>

From matti.picus at gmail.com  Fri Aug  5 03:06:02 2016
From: matti.picus at gmail.com (Matti Picus)
Date: Fri, 5 Aug 2016 10:06:02 +0300
Subject: [Numpy-discussion] test_closing_fid (in test_io.py) on PyPy
Message-ID: <7888be5c-52eb-4d6d-00b5-b31128f8cd0d@gmail.com>

test_closing_fid essentially calls this to ensure close() is called when 
a NpzFile object goes out of context:

for i in range(1, 1025):

     np.load(tmp)["data"]


This raises a ResourceWarning on python 3, and fails on pypy since the 
garbage collector works differently.

It seems to be a classic example of wrongly using np.load(), but I'm 
sure it is a common use case.

I can submit a pull request to skip on pypy, or should this be solved in 
a more substantial way?

Matti



From florin.papa at intel.com  Fri Aug  5 03:42:35 2016
From: florin.papa at intel.com (Papa, Florin)
Date: Fri, 5 Aug 2016 07:42:35 +0000
Subject: [Numpy-discussion] NumPy in PyPy
Message-ID: <3A375A669FBEFF45B6B60E689636EDCA09C0BDDA@IRSMSX101.ger.corp.intel.com>

Hi,

This is Florin Papa from the Dynamic Scripting Languages Optimizations team in Intel Corporation.

Our team is working on optimizing the PyPy interpreter and part of this work is to find and fix incompatibilities between NumPy and PyPy. Does anyone have knowledge of real life workloads that use NumPy and cannot be run using PyPy?

We are also interested in creating a repository with relevant benchmarks for real world usage of NumPy, like GUPB for CPython, but we have not found such workloads for NumPy.

Thank you,
Florin
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160805/bf18a61a/attachment.html>

From pav at iki.fi  Fri Aug  5 13:35:44 2016
From: pav at iki.fi (Pauli Virtanen)
Date: Fri, 5 Aug 2016 17:35:44 +0000 (UTC)
Subject: [Numpy-discussion] test_closing_fid (in test_io.py) on PyPy
References: <7888be5c-52eb-4d6d-00b5-b31128f8cd0d@gmail.com>
Message-ID: <no2ipf$8ks$1@blaine.gmane.org>

Fri, 05 Aug 2016 10:06:02 +0300, Matti Picus kirjoitti:
[clip]
> I can submit a pull request to skip on pypy, or should this be solved in
> a more substantial way?

Should also be safe to just skip it on Pypy, it's testing that the wrong 
way to use np.load also works on CPython.

-- 
Pauli Virtanen



From ben.v.root at gmail.com  Fri Aug  5 17:20:48 2016
From: ben.v.root at gmail.com (Benjamin Root)
Date: Fri, 5 Aug 2016 17:20:48 -0400
Subject: [Numpy-discussion] NumPy in PyPy
In-Reply-To: <3A375A669FBEFF45B6B60E689636EDCA09C0BDDA@IRSMSX101.ger.corp.intel.com>
References: <3A375A669FBEFF45B6B60E689636EDCA09C0BDDA@IRSMSX101.ger.corp.intel.com>
Message-ID: <CANNq6Fm8cJMAiMGUjdumm5cKvqFdJbXbf+hCBXXjUe82mjxgJw@mail.gmail.com>

Don't know if it is what you are looking for, but NumPy has a built-in
suite of benchmarks:
http://docs.scipy.org/doc/numpy/reference/generated/numpy.testing.Tester.bench.html

Also, some projects have taken to utilizing the "airspeed velocity" utility
to track benchmarking stats for their projects. I know astropy utilizes it.
So, maybe their benchmarks might be a good starting point since they
utilize numpy heavily?

Cheers!
Ben Root


On Fri, Aug 5, 2016 at 3:42 AM, Papa, Florin <florin.papa at intel.com> wrote:

> Hi,
>
>
>
> This is Florin Papa from the Dynamic Scripting Languages Optimizations
> team in Intel Corporation.
>
>
>
> Our team is working on optimizing the PyPy interpreter and part of this
> work is to find and fix incompatibilities between NumPy and PyPy. Does
> anyone have knowledge of real life workloads that use NumPy and cannot be
> run using PyPy?
>
>
>
> We are also interested in creating a repository with relevant benchmarks
> for real world usage of NumPy, like GUPB for CPython, but we have not found
> such workloads for NumPy.
>
>
>
> Thank you,
>
> Florin
>
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160805/d5048137/attachment.html>

From ralf.gommers at gmail.com  Sat Aug  6 04:05:24 2016
From: ralf.gommers at gmail.com (Ralf Gommers)
Date: Sat, 6 Aug 2016 20:05:24 +1200
Subject: [Numpy-discussion] NumPy in PyPy
In-Reply-To: <CANNq6Fm8cJMAiMGUjdumm5cKvqFdJbXbf+hCBXXjUe82mjxgJw@mail.gmail.com>
References: <3A375A669FBEFF45B6B60E689636EDCA09C0BDDA@IRSMSX101.ger.corp.intel.com>
 <CANNq6Fm8cJMAiMGUjdumm5cKvqFdJbXbf+hCBXXjUe82mjxgJw@mail.gmail.com>
Message-ID: <CABL7CQja0caTYJnvf24utm9g21+FPviO1Bt6puDLN7dN7cRu6A@mail.gmail.com>

On Sat, Aug 6, 2016 at 9:20 AM, Benjamin Root <ben.v.root at gmail.com> wrote:

> Don't know if it is what you are looking for, but NumPy has a built-in
> suite of benchmarks: http://docs.scipy.org/doc/numpy/reference/generated/
> numpy.testing.Tester.bench.html
>

That's the very old (now unused) benchmark runner. Numpy has had an ASV
test suite for a while, see
https://github.com/numpy/numpy/tree/master/benchmarks for how to run it.


> Also, some projects have taken to utilizing the "airspeed velocity"
> utility to track benchmarking stats for their projects. I know astropy
> utilizes it. So, maybe their benchmarks might be a good starting point
> since they utilize numpy heavily?
>
> Cheers!
> Ben Root
>
>
> On Fri, Aug 5, 2016 at 3:42 AM, Papa, Florin <florin.papa at intel.com>
> wrote:
>
>> Hi,
>>
>>
>>
>> This is Florin Papa from the Dynamic Scripting Languages Optimizations
>> team in Intel Corporation.
>>
>>
>>
>> Our team is working on optimizing the PyPy interpreter and part of this
>> work is to find and fix incompatibilities between NumPy and PyPy. Does
>> anyone have knowledge of real life workloads that use NumPy and cannot be
>> run using PyPy?
>>
>>
>>
>> We are also interested in creating a repository with relevant benchmarks
>> for real world usage of NumPy, like GUPB for CPython, but we have not found
>> such workloads for NumPy.
>>
>
The approach of GUPB is interesting (the whole application part that is,
the rest looks much more cumbersome than ASV benchmarks), but of course
easier to create for Python than for Numpy. You'd need to find whole
applications that spend most of their time in numpy but not in too small a
set of numpy functions. Maybe benchmark suites of other projects aren't
such a bad idea for that. Or spend a bit of time collecting relevant
published ipython notebooks.

Ralf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160806/c70099b0/attachment.html>

From akborder at gmail.com  Mon Aug  8 09:11:17 2016
From: akborder at gmail.com (Anakim Border)
Date: Mon, 8 Aug 2016 15:11:17 +0200
Subject: [Numpy-discussion] Views and Increments
Message-ID: <CAFurOt+j=8HgLkmeAbx6ZD=M+h3WUs_VX2YJ_USFwxdLUNB30A@mail.gmail.com>

Dear List,

I'm experimenting with views and array indexing. I have written two code
blocks that I was expecting to produce the same result.

First try:

>>> a = np.arange(10)
>>> b = a[np.array([1,6,5])]
>>> b += 1
>>> a
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])


Alternative version:

>>> a = np.arange(10)
>>> a[np.array([1,6,5])] += 1
>>> a
array([0, 2, 2, 3, 4, 6, 7, 7, 8, 9])


I understand what is happening in the first case. In fact, the
documentation is quite clear on the subject:

For all cases of index arrays, what is returned is a copy of the original
data, not a view as one gets for slices.

What about the second case? There, I'm not keeping a reference to the
intermediate copy (b, in the first example). Still, I don't see why the
update (to the copy) is propagating to the original array. Is there any
implementation detail that I'm missing?

Best,
  ab
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160808/51dd5fc3/attachment.html>

From sebastian at sipsolutions.net  Mon Aug  8 09:21:15 2016
From: sebastian at sipsolutions.net (Sebastian Berg)
Date: Mon, 08 Aug 2016 15:21:15 +0200
Subject: [Numpy-discussion] Views and Increments
In-Reply-To: <CAFurOt+j=8HgLkmeAbx6ZD=M+h3WUs_VX2YJ_USFwxdLUNB30A@mail.gmail.com>
References: <CAFurOt+j=8HgLkmeAbx6ZD=M+h3WUs_VX2YJ_USFwxdLUNB30A@mail.gmail.com>
Message-ID: <1470662475.20080.1.camel@sipsolutions.net>

On Mo, 2016-08-08 at 15:11 +0200, Anakim Border wrote:
> Dear List,
> 
> I'm experimenting with views and array indexing. I have written two
> code blocks that I was expecting to produce the same result.
> 
> First try:
> 
> >>> a = np.arange(10)
> >>> b = a[np.array([1,6,5])]
> >>> b += 1
> >>> a
> array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
> 
> 
> Alternative version:
> 
> >>> a = np.arange(10)
> >>> a[np.array([1,6,5])] += 1
> >>> a
> array([0, 2, 2, 3, 4, 6, 7, 7, 8, 9])
> 
> 
> I understand what is happening in the first case. In fact, the
> documentation is quite clear on the subject:
> 
> For all cases of index arrays, what is returned is a copy of the
> original data, not a view as one gets for slices.
> 
> What about the second case? There, I'm not keeping a reference to the
> intermediate copy (b, in the first example). Still, I don't see why
> the update (to the copy) is propagating to the original array. Is
> there any implementation detail that I'm missing?
> 

The second case translates to:

tmp =?a[np.array([1,6,5])] + 1
a[np.array([1,6,5])] = tmp

this is done by python, without any interplay of numpy at all. Which is
different from `arr += 1`, which is specifically defined and translates
to `np.add(arr, 1, out=arr)`.

- Sebastian


> Best,
> ? ab
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160808/f15caa1e/attachment.sig>

From bennyrowland at mac.com  Mon Aug  8 11:35:25 2016
From: bennyrowland at mac.com (Ben Rowland)
Date: Mon, 08 Aug 2016 11:35:25 -0400
Subject: [Numpy-discussion] PR 7918 apply_along_axis()
Message-ID: <B77CC7BA-3F20-4518-86FF-911D141E6EC8@mac.com>

Hi list,

This is both my first post to this list and first pull request for numpy so apologies if this is not the right list or for any other mistakes.

Here is the link to my pull request: https://github.com/numpy/numpy/pull/7918 <https://github.com/numpy/numpy/pull/7918>

It is a simple modification to the numpy.apply_along_axis() function in numpy/lib/shape_base.py to allow it to deal better with ndarray subclasses. My use case is this: I have an ndarray subclass storing sets of temporal data which has additional metadata properties such as the time between data points. I frequently want to apply the same function to each time course in a set and apply_along_axis() is a compact and efficient way to do this. However there are two behaviours that I require which are not provided by the current numpy master:
1) apply_along_axis returns the same subclass as it was called upon, so that I don?t lose all the metadata that it went in with.
2) fund1d calls inside apply_along_axis should receive 1d slices of the same subclass as the supplied whole array, so that they can make use of the metadata in performing their function.

To achieve these two behaviours requires modifying only three lines of code:
1) At the start of the function, the input is converted to a bare ndarray by the function numpy.asarray(), I replace this with the subclass friendly numpy.asanyarray()
2) Depending on whether func1d returns a scalar or a vector there are two different code paths constructing and returning the output array. In each case I call __array_wrap__ on the input array, passing the output array to allow it to be updated with all the metadata of the original array.

I have also implemented two new tests for this functionality. The first is very simple and merely calls apply_along_axis on a numpy.matrix class and checks that the result is also a numpy.matrix. The second is slightly more involved as it requires defining a MinimalSubclass class which adds a data member to the bare ndarray, then a function which returns that member. apply_along_axis() is then called passing in an instance of the subclass and the function to show that the slices passed to func1d preserve the data member.

I would welcome any comment or constructive criticism.

Ben Rowland
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160808/f3ced424/attachment.html>

From m.h.vankerkwijk at gmail.com  Mon Aug  8 12:32:36 2016
From: m.h.vankerkwijk at gmail.com (Marten van Kerkwijk)
Date: Mon, 8 Aug 2016 12:32:36 -0400
Subject: [Numpy-discussion] Views and Increments
In-Reply-To: <CAFurOt+j=8HgLkmeAbx6ZD=M+h3WUs_VX2YJ_USFwxdLUNB30A@mail.gmail.com>
References: <CAFurOt+j=8HgLkmeAbx6ZD=M+h3WUs_VX2YJ_USFwxdLUNB30A@mail.gmail.com>
Message-ID: <CAJNV+9t=4Dk3p94HOdoA-NOytD1XeRY4hRzwXPtKBVb2s8oqng@mail.gmail.com>

Hi Anakim,

The difference is really in the code path that gets taken: in the first
case, you go through `a.__getitem__(np.array([1,6,5])`, in the second
through `a.__setitem__(...)`. The increments would not work if you added an
extra indexing to it, as in:
```
a[np.array([1,6,5])][:] += 1
```

?(which would do `a.__getitem__(...).__setitem__(slice(None))`)

Hope this clarifies it,

Marten
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160808/2c159f40/attachment.html>

From shoyer at gmail.com  Mon Aug  8 16:08:46 2016
From: shoyer at gmail.com (Stephan Hoyer)
Date: Mon, 8 Aug 2016 13:08:46 -0700
Subject: [Numpy-discussion] Views and Increments
In-Reply-To: <CAFurOt+j=8HgLkmeAbx6ZD=M+h3WUs_VX2YJ_USFwxdLUNB30A@mail.gmail.com>
References: <CAFurOt+j=8HgLkmeAbx6ZD=M+h3WUs_VX2YJ_USFwxdLUNB30A@mail.gmail.com>
Message-ID: <CAEQ_Tvf3ZKko8GMST4C0RdRVzMfFLfrB3f7jqZC1Ez=RQBNkvQ@mail.gmail.com>

On Mon, Aug 8, 2016 at 6:11 AM, Anakim Border <akborder at gmail.com> wrote:

> Alternative version:
>
> >>> a = np.arange(10)
> >>> a[np.array([1,6,5])] += 1
> >>> a
> array([0, 2, 2, 3, 4, 6, 7, 7, 8, 9])
>

I haven't checked, but a likely explanation is that Python itself
interprets a[b] += c as a[b] = a[b] + c.

Python has special methods for inplace assignment (__setitem__) and inplace
arithmetic (__iadd__) but no special methods for inplace arithmetic and
assignment at the same time, so this is really out of NumPy's control here.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160808/cf81b142/attachment.html>

From chris.barker at noaa.gov  Mon Aug  8 18:11:37 2016
From: chris.barker at noaa.gov (Chris Barker)
Date: Mon, 8 Aug 2016 15:11:37 -0700
Subject: [Numpy-discussion] NumPy in PyPy
In-Reply-To: <CABL7CQja0caTYJnvf24utm9g21+FPviO1Bt6puDLN7dN7cRu6A@mail.gmail.com>
References: <3A375A669FBEFF45B6B60E689636EDCA09C0BDDA@IRSMSX101.ger.corp.intel.com>
 <CANNq6Fm8cJMAiMGUjdumm5cKvqFdJbXbf+hCBXXjUe82mjxgJw@mail.gmail.com>
 <CABL7CQja0caTYJnvf24utm9g21+FPviO1Bt6puDLN7dN7cRu6A@mail.gmail.com>
Message-ID: <CALGmxEL2s5tgEpUyGi1AEKKMdH1jVx_mawOF0W7sxP6653Cjuw@mail.gmail.com>

>
> On Fri, Aug 5, 2016 at 3:42 AM, Papa, Florin <florin.papa at intel.com>
>> wrote:
>>
>>>  Does anyone have knowledge of real life workloads that use NumPy and
>>> cannot be run using PyPy?
>>>
>>>
>>>
>>> We are also interested in creating a repository with relevant benchmarks
>>> for real world usage of NumPy,
>>>
>>
We have a numpy --  heavy app. bu tit, like many others, I'm sure, also
relies heavily on Cython-wrapped C++ code, as well as pure Cython
extensions.

As well as many other packages that are also wrappers around C libs, Cython
-optimized, etc.

I've never tried to run it under PyPy I've always assumed it's a
non-starter.

Is there any hope?


If you are curious:

https://github.com/NOAA-ORR-ERD/PyGnome

-CHB



-- 

Christopher Barker, Ph.D.
Oceanographer

Emergency Response Division
NOAA/NOS/OR&R            (206) 526-6959   voice
7600 Sand Point Way NE   (206) 526-6329   fax
Seattle, WA  98115       (206) 526-6317   main reception

Chris.Barker at noaa.gov
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160808/8398d896/attachment.html>

From florin.papa at intel.com  Tue Aug  9 06:49:14 2016
From: florin.papa at intel.com (Papa, Florin)
Date: Tue, 9 Aug 2016 10:49:14 +0000
Subject: [Numpy-discussion] NumPy in PyPy
In-Reply-To: <CALGmxEL2s5tgEpUyGi1AEKKMdH1jVx_mawOF0W7sxP6653Cjuw@mail.gmail.com>
References: <3A375A669FBEFF45B6B60E689636EDCA09C0BDDA@IRSMSX101.ger.corp.intel.com>
 <CANNq6Fm8cJMAiMGUjdumm5cKvqFdJbXbf+hCBXXjUe82mjxgJw@mail.gmail.com>
 <CABL7CQja0caTYJnvf24utm9g21+FPviO1Bt6puDLN7dN7cRu6A@mail.gmail.com>
 <CALGmxEL2s5tgEpUyGi1AEKKMdH1jVx_mawOF0W7sxP6653Cjuw@mail.gmail.com>
Message-ID: <3A375A669FBEFF45B6B60E689636EDCA09C0C954@IRSMSX101.ger.corp.intel.com>

> We have a numpy --  heavy app. bu tit, like many others, I'm sure, also relies heavily on Cython-wrapped C++ code, as well as pure Cython extensions.
> 
> As well as many other packages that are also wrappers around C libs, Cython -optimized, etc.
> 
> I've never tried to run it under PyPy I've always assumed it's a non-starter.
> 
> Is there any hope?
> 
> 
> If you are curious:
> 
> https://github.com/NOAA-ORR-ERD/PyGnome
> 
> -CHB

Hi Christopher,

Thank you for the information, I will investigate whether we can use PyPy for this workload.

Regards,
Florin

From florin.papa at intel.com  Tue Aug  9 07:03:11 2016
From: florin.papa at intel.com (Papa, Florin)
Date: Tue, 9 Aug 2016 11:03:11 +0000
Subject: [Numpy-discussion] NumPy in PyPy
In-Reply-To: <CABL7CQja0caTYJnvf24utm9g21+FPviO1Bt6puDLN7dN7cRu6A@mail.gmail.com>
References: <3A375A669FBEFF45B6B60E689636EDCA09C0BDDA@IRSMSX101.ger.corp.intel.com>
 <CANNq6Fm8cJMAiMGUjdumm5cKvqFdJbXbf+hCBXXjUe82mjxgJw@mail.gmail.com>
 <CABL7CQja0caTYJnvf24utm9g21+FPviO1Bt6puDLN7dN7cRu6A@mail.gmail.com>
Message-ID: <3A375A669FBEFF45B6B60E689636EDCA09C0C975@IRSMSX101.ger.corp.intel.com>


On Sat, Aug 6, 2016 at 9:20 AM, Benjamin Root <ben.v.root at gmail.com> wrote:
> Don't know if it is what you are looking for, but NumPy has a built-in suite of benchmarks: http://docs.scipy.org/doc/numpy/reference/generated/numpy.testing.Tester.bench.html

> That's the very old (now unused) benchmark runner. Numpy has had an ASV test suite for a while, see https://github.com/numpy/numpy/tree/master/benchmarks for how to run it.
?
> Also, some projects have taken to utilizing the "airspeed velocity" utility to track benchmarking stats for their projects. I know astropy utilizes it. So, maybe their benchmarks might be a good starting point since they utilize numpy heavily?

> Cheers!
> Ben Root

>> On Fri, Aug 5, 2016 at 3:42 AM, Papa, Florin <florin.papa at intel.com> wrote:
>> Hi,
?
>> This is Florin Papa from the Dynamic Scripting Languages Optimizations team in Intel Corporation.
?
>> Our team is working on optimizing the PyPy interpreter and part of this work is to find and fix incompatibilities between NumPy and PyPy. Does anyone have knowledge of real life workloads that use NumPy and cannot be run using PyPy?
?
>> We are also interested in creating a repository with relevant benchmarks for real world usage of NumPy, like GUPB for CPython, but we have not found such workloads for NumPy.
?
>> The approach of GUPB is interesting (the whole application part that is, the rest looks much more cumbersome than ASV benchmarks), but of course easier to create for Python than for Numpy. You'd need to find whole applications that spend most of their time in numpy but not in too small a set of numpy functions. Maybe benchmark suites of other projects aren't such a bad idea for that. Or spend a bit of time collecting relevant published ipython notebooks.
>> Ralf

Astropy definitely looks like a good candidate for a real life workload. Thank you also for the useful information on ASV.

Regards,
Florin

From morph at debian.org  Sun Aug 14 13:11:08 2016
From: morph at debian.org (Sandro Tosi)
Date: Sun, 14 Aug 2016 18:11:08 +0100
Subject: [Numpy-discussion] Numpy 1.11.2
In-Reply-To: <CAB6mnxKz0EEGMgDSVqRtT8V+9VMP7yAU1pHo1K36OgkXS4PX0g@mail.gmail.com>
References: <CAB6mnxKz0EEGMgDSVqRtT8V+9VMP7yAU1pHo1K36OgkXS4PX0g@mail.gmail.com>
Message-ID: <CAB4XWXzQf2cOLjMNN3qSowJUHQ+40aaVFSF505EnkZFp=V_cSQ@mail.gmail.com>

hey there, what happened here? do you still plan to release a 1.11.2rc1 soon?

On Wed, Aug 3, 2016 at 9:09 PM, Charles R Harris
<charlesr.harris at gmail.com> wrote:
> Hi All,
>
> I would like to release Numpy 1.11.2rc1 this weekend. It will contain a few
> small fixes and enhancements for windows and the last Scipy release. If
> there are any pending PRs that you think should go in or be backported for
> this release, please speak up.
>
> Chuck
>
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>



-- 
Sandro "morph" Tosi
My website: http://sandrotosi.me/
Me at Debian: http://wiki.debian.org/SandroTosi
G+: https://plus.google.com/u/0/+SandroTosi


From charlesr.harris at gmail.com  Sun Aug 14 20:12:33 2016
From: charlesr.harris at gmail.com (Charles R Harris)
Date: Sun, 14 Aug 2016 18:12:33 -0600
Subject: [Numpy-discussion] Numpy 1.11.2
In-Reply-To: <CAB4XWXzQf2cOLjMNN3qSowJUHQ+40aaVFSF505EnkZFp=V_cSQ@mail.gmail.com>
References: <CAB6mnxKz0EEGMgDSVqRtT8V+9VMP7yAU1pHo1K36OgkXS4PX0g@mail.gmail.com>
 <CAB4XWXzQf2cOLjMNN3qSowJUHQ+40aaVFSF505EnkZFp=V_cSQ@mail.gmail.com>
Message-ID: <CAB6mnx+ZnM-D1U8S6w6nZ7o6RWZ3TiVTdFDaZGbfpshCQHETRw@mail.gmail.com>

"Events, dear boy, events" ;) There were a couple of bugs that turned up at
the last moment that needed fixing. At the moment there are two, possibly
three, bugs that need finishing off.

   - A fix for compilation on PPC running RHEL 7.2 (done, but not verified)
   - Roll back Numpy reload error: more than one project was reloading.
   - Maybe fix crash for quicksort of object arrays with bogus comparison.

Chuck


On Sun, Aug 14, 2016 at 11:11 AM, Sandro Tosi <morph at debian.org> wrote:

> hey there, what happened here? do you still plan to release a 1.11.2rc1
> soon?
>
> On Wed, Aug 3, 2016 at 9:09 PM, Charles R Harris
> <charlesr.harris at gmail.com> wrote:
> > Hi All,
> >
> > I would like to release Numpy 1.11.2rc1 this weekend. It will contain a
> few
> > small fixes and enhancements for windows and the last Scipy release. If
> > there are any pending PRs that you think should go in or be backported
> for
> > this release, please speak up.
> >
> > Chuck
> >
> > _______________________________________________
> > NumPy-Discussion mailing list
> > NumPy-Discussion at scipy.org
> > https://mail.scipy.org/mailman/listinfo/numpy-discussion
> >
>
>
>
> --
> Sandro "morph" Tosi
> My website: http://sandrotosi.me/
> Me at Debian: http://wiki.debian.org/SandroTosi
> G+: https://plus.google.com/u/0/+SandroTosi
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160814/49e3b8f5/attachment.html>

From alan.isaac at gmail.com  Mon Aug 15 15:04:01 2016
From: alan.isaac at gmail.com (Alan Isaac)
Date: Mon, 15 Aug 2016 15:04:01 -0400
Subject: [Numpy-discussion] weighted random choice in Python
Message-ID: <1bf63320-0881-8097-a9a8-46c6a6ba6b7e@gmail.com>

It seems that there may soon be movement on this enhancement request:
http://bugs.python.org/issue18844
The API is currently under discussion.
If this decision might interact with NumPy in any way,
it would be good to have that documented now.
(See Raymond Hettinger's comment today.)

Cheers,
Alan Isaac


From ralf.gommers at gmail.com  Tue Aug 16 04:06:32 2016
From: ralf.gommers at gmail.com (Ralf Gommers)
Date: Tue, 16 Aug 2016 20:06:32 +1200
Subject: [Numpy-discussion] weighted random choice in Python
In-Reply-To: <1bf63320-0881-8097-a9a8-46c6a6ba6b7e@gmail.com>
References: <1bf63320-0881-8097-a9a8-46c6a6ba6b7e@gmail.com>
Message-ID: <CABL7CQhsP9OjoXP0AMso7ZQCjWovmCiumhsV_Xam792nT4Z94Q@mail.gmail.com>

On Tue, Aug 16, 2016 at 7:04 AM, Alan Isaac <alan.isaac at gmail.com> wrote:

> It seems that there may soon be movement on this enhancement request:
> http://bugs.python.org/issue18844
> The API is currently under discussion.
> If this decision might interact with NumPy in any way,
> it would be good to have that documented now.
> (See Raymond Hettinger's comment today.)
>

Doesn't look like it interacts with numpy.random. The whole enhancement
request doesn't look very interesting imho.

Ralf
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160816/5ac7744d/attachment.html>

From mail at telenczuk.pl  Tue Aug 16 08:28:37 2016
From: mail at telenczuk.pl (mail at telenczuk.pl)
Date: Tue, 16 Aug 2016 14:28:37 +0200
Subject: [Numpy-discussion] Euroscipy
In-Reply-To: <CAHRdPFzEEHFZWUhevSdtJV2QwSX6S4KvEJSJQ0NG9rtda+KiyA@mail.gmail.com>
References: <1470158553.5372.2.camel@sipsolutions.net>
 <CABL7CQjTgo0yDvasZ5vP6jgs0Sa0-S-SY7-Gr=YwU_Mj44wPKg@mail.gmail.com>
 <CAHRdPFzEEHFZWUhevSdtJV2QwSX6S4KvEJSJQ0NG9rtda+KiyA@mail.gmail.com>
Message-ID: <57b306f53aebe_44ce14f7e54c9@Pct-EqAlain-Z30.notmuch>

Hi all,

I will present advanced NumPy tutorial at EuroScipy next week [1]. Is anyone going to attend the tutorials on Tuesday? If so, would you be interested in helping during my tutorial (mainly helping the participants with installation and the exercises). Your help would be much appreciated.

Preliminary materials for the tutorial are  on Github [2]. I have only 1.5h so I won't cover all lessons. My plan is to do lessons 1-5 (broadcasting, shape and strides, dtypes and structured array, intro to xarray) . Your feedback is most welcome!

Thanks in advance!

Cheers,

Bartosz

[1] https://www.euroscipy.org/2016/schedule/sessions/5/
[2] https://github.com/btel/2016-erlangen-euroscipy-advanced-numpy


From alan.isaac at gmail.com  Tue Aug 16 08:52:07 2016
From: alan.isaac at gmail.com (Alan Isaac)
Date: Tue, 16 Aug 2016 08:52:07 -0400
Subject: [Numpy-discussion] weighted random choice in Python
In-Reply-To: <CABL7CQhsP9OjoXP0AMso7ZQCjWovmCiumhsV_Xam792nT4Z94Q@mail.gmail.com>
References: <1bf63320-0881-8097-a9a8-46c6a6ba6b7e@gmail.com>
 <CABL7CQhsP9OjoXP0AMso7ZQCjWovmCiumhsV_Xam792nT4Z94Q@mail.gmail.com>
Message-ID: <157814d7-6825-da88-ebc5-3116d229a203@gmail.com>

On 8/16/2016 4:06 AM, Ralf Gommers wrote:
> The whole enhancement request doesn't look very interesting imho.


Because the functionality is already in NumPy,
or because it is easily user-written?

Alan



From matthew.gidden at gmail.com  Tue Aug 16 09:36:01 2016
From: matthew.gidden at gmail.com (Matthew Gidden)
Date: Tue, 16 Aug 2016 15:36:01 +0200
Subject: [Numpy-discussion] Euroscipy
In-Reply-To: <57b306f53aebe_44ce14f7e54c9@Pct-EqAlain-Z30.notmuch>
References: <1470158553.5372.2.camel@sipsolutions.net>
 <CABL7CQjTgo0yDvasZ5vP6jgs0Sa0-S-SY7-Gr=YwU_Mj44wPKg@mail.gmail.com>
 <CAHRdPFzEEHFZWUhevSdtJV2QwSX6S4KvEJSJQ0NG9rtda+KiyA@mail.gmail.com>
 <57b306f53aebe_44ce14f7e54c9@Pct-EqAlain-Z30.notmuch>
Message-ID: <CAHRdPFx23qnyABkQ=3UKGdO+BsMroKGnnT4HkqDH2wdiimhFqA@mail.gmail.com>

Hi Bartosz,

I'll happily help. I normally teach Software Carpentry lessons [1], so have
experience in the workshop environment. I'll take a look over what you
prepared.

Cheers,
Matt

[1] http://software-carpentry.org/lessons/

On Tue, Aug 16, 2016 at 2:28 PM, <mail at telenczuk.pl> wrote:

> Hi all,
>
> I will present advanced NumPy tutorial at EuroScipy next week [1]. Is
> anyone going to attend the tutorials on Tuesday? If so, would you be
> interested in helping during my tutorial (mainly helping the participants
> with installation and the exercises). Your help would be much appreciated.
>
> Preliminary materials for the tutorial are  on Github [2]. I have only
> 1.5h so I won't cover all lessons. My plan is to do lessons 1-5
> (broadcasting, shape and strides, dtypes and structured array, intro to
> xarray) . Your feedback is most welcome!
>
> Thanks in advance!
>
> Cheers,
>
> Bartosz
>
> [1] https://www.euroscipy.org/2016/schedule/sessions/5/
> [2] https://github.com/btel/2016-erlangen-euroscipy-advanced-numpy
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160816/cefe852d/attachment.html>

From stuarteberg at gmail.com  Fri Aug 19 11:29:29 2016
From: stuarteberg at gmail.com (Stuart Berg)
Date: Fri, 19 Aug 2016 11:29:29 -0400
Subject: [Numpy-discussion] How to trigger warnings for integer division in
	python 2
Message-ID: <CALi1z7P+qtDUv4FqpnCUPY78S0h3UXuD0HnEA8BwCLyZ8pnb9Q@mail.gmail.com>

Hi,

To help people migrate their code bases from Python 2 to Python 3, the
python interpreter has a handy option '-3' that issues warnings at
runtime.  One of the warnings is for integer division:

$ echo "print 3/2" > /tmp/foo.py
$ python -3 /tmp/foo.py
/tmp/foo.py:1: DeprecationWarning: classic int division
  print 3/2
1

But no warnings are shown for division of numpy arrays, e.g. for a
statement like this:
print np.array([3]) / np.array([2])

I see that np.seterr can be used to issue certain types of division
warnings, but not this one.  Is there a way to activate integer division
warnings?  It would really help me migrate my application to Python 3.

Thanks,
Stuart
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160819/724ddd46/attachment.html>

From sebastian at sipsolutions.net  Fri Aug 19 13:03:15 2016
From: sebastian at sipsolutions.net (Sebastian Berg)
Date: Fri, 19 Aug 2016 19:03:15 +0200
Subject: [Numpy-discussion] How to trigger warnings for integer division
 in python 2
In-Reply-To: <CALi1z7P+qtDUv4FqpnCUPY78S0h3UXuD0HnEA8BwCLyZ8pnb9Q@mail.gmail.com>
References: <CALi1z7P+qtDUv4FqpnCUPY78S0h3UXuD0HnEA8BwCLyZ8pnb9Q@mail.gmail.com>
Message-ID: <1471626195.9335.5.camel@sipsolutions.net>

On Fr, 2016-08-19 at 11:29 -0400, Stuart Berg wrote:
> Hi,
> 
> To help people migrate their code bases from Python 2 to Python 3,
> the python interpreter has a handy option '-3' that issues warnings
> at runtime.? One of the warnings is for integer division:
> 
> $ echo "print 3/2" > /tmp/foo.py
> $ python -3 /tmp/foo.py
> /tmp/foo.py:1: DeprecationWarning: classic int division
> ? print 3/2
> 1
> 
> But no warnings are shown for division of numpy arrays, e.g. for a
> statement like this:
> print np.array([3]) / np.array([2])
> 
> I see that np.seterr can be used to issue certain types of division
> warnings, but not this one.? Is there a way to activate integer
> division warnings?? It would really help me migrate my application to
> Python 3.
> 

I don't think numpy implements any py3kwarnings. It seems that it would
be possible though. On newer versions we got more strict about using
floats instead of ints, so some of these places might follow up with a
warning quickly.
I guess the question is whether we should aim to add at least some of these warnings and someone is willing to put in the effort (I suppose it is likely only a few places). I am not sure how easy they are on the C side, but probably not difficult at all.

- Sebastian




> Thanks,
> Stuart
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160819/5133f526/attachment.sig>

From stuarteberg at gmail.com  Fri Aug 19 13:33:01 2016
From: stuarteberg at gmail.com (Stuart Berg)
Date: Fri, 19 Aug 2016 13:33:01 -0400
Subject: [Numpy-discussion] How to trigger warnings for integer division
 in python 2
In-Reply-To: <1471626195.9335.5.camel@sipsolutions.net>
References: <CALi1z7P+qtDUv4FqpnCUPY78S0h3UXuD0HnEA8BwCLyZ8pnb9Q@mail.gmail.com>
 <1471626195.9335.5.camel@sipsolutions.net>
Message-ID: <CALi1z7N-6KuTeSyEC6GwDOz6+kJiAb5SUy8bYq1SMN5kEbRZFQ@mail.gmail.com>

> I guess the question is whether we should aim to add at least some of
> these warnings


As you can probably guess, I think such a feature would be very useful :-)
When my team started planning our migration to Python 3, we quickly
realized that the integer division issue was our biggest risk.  It has the
potential for subtle bugs to remain hidden in our code base long after the
migration, whereas most of the other Python 3 changes will be easier to
identify.


> and someone is willing to put in the effort (I suppose it is likely only a
> few places).
>
I am not sure how easy they are on the C side, but probably not difficult
> at all.


OK, good to hear that it shouldn't be difficult.  I've opened the following
issue on the github repo, to continue this discussion and hopefully attract
a volunteer. :-)

https://github.com/numpy/numpy/issues/7949

Best regards,
Stuart
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160819/616f772d/attachment.html>

From alan.isaac at gmail.com  Sat Aug 20 16:16:14 2016
From: alan.isaac at gmail.com (Alan Isaac)
Date: Sat, 20 Aug 2016 16:16:14 -0400
Subject: [Numpy-discussion] coordinate bounds
Message-ID: <e58ac269-98c8-271b-22bf-1c73962ef935@gmail.com>

Is there a numpy equivalent to Mma's CoordinateBounds command?
http://reference.wolfram.com/language/ref/CoordinateBounds.html

Thanks,
Alan Isaac


From robert.kern at gmail.com  Sat Aug 20 17:14:38 2016
From: robert.kern at gmail.com (Robert Kern)
Date: Sat, 20 Aug 2016 22:14:38 +0100
Subject: [Numpy-discussion] coordinate bounds
In-Reply-To: <e58ac269-98c8-271b-22bf-1c73962ef935@gmail.com>
References: <e58ac269-98c8-271b-22bf-1c73962ef935@gmail.com>
Message-ID: <CAF6FJiv+cRHM8H5QDhixiD56pHOgWRQTERLAmZmTwS1_zoOUzg@mail.gmail.com>

On Sat, Aug 20, 2016 at 9:16 PM, Alan Isaac <alan.isaac at gmail.com> wrote:
>
> Is there a numpy equivalent to Mma's CoordinateBounds command?
> http://reference.wolfram.com/language/ref/CoordinateBounds.html

The first signature can be computed like so:

  np.transpose([coords.min(axis=0), coords.max(axis=0)])

--
Robert Kern
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160820/fd68f515/attachment.html>

From dipugee at gmail.com  Sun Aug 21 13:46:23 2016
From: dipugee at gmail.com (=?utf-8?Q?Dipankar_=E2=80=9CDipu=E2=80=9D_Ganguly?=)
Date: Sun, 21 Aug 2016 10:46:23 -0700
Subject: [Numpy-discussion] NumPy-Discussion Digest, Vol 119, Issue 11
In-Reply-To: <mailman.5.1471780801.18876.numpy-discussion@scipy.org>
References: <mailman.5.1471780801.18876.numpy-discussion@scipy.org>
Message-ID: <E7D4D6DC-1EEC-4367-BE34-CE4C0078FD25@gmail.com>

Is there a way to use Wolframs? Mathematica 11 within IPython on Jupyter running on Anaconda?s Navigator on a Mac with OS 10.11.6? Failing that, what Python package would give me those capabilities?

Thanks.

Dipu


Dipankar Ganguly
Consultant: Strategy/Technology/Commercialization
Bothell, WA
Cell: 408-203-8814
email: dipugee at gmail.com
http://www.linkedin.com/in/dipugee 









> On Aug 21, 2016, at 5:00 AM, numpy-discussion-request at scipy.org wrote:
> 
> Send NumPy-Discussion mailing list submissions to
> 	numpy-discussion at scipy.org
> 
> To subscribe or unsubscribe via the World Wide Web, visit
> 	https://mail.scipy.org/mailman/listinfo/numpy-discussion
> or, via email, send a message with subject or body 'help' to
> 	numpy-discussion-request at scipy.org
> 
> You can reach the person managing the list at
> 	numpy-discussion-owner at scipy.org
> 
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of NumPy-Discussion digest..."
> 
> 
> Today's Topics:
> 
>   1. coordinate bounds (Alan Isaac)
>   2. Re: coordinate bounds (Robert Kern)
> 
> 
> ----------------------------------------------------------------------
> 
> Message: 1
> Date: Sat, 20 Aug 2016 16:16:14 -0400
> From: Alan Isaac <alan.isaac at gmail.com>
> To: Discussion of Numerical Python <numpy-discussion at scipy.org>
> Subject: [Numpy-discussion] coordinate bounds
> Message-ID: <e58ac269-98c8-271b-22bf-1c73962ef935 at gmail.com>
> Content-Type: text/plain; charset=utf-8; format=flowed
> 
> Is there a numpy equivalent to Mma's CoordinateBounds command?
> http://reference.wolfram.com/language/ref/CoordinateBounds.html
> 
> Thanks,
> Alan Isaac
> 
> 
> ------------------------------
> 
> Message: 2
> Date: Sat, 20 Aug 2016 22:14:38 +0100
> From: Robert Kern <robert.kern at gmail.com>
> To: Discussion of Numerical Python <numpy-discussion at scipy.org>
> Subject: Re: [Numpy-discussion] coordinate bounds
> Message-ID:
> 	<CAF6FJiv+cRHM8H5QDhixiD56pHOgWRQTERLAmZmTwS1_zoOUzg at mail.gmail.com>
> Content-Type: text/plain; charset="utf-8"
> 
> On Sat, Aug 20, 2016 at 9:16 PM, Alan Isaac <alan.isaac at gmail.com> wrote:
>> 
>> Is there a numpy equivalent to Mma's CoordinateBounds command?
>> http://reference.wolfram.com/language/ref/CoordinateBounds.html
> 
> The first signature can be computed like so:
> 
>  np.transpose([coords.min(axis=0), coords.max(axis=0)])
> 
> --
> Robert Kern
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: <https://mail.scipy.org/pipermail/numpy-discussion/attachments/20160820/fd68f515/attachment-0001.html>
> 
> ------------------------------
> 
> Subject: Digest Footer
> 
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
> 
> 
> ------------------------------
> 
> End of NumPy-Discussion Digest, Vol 119, Issue 11
> *************************************************

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160821/261955b4/attachment.html>

From njs at pobox.com  Sun Aug 21 18:53:43 2016
From: njs at pobox.com (Nathaniel Smith)
Date: Sun, 21 Aug 2016 15:53:43 -0700
Subject: [Numpy-discussion] NumPy-Discussion Digest, Vol 119, Issue 11
In-Reply-To: <E7D4D6DC-1EEC-4367-BE34-CE4C0078FD25@gmail.com>
References: <mailman.5.1471780801.18876.numpy-discussion@scipy.org>
 <E7D4D6DC-1EEC-4367-BE34-CE4C0078FD25@gmail.com>
Message-ID: <CAPJVwBksKF5nEhFVZM9vfAizprJppjT7GEgnYZK2=byMf3bHDg@mail.gmail.com>

On Aug 21, 2016 10:46 AM, "Dipankar ?Dipu? Ganguly" <dipugee at gmail.com>
wrote:
>
> Is there a way to use Wolframs? Mathematica 11 within IPython on Jupyter
running on Anaconda?s Navigator on a Mac with OS 10.11.6? Failing that,
what Python package would give me those capabilities?

This has nothing to do with numpy, so I'm afraid you're in the wrong place.
You might try asking the Jupyter developers, or looking into Sage.

-n
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160821/22bb2b00/attachment.html>

From oleksandr.pavlyk at intel.com  Wed Aug 24 15:41:06 2016
From: oleksandr.pavlyk at intel.com (Pavlyk, Oleksandr)
Date: Wed, 24 Aug 2016 19:41:06 +0000
Subject: [Numpy-discussion] numpy.distutils issue
Message-ID: <4C9EDA7282E297428F3986994EB0FBD3856C71@ORSMSX110.amr.corp.intel.com>

Hi All,

According to the documentation page:

   http://docs.scipy.org/doc/numpy/reference/distutils.html

Function add_library allows the following keywords:
      extra_f77_compiler_args
      extra_f90_compiler_args

however setting them seem to have no effect for my extension. Digging deeper, I discovered,
the documentation is inconsistent with the implementation, as per

https://github.com/numpy/numpy/blob/v1.11.0/numpy/distutils/fcompiler/__init__.py#L569

https://github.com/numpy/numpy/blob/v1.11.0/numpy/distutils/fcompiler/__init__.py#L583

And indeed, setting extra_f77_compile_arg has the effect I was looking for.

Fixing it is easy, but I am less certain whether we should fix the docs, or the code.

Given that add_extension lists extra_compile_args, extra_f77_compile_args, etc, I would think it
Is the documentation that need to change.

Please confirm, and I will open up a pull request for this.

Thank you,
Oleksandr
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160824/d0bb81f0/attachment.html>

From charlesr.harris at gmail.com  Wed Aug 24 19:05:48 2016
From: charlesr.harris at gmail.com (Charles R Harris)
Date: Wed, 24 Aug 2016 17:05:48 -0600
Subject: [Numpy-discussion] numpy.distutils issue
In-Reply-To: <4C9EDA7282E297428F3986994EB0FBD3856C71@ORSMSX110.amr.corp.intel.com>
References: <4C9EDA7282E297428F3986994EB0FBD3856C71@ORSMSX110.amr.corp.intel.com>
Message-ID: <CAB6mnxKgmOFT8TKCRZbEf2tK=-0AWpmC3cYFO2o5PvxXydUPUw@mail.gmail.com>

On Wed, Aug 24, 2016 at 1:41 PM, Pavlyk, Oleksandr <
oleksandr.pavlyk at intel.com> wrote:

> Hi All,
>
>
>
> According to the documentation page:
>
>
>
>    http://docs.scipy.org/doc/numpy/reference/distutils.html
>
>
>
> Function add_library allows the following keywords:
>
>       extra_f77_compiler_args
>
>       extra_f90_compiler_args
>
>
>
> however setting them seem to have no effect for my extension. Digging
> deeper, I discovered,
>
> the documentation is inconsistent with the implementation, as per
>
>
>
> https://github.com/numpy/numpy/blob/v1.11.0/numpy/
> distutils/fcompiler/__init__.py#L569
>
>
>
> https://github.com/numpy/numpy/blob/v1.11.0/numpy/
> distutils/fcompiler/__init__.py#L583
>
>
>
> And indeed, setting extra_f77_compile_arg has the effect I was looking
> for.
>
> Fixing it is easy, but I am less certain whether we should fix the docs,
> or the code.
>
>
>
> Given that add_extension lists extra_compile_args, extra_f77_compile_args,
> etc, I would think it
>
> Is the documentation that need to change.
>
>
>
> Please confirm, and I will open up a pull request for this.
>

That's rather unfortunate, "compiler" would be better than "compile", but
it is best to document the actual behavior. If we later settle on changing
the argument we can do that, but it is a more involved process.

Chuck

>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160824/dc71bf0d/attachment.html>

From charlesr.harris at gmail.com  Wed Aug 24 19:18:35 2016
From: charlesr.harris at gmail.com (Charles R Harris)
Date: Wed, 24 Aug 2016 17:18:35 -0600
Subject: [Numpy-discussion] numpy.distutils issue
In-Reply-To: <CAB6mnxKgmOFT8TKCRZbEf2tK=-0AWpmC3cYFO2o5PvxXydUPUw@mail.gmail.com>
References: <4C9EDA7282E297428F3986994EB0FBD3856C71@ORSMSX110.amr.corp.intel.com>
 <CAB6mnxKgmOFT8TKCRZbEf2tK=-0AWpmC3cYFO2o5PvxXydUPUw@mail.gmail.com>
Message-ID: <CAB6mnxLVO0QZOS9stQVtJTd5fNo9b4U-0RNUz7mtFeV9E8rV=g@mail.gmail.com>

On Wed, Aug 24, 2016 at 5:05 PM, Charles R Harris <charlesr.harris at gmail.com
> wrote:

>
>
> On Wed, Aug 24, 2016 at 1:41 PM, Pavlyk, Oleksandr <
> oleksandr.pavlyk at intel.com> wrote:
>
>> Hi All,
>>
>>
>>
>> According to the documentation page:
>>
>>
>>
>>    http://docs.scipy.org/doc/numpy/reference/distutils.html
>>
>>
>>
>> Function add_library allows the following keywords:
>>
>>       extra_f77_compiler_args
>>
>>       extra_f90_compiler_args
>>
>>
>>
>> however setting them seem to have no effect for my extension. Digging
>> deeper, I discovered,
>>
>> the documentation is inconsistent with the implementation, as per
>>
>>
>>
>> https://github.com/numpy/numpy/blob/v1.11.0/numpy/distutils/
>> fcompiler/__init__.py#L569
>>
>>
>>
>> https://github.com/numpy/numpy/blob/v1.11.0/numpy/distutils/
>> fcompiler/__init__.py#L583
>>
>>
>>
>> And indeed, setting extra_f77_compile_arg has the effect I was looking
>> for.
>>
>> Fixing it is easy, but I am less certain whether we should fix the docs,
>> or the code.
>>
>>
>>
>> Given that add_extension lists extra_compile_args,
>> extra_f77_compile_args, etc, I would think it
>>
>> Is the documentation that need to change.
>>
>>
>>
>> Please confirm, and I will open up a pull request for this.
>>
>
> That's rather unfortunate, "compiler" would be better than "compile", but
> it is best to document the actual behavior. If we later settle on changing
> the argument we can do that, but it is a more involved process.
>
>
Although I suppose we could allow either in the future.

Chuck

>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160824/c6197761/attachment.html>

From jfoxrabinovitz at gmail.com  Thu Aug 25 10:36:50 2016
From: jfoxrabinovitz at gmail.com (Joseph Fox-Rabinovitz)
Date: Thu, 25 Aug 2016 10:36:50 -0400
Subject: [Numpy-discussion] Indexing issue with ndarrays
Message-ID: <CAAa1KPY2uqrbgUrL9Mrn8E62FwhaJyxyu2Zo3+gPCh34m6519w@mail.gmail.com>

This issue recently came up on Stack Overflow:
http://stackoverflow.com/questions/39145795/masking-a-series-with-a-boolean-array.
The poster attempted to index an ndarray with a pandas boolean Series
object (all False), but the result was as if he had indexed with an array
of integer zeros.

Can someone explain this behavior? I can see two obvious possibilities:

   1. ndarray checks if the input to __getitem__ is of exactly the right
   type, not using instanceof.
   2. pandas actually uses a wider datatype than boolean internally, so
   indexing with the series is in fact indexing with an integer array.

In my attempt to reproduce the poster's results, I got the following
warning:

FutureWarning: in the future, boolean array-likes will be handled as a
boolean array index

This indicates that the issue is probably #1 and that a fix is already on
the way. Please correct me if I am wrong. Also, where does the code for
ndarray.__getitem__ live?

Thanks,

    -Joe
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160825/6477c42d/attachment.html>

From sebastian at sipsolutions.net  Thu Aug 25 16:37:45 2016
From: sebastian at sipsolutions.net (Sebastian Berg)
Date: Thu, 25 Aug 2016 22:37:45 +0200
Subject: [Numpy-discussion] Indexing issue with ndarrays
In-Reply-To: <CAAa1KPY2uqrbgUrL9Mrn8E62FwhaJyxyu2Zo3+gPCh34m6519w@mail.gmail.com>
References: <CAAa1KPY2uqrbgUrL9Mrn8E62FwhaJyxyu2Zo3+gPCh34m6519w@mail.gmail.com>
Message-ID: <1472157465.1331.1.camel@sipsolutions.net>

On Do, 2016-08-25 at 10:36 -0400, Joseph Fox-Rabinovitz wrote:
> This issue recently came up on Stack Overflow: http://stackoverflow.c
> om/questions/39145795/masking-a-series-with-a-boolean-array. The
> poster attempted to index an ndarray with a pandas boolean Series
> object (all False), but the result was as if he had indexed with an
> array of integer zeros.
> 
> Can someone explain this behavior? I can see two obvious
> possibilities:
> ndarray checks if the input to __getitem__ is of exactly the right
> type, not using instanceof.
> pandas actually uses a wider datatype than boolean internally, so
> indexing with the series is in fact indexing with an integer array.

You are overthinking it ;). The reason is quite simply that the logic
used to be:

?* Boolean array? -> think about boolean indexing.
?* Everything "array-like" (not caught earlier) -> convert to `intp`
array and do integer indexing.

Now you might wonder why, but probably it is quite simply because
boolean indexing was tagged on later.

- Sebastian


> In my attempt to reproduce the poster's results, I got the following
> warning:
> FutureWarning: in the future, boolean array-likes will be handled as
> a boolean array index
> This indicates that the issue is probably #1 and that a fix is
> already on the way. Please correct me if I am wrong. Also, where does
> the code for ndarray.__getitem__ live?
> Thanks,
> ??? -Joe
> 
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160825/f6296be0/attachment.sig>

From jfoxrabinovitz at gmail.com  Fri Aug 26 09:57:22 2016
From: jfoxrabinovitz at gmail.com (Joseph Fox-Rabinovitz)
Date: Fri, 26 Aug 2016 09:57:22 -0400
Subject: [Numpy-discussion] Indexing issue with ndarrays
In-Reply-To: <1472157465.1331.1.camel@sipsolutions.net>
References: <CAAa1KPY2uqrbgUrL9Mrn8E62FwhaJyxyu2Zo3+gPCh34m6519w@mail.gmail.com>
 <1472157465.1331.1.camel@sipsolutions.net>
Message-ID: <CAAa1KPaPLpEpWdNuCdRAiojmThd+VGjXnn2QLrda-=kpoamCOg@mail.gmail.com>

On Thu, Aug 25, 2016 at 4:37 PM, Sebastian Berg <sebastian at sipsolutions.net>
wrote:

> On Do, 2016-08-25 at 10:36 -0400, Joseph Fox-Rabinovitz wrote:
> > This issue recently came up on Stack Overflow: http://stackoverflow.c
> > om/questions/39145795/masking-a-series-with-a-boolean-array. The
> > poster attempted to index an ndarray with a pandas boolean Series
> > object (all False), but the result was as if he had indexed with an
> > array of integer zeros.
> >
> > Can someone explain this behavior? I can see two obvious
> > possibilities:
> > ndarray checks if the input to __getitem__ is of exactly the right
> > type, not using instanceof.
> > pandas actually uses a wider datatype than boolean internally, so
> > indexing with the series is in fact indexing with an integer array.
>
> You are overthinking it ;). The reason is quite simply that the logic
> used to be:
>
>  * Boolean array? -> think about boolean indexing.
>  * Everything "array-like" (not caught earlier) -> convert to `intp`
> array and do integer indexing.
>
> Now you might wonder why, but probably it is quite simply because
> boolean indexing was tagged on later.
>
> - Sebastian
>
>
> > In my attempt to reproduce the poster's results, I got the following
> > warning:
> > FutureWarning: in the future, boolean array-likes will be handled as
> > a boolean array index
> > This indicates that the issue is probably #1 and that a fix is
> > already on the way. Please correct me if I am wrong. Also, where does
> > the code for ndarray.__getitem__ live?
> > Thanks,
> >     -Joe
> >
> > _______________________________________________
> > NumPy-Discussion mailing list
> > NumPy-Discussion at scipy.org
> > https://mail.scipy.org/mailman/listinfo/numpy-discussion
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>
>
This makes perfect sense. I would like to help fix it if a fix is desired
and has not been done already. Could you point me to where the "Boolean
array?, etc." decision happens? I have had trouble navigating to
`__getitem__` (which I assume is somewhere in np.core.multiarray C code.

    -Joe
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160826/e3ab1d15/attachment.html>

From sebastian at sipsolutions.net  Fri Aug 26 10:08:14 2016
From: sebastian at sipsolutions.net (Sebastian Berg)
Date: Fri, 26 Aug 2016 16:08:14 +0200
Subject: [Numpy-discussion] Indexing issue with ndarrays
In-Reply-To: <CAAa1KPaPLpEpWdNuCdRAiojmThd+VGjXnn2QLrda-=kpoamCOg@mail.gmail.com>
References: <CAAa1KPY2uqrbgUrL9Mrn8E62FwhaJyxyu2Zo3+gPCh34m6519w@mail.gmail.com>
 <1472157465.1331.1.camel@sipsolutions.net>
 <CAAa1KPaPLpEpWdNuCdRAiojmThd+VGjXnn2QLrda-=kpoamCOg@mail.gmail.com>
Message-ID: <1472220494.14163.9.camel@sipsolutions.net>

On Fr, 2016-08-26 at 09:57 -0400, Joseph Fox-Rabinovitz wrote:
> 
> 
> On Thu, Aug 25, 2016 at 4:37 PM, Sebastian Berg <sebastian at sipsolutio
> ns.net> wrote:
> > On Do, 2016-08-25 at 10:36 -0400, Joseph Fox-Rabinovitz wrote:
> > > This issue recently came up on Stack Overflow: http://stackoverfl
> > ow.c
> > > om/questions/39145795/masking-a-series-with-a-boolean-array. The
> > > poster attempted to index an ndarray with a pandas boolean Series
> > > object (all False), but the result was as if he had indexed with
> > an
> > > array of integer zeros.
> > >
> > > Can someone explain this behavior? I can see two obvious
> > > possibilities:
> > > ndarray checks if the input to __getitem__ is of exactly the
> > right
> > > type, not using instanceof.
> > > pandas actually uses a wider datatype than boolean internally, so
> > > indexing with the series is in fact indexing with an integer
> > array.
> > 
> > You are overthinking it ;). The reason is quite simply that the
> > logic
> > used to be:
> > 
> > ?* Boolean array? -> think about boolean indexing.
> > ?* Everything "array-like" (not caught earlier) -> convert to
> > `intp`
> > array and do integer indexing.
> > 
> > Now you might wonder why, but probably it is quite simply because
> > boolean indexing was tagged on later.
> > 
> > - Sebastian
> > 
> > 
> > > In my attempt to reproduce the poster's results, I got the
> > following
> > > warning:
> > > FutureWarning: in the future, boolean array-likes will be handled
> > as
> > > a boolean array index
> > > This indicates that the issue is probably #1 and that a fix is
> > > already on the way. Please correct me if I am wrong. Also, where
> > does
> > > the code for ndarray.__getitem__ live?
> > > Thanks,
> > > ??? -Joe
> > >
> > > _______________________________________________
> > > NumPy-Discussion mailing list
> > > NumPy-Discussion at scipy.org
> > > https://mail.scipy.org/mailman/listinfo/numpy-discussion
> > _______________________________________________
> > NumPy-Discussion mailing list
> > NumPy-Discussion at scipy.org
> > https://mail.scipy.org/mailman/listinfo/numpy-discussion
> > 
> This makes perfect sense. I would like to help fix it if a fix is
> desired and has not been done already. Could you point me to where
> the "Boolean array?, etc." decision happens? I have had trouble
> navigating to `__getitem__` (which I assume is somewhere in
> np.core.multiarray C code.
> 

As the warning says, it already is fixed in a sense (we just have to
move forward with the deprecation, which you can maybe actually do at
this time). This is all in the mapping.c stuff, without checking, there
is a function called something like "prepare index" which goes through
all the different types of indexing objects. It should be pretty
straight forward to find the warning.

The actual old behaviour where this behaviour originated in was a
completely different code base though (you would have to check out some
pre 1.9 version of numpy if you are interested in archeology.

- Sebastian




> ??? -Joe
> 
> 
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160826/af9ca201/attachment.sig>

From mail at telenczuk.pl  Mon Aug 29 07:43:03 2016
From: mail at telenczuk.pl (mail at telenczuk.pl)
Date: Mon, 29 Aug 2016 13:43:03 +0200
Subject: [Numpy-discussion] Which NumPy/Numpy/numpy spelling?
Message-ID: <57c41fc7bf6ce_4222a36c5419@Pct-EqAlain-Z30.notmuch>

Hi all,

What is the official spelling of NumPy/Numpy/numpy? 

The documentation is not consistent and it mixes both NumPy and Numpy. For example, the reference manual uses both spellings in the introduction paragraph (http://docs.scipy.org/doc/numpy/reference/):

    "This reference manual details functions, modules, and objects included in Numpy, describing what they are and what they do. For learning how to use NumPy, see also NumPy User Guide."

However, in all docs taken together "NumPy" is most frequently used (74%):

% find . -name "*.rst" -exec grep Numpy -ow {} \; | wc -l
161
% find . -name "*.rst" -exec grep NumPy -ow {} \; | wc -l
471

I also reported it as an issue: https://github.com/numpy/numpy/issues/7986

Yours,

Bartosz


From hodge at stsci.edu  Mon Aug 29 08:39:59 2016
From: hodge at stsci.edu (Phil Hodge)
Date: Mon, 29 Aug 2016 08:39:59 -0400
Subject: [Numpy-discussion] Which NumPy/Numpy/numpy spelling?
In-Reply-To: <57c41fc7bf6ce_4222a36c5419@Pct-EqAlain-Z30.notmuch>
References: <57c41fc7bf6ce_4222a36c5419@Pct-EqAlain-Z30.notmuch>
Message-ID: <0fbc81f4-c384-1e0d-6876-09a665caf628@stsci.edu>

On 08/29/2016 07:43 AM, mail at telenczuk.pl wrote:
> What is the official spelling of NumPy/Numpy/numpy?

IMHO it should be written numpy, because ...

 >>> import NumPy
Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
ImportError: No module named NumPy
 >>> import Numpy
Traceback (most recent call last):
   File "<stdin>", line 1, in <module>
ImportError: No module named Numpy
 >>> import numpy
 >>>

Phil


From mail at telenczuk.pl  Mon Aug 29 09:22:13 2016
From: mail at telenczuk.pl (Bartosz Telenczuk)
Date: Mon, 29 Aug 2016 15:22:13 +0200
Subject: [Numpy-discussion] Which NumPy/Numpy/numpy spelling?
In-Reply-To: <0fbc81f4-c384-1e0d-6876-09a665caf628@stsci.edu>
References: <57c41fc7bf6ce_4222a36c5419@Pct-EqAlain-Z30.notmuch>
 <0fbc81f4-c384-1e0d-6876-09a665caf628@stsci.edu>
Message-ID: <57c43705eee41_4222a36c5420@Pct-EqAlain-Z30.notmuch>

Hi,

I would not mind any choice as long as it's consistent.

I agree that using all-lowercase spelling may avoid some common errors. However,
PEP8 requires all module/package names to be lower case [1].  If we force the
name of the library and the corresponding package to be the same, all Python
libraries would be named in lowercase. This would not be the best choice for
libraries, which have multi-component names (like NumPy = Numerical Python).

Note also that both the Wikipedia page [2] and the official NumPy logo [3] use
"NumPy" spelling.

Some other popular [4] libraries use similar dichotomies:

- Django - import django
- Cython - import cython
- PyYAML - import yaml
- scikit-learn - import sklearn

On the other hand all standard Python libraries are lower-case named.

Cheers,

Bartosz

[1] https://www.python.org/dev/peps/pep-0008/#package-and-module-names
[2] https://en.wikipedia.org/wiki/NumPy
[3] http://www.numpy.org/_static/numpy_logo.png
[4] http://pypi-ranking.info/alltime

> On 08/29/2016 07:43 AM, mail at telenczuk.pl wrote:
> > What is the official spelling of NumPy/Numpy/numpy?
> 
> IMHO it should be written numpy, because ...
> 
>  >>> import NumPy
> Traceback (most recent call last):
>    File "<stdin>", line 1, in <module>
> ImportError: No module named NumPy
>  >>> import Numpy
> Traceback (most recent call last):
>    File "<stdin>", line 1, in <module>
> ImportError: No module named Numpy
>  >>> import numpy
>  >>>
> 
> Phil
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion




From josef.pktd at gmail.com  Mon Aug 29 12:56:50 2016
From: josef.pktd at gmail.com (josef.pktd at gmail.com)
Date: Mon, 29 Aug 2016 12:56:50 -0400
Subject: [Numpy-discussion] Which NumPy/Numpy/numpy spelling?
In-Reply-To: <57c43705eee41_4222a36c5420@Pct-EqAlain-Z30.notmuch>
References: <57c41fc7bf6ce_4222a36c5419@Pct-EqAlain-Z30.notmuch>
 <0fbc81f4-c384-1e0d-6876-09a665caf628@stsci.edu>
 <57c43705eee41_4222a36c5420@Pct-EqAlain-Z30.notmuch>
Message-ID: <CAMMTP+Cm3zyL8hqbpzFmSLTZWnsWpQam69oOzAjXou=t-SEw5Q@mail.gmail.com>

https://mail.scipy.org/pipermail/scipy-user/2010-June/025756.html

"NumPy and SciPy to refer to the projects. numpy and scipy to refer to
the packages, specifically. When in doubt, use the former."

I thought there was also another discussion about capital letters, but
I don't find it.

Josef


On Mon, Aug 29, 2016 at 9:22 AM, Bartosz Telenczuk <mail at telenczuk.pl> wrote:
> Hi,
>
> I would not mind any choice as long as it's consistent.
>
> I agree that using all-lowercase spelling may avoid some common errors. However,
> PEP8 requires all module/package names to be lower case [1].  If we force the
> name of the library and the corresponding package to be the same, all Python
> libraries would be named in lowercase. This would not be the best choice for
> libraries, which have multi-component names (like NumPy = Numerical Python).
>
> Note also that both the Wikipedia page [2] and the official NumPy logo [3] use
> "NumPy" spelling.
>
> Some other popular [4] libraries use similar dichotomies:
>
> - Django - import django
> - Cython - import cython
> - PyYAML - import yaml
> - scikit-learn - import sklearn
>
> On the other hand all standard Python libraries are lower-case named.
>
> Cheers,
>
> Bartosz
>
> [1] https://www.python.org/dev/peps/pep-0008/#package-and-module-names
> [2] https://en.wikipedia.org/wiki/NumPy
> [3] http://www.numpy.org/_static/numpy_logo.png
> [4] http://pypi-ranking.info/alltime
>
>> On 08/29/2016 07:43 AM, mail at telenczuk.pl wrote:
>> > What is the official spelling of NumPy/Numpy/numpy?
>>
>> IMHO it should be written numpy, because ...
>>
>>  >>> import NumPy
>> Traceback (most recent call last):
>>    File "<stdin>", line 1, in <module>
>> ImportError: No module named NumPy
>>  >>> import Numpy
>> Traceback (most recent call last):
>>    File "<stdin>", line 1, in <module>
>> ImportError: No module named Numpy
>>  >>> import numpy
>>  >>>
>>
>> Phil
>> _______________________________________________
>> NumPy-Discussion mailing list
>> NumPy-Discussion at scipy.org
>> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>
>
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion


From ben.v.root at gmail.com  Mon Aug 29 13:47:46 2016
From: ben.v.root at gmail.com (Benjamin Root)
Date: Mon, 29 Aug 2016 13:47:46 -0400
Subject: [Numpy-discussion] Which NumPy/Numpy/numpy spelling?
In-Reply-To: <CAMMTP+Cm3zyL8hqbpzFmSLTZWnsWpQam69oOzAjXou=t-SEw5Q@mail.gmail.com>
References: <57c41fc7bf6ce_4222a36c5419@Pct-EqAlain-Z30.notmuch>
 <0fbc81f4-c384-1e0d-6876-09a665caf628@stsci.edu>
 <57c43705eee41_4222a36c5420@Pct-EqAlain-Z30.notmuch>
 <CAMMTP+Cm3zyL8hqbpzFmSLTZWnsWpQam69oOzAjXou=t-SEw5Q@mail.gmail.com>
Message-ID: <CANNq6FkMtATdcyS3VZcmmT1YBFJ9-AJgvY8bR=tGqmhV8BW=5Q@mail.gmail.com>

There was similar discussion almost two years ago with respect to
capitalization of matplotlib in prose. Most of the time, it was lower-case
in our documentation, but then the question was if it should be upper-case
at the beginning of the sentence... or should it always be upper-cased like
a proper noun. I don't think a clear consensus was reached, but I ended up
treating it as a proper noun in my book.

I am also pretty sure I used "NumPy" in most places, even for "NumPy
arrays", which still looks weird to me when I go back over my book.

Ben Root


On Mon, Aug 29, 2016 at 12:56 PM, <josef.pktd at gmail.com> wrote:

> https://mail.scipy.org/pipermail/scipy-user/2010-June/025756.html
>
> "NumPy and SciPy to refer to the projects. numpy and scipy to refer to
> the packages, specifically. When in doubt, use the former."
>
> I thought there was also another discussion about capital letters, but
> I don't find it.
>
> Josef
>
>
> On Mon, Aug 29, 2016 at 9:22 AM, Bartosz Telenczuk <mail at telenczuk.pl>
> wrote:
> > Hi,
> >
> > I would not mind any choice as long as it's consistent.
> >
> > I agree that using all-lowercase spelling may avoid some common errors.
> However,
> > PEP8 requires all module/package names to be lower case [1].  If we
> force the
> > name of the library and the corresponding package to be the same, all
> Python
> > libraries would be named in lowercase. This would not be the best choice
> for
> > libraries, which have multi-component names (like NumPy = Numerical
> Python).
> >
> > Note also that both the Wikipedia page [2] and the official NumPy logo
> [3] use
> > "NumPy" spelling.
> >
> > Some other popular [4] libraries use similar dichotomies:
> >
> > - Django - import django
> > - Cython - import cython
> > - PyYAML - import yaml
> > - scikit-learn - import sklearn
> >
> > On the other hand all standard Python libraries are lower-case named.
> >
> > Cheers,
> >
> > Bartosz
> >
> > [1] https://www.python.org/dev/peps/pep-0008/#package-and-module-names
> > [2] https://en.wikipedia.org/wiki/NumPy
> > [3] http://www.numpy.org/_static/numpy_logo.png
> > [4] http://pypi-ranking.info/alltime
> >
> >> On 08/29/2016 07:43 AM, mail at telenczuk.pl wrote:
> >> > What is the official spelling of NumPy/Numpy/numpy?
> >>
> >> IMHO it should be written numpy, because ...
> >>
> >>  >>> import NumPy
> >> Traceback (most recent call last):
> >>    File "<stdin>", line 1, in <module>
> >> ImportError: No module named NumPy
> >>  >>> import Numpy
> >> Traceback (most recent call last):
> >>    File "<stdin>", line 1, in <module>
> >> ImportError: No module named Numpy
> >>  >>> import numpy
> >>  >>>
> >>
> >> Phil
> >> _______________________________________________
> >> NumPy-Discussion mailing list
> >> NumPy-Discussion at scipy.org
> >> https://mail.scipy.org/mailman/listinfo/numpy-discussion
> >
> >
> > _______________________________________________
> > NumPy-Discussion mailing list
> > NumPy-Discussion at scipy.org
> > https://mail.scipy.org/mailman/listinfo/numpy-discussion
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160829/9b15087f/attachment.html>

From stefanv at berkeley.edu  Tue Aug 30 15:17:54 2016
From: stefanv at berkeley.edu (Stefan van der Walt)
Date: Tue, 30 Aug 2016 12:17:54 -0700
Subject: [Numpy-discussion] Which NumPy/Numpy/numpy spelling?
In-Reply-To: <57c41fc7bf6ce_4222a36c5419@Pct-EqAlain-Z30.notmuch>
References: <57c41fc7bf6ce_4222a36c5419@Pct-EqAlain-Z30.notmuch>
Message-ID: <1472584674.1559543.710714393.59C5761E@webmail.messagingengine.com>

On Mon, Aug 29, 2016, at 04:43, mail at telenczuk.pl wrote:
> The documentation is not consistent and it mixes both NumPy and Numpy.
> For example, the reference manual uses both spellings in the introduction
> paragraph (http://docs.scipy.org/doc/numpy/reference/):
> 
>     "This reference manual details functions, modules, and objects
>     included in Numpy, describing what they are and what they do. For
>     learning how to use NumPy, see also NumPy User Guide."

That's technically a bug: the official spelling is NumPy.  But, no one
really cares :)

St?fan



From sebastian at sipsolutions.net  Wed Aug 31 04:30:06 2016
From: sebastian at sipsolutions.net (Sebastian Berg)
Date: Wed, 31 Aug 2016 10:30:06 +0200
Subject: [Numpy-discussion] Which NumPy/Numpy/numpy spelling?
In-Reply-To: <1472584674.1559543.710714393.59C5761E@webmail.messagingengine.com>
References: <57c41fc7bf6ce_4222a36c5419@Pct-EqAlain-Z30.notmuch>
 <1472584674.1559543.710714393.59C5761E@webmail.messagingengine.com>
Message-ID: <1472632206.22852.0.camel@sipsolutions.net>

On Di, 2016-08-30 at 12:17 -0700, Stefan van der Walt wrote:
> On Mon, Aug 29, 2016, at 04:43, mail at telenczuk.pl wrote:
> > 
> > The documentation is not consistent and it mixes both NumPy and
> > Numpy.
> > For example, the reference manual uses both spellings in the
> > introduction
> > paragraph (http://docs.scipy.org/doc/numpy/reference/):
> > 
> > ????"This reference manual details functions, modules, and objects
> > ????included in Numpy, describing what they are and what they do.
> > For
> > ????learning how to use NumPy, see also NumPy User Guide."
> That's technically a bug: the official spelling is NumPy.??But, no
> one
> really cares :)
> 

I like the fact that this is all posted in:?[Numpy-discussion] ;).

- Sebastian

> St?fan
> 
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/c7f971a6/attachment.sig>

From mibieri at gmail.com  Wed Aug 31 07:28:21 2016
From: mibieri at gmail.com (Michael Bieri)
Date: Wed, 31 Aug 2016 13:28:21 +0200
Subject: [Numpy-discussion] State-of-the-art to use a C/C++ library from
	Python
Message-ID: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>

Hi all

There are several ways on how to use C/C++ code from Python with NumPy, as
given in http://docs.scipy.org/doc/numpy/user/c-info.html . Furthermore,
there's at least pybind11.

I'm not quite sure which approach is state-of-the-art as of 2016. How would
you do it if you had to make a C/C++ library available in Python right now?

In my case, I have a C library with some scientific functions on matrices
and vectors. You will typically call a few functions to configure the
computation, then hand over some pointers to existing buffers containing
vector data, then start the computation, and finally read back the data.
The library also can use MPI to parallelize.

Best regards,
Michael
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/3fd4ba9b/attachment.html>

From mviljamaa at kapsi.fi  Wed Aug 31 07:28:18 2016
From: mviljamaa at kapsi.fi (Matti Viljamaa)
Date: Wed, 31 Aug 2016 14:28:18 +0300
Subject: [Numpy-discussion] Include last element when subindexing numpy
	arrays?
Message-ID: <75413495-B5EB-4973-A0C1-2307C986D1F3@kapsi.fi>

Is there a clean way to include the last element when subindexing numpy arrays?
Since the default behaviour of numpy arrays is to omit the ?stop index?.

So for,

>>> A
array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
>>> A[0:5]
array([0, 1, 2, 3, 4])



From mviljamaa at kapsi.fi  Wed Aug 31 08:14:30 2016
From: mviljamaa at kapsi.fi (Matti Viljamaa)
Date: Wed, 31 Aug 2016 15:14:30 +0300
Subject: [Numpy-discussion] Why np.fft.rfftfreq only returns up to Nyqvist?
Message-ID: <D874B686-9B8D-475A-B388-42F98FD5A377@kapsi.fi>

What?s the reasonability of np.fft.rfftfreq returning frequencies only up to Nyquist, rather than for the full sample rate?

From robert.kern at gmail.com  Wed Aug 31 08:22:28 2016
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 31 Aug 2016 13:22:28 +0100
Subject: [Numpy-discussion] Why np.fft.rfftfreq only returns up to
	Nyqvist?
In-Reply-To: <D874B686-9B8D-475A-B388-42F98FD5A377@kapsi.fi>
References: <D874B686-9B8D-475A-B388-42F98FD5A377@kapsi.fi>
Message-ID: <CAF6FJiu9-qtFPhjQ-Vw661MzuHZytX=WT1w4b=bAwBZwx3NERw@mail.gmail.com>

On Wed, Aug 31, 2016 at 1:14 PM, Matti Viljamaa <mviljamaa at kapsi.fi> wrote:
>
> What?s the reasonability of np.fft.rfftfreq returning frequencies only up
to Nyquist, rather than for the full sample rate?

The answer to the question that you asked is that np.fft.rfft() only
computes values for frequencies only up to Nyquist, so np.fft.rfftfreq()
must give you the frequencies to match. I'm not sure if there is another
misunderstanding lurking that needs to be clarified.

--
Robert Kern
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/9cf3d20a/attachment.html>

From robert.kern at gmail.com  Wed Aug 31 08:22:54 2016
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 31 Aug 2016 13:22:54 +0100
Subject: [Numpy-discussion] Include last element when subindexing numpy
	arrays?
In-Reply-To: <75413495-B5EB-4973-A0C1-2307C986D1F3@kapsi.fi>
References: <75413495-B5EB-4973-A0C1-2307C986D1F3@kapsi.fi>
Message-ID: <CAF6FJitoMvTPaKDr6+60op_c8D3jhJMsPZBOGmxM7TbE5A20_w@mail.gmail.com>

On Wed, Aug 31, 2016 at 12:28 PM, Matti Viljamaa <mviljamaa at kapsi.fi> wrote:
>
> Is there a clean way to include the last element when subindexing numpy
arrays?
> Since the default behaviour of numpy arrays is to omit the ?stop index?.
>
> So for,
>
> >>> A
> array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
> >>> A[0:5]
> array([0, 1, 2, 3, 4])

A[5:]

--
Robert Kern
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/1c3d0953/attachment.html>

From robert.kern at gmail.com  Wed Aug 31 08:23:57 2016
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 31 Aug 2016 13:23:57 +0100
Subject: [Numpy-discussion] State-of-the-art to use a C/C++ library from
	Python
In-Reply-To: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
References: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
Message-ID: <CAF6FJiutvrgqZAVuYYEG8O2MM1J_Q19ACaea+1Ysiwcrm+i6Aw@mail.gmail.com>

On Wed, Aug 31, 2016 at 12:28 PM, Michael Bieri <mibieri at gmail.com> wrote:
>
> Hi all
>
> There are several ways on how to use C/C++ code from Python with NumPy,
as given in http://docs.scipy.org/doc/numpy/user/c-info.html . Furthermore,
there's at least pybind11.
>
> I'm not quite sure which approach is state-of-the-art as of 2016. How
would you do it if you had to make a C/C++ library available in Python
right now?
>
> In my case, I have a C library with some scientific functions on matrices
and vectors. You will typically call a few functions to configure the
computation, then hand over some pointers to existing buffers containing
vector data, then start the computation, and finally read back the data.
The library also can use MPI to parallelize.

I usually reach for Cython:

http://cython.org/
http://docs.cython.org/en/latest/src/userguide/memoryviews.html

--
Robert Kern
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/99125154/attachment.html>

From mviljamaa at kapsi.fi  Wed Aug 31 08:34:54 2016
From: mviljamaa at kapsi.fi (Matti Viljamaa)
Date: Wed, 31 Aug 2016 15:34:54 +0300
Subject: [Numpy-discussion] Include last element when subindexing numpy
	arrays?
In-Reply-To: <CAF6FJitoMvTPaKDr6+60op_c8D3jhJMsPZBOGmxM7TbE5A20_w@mail.gmail.com>
References: <75413495-B5EB-4973-A0C1-2307C986D1F3@kapsi.fi>
 <CAF6FJitoMvTPaKDr6+60op_c8D3jhJMsPZBOGmxM7TbE5A20_w@mail.gmail.com>
Message-ID: <931CA2D0-5618-4C87-8F85-F5E3B0112795@kapsi.fi>


> On 31 Aug 2016, at 15:22, Robert Kern <robert.kern at gmail.com> wrote:
> 
> On Wed, Aug 31, 2016 at 12:28 PM, Matti Viljamaa <mviljamaa at kapsi.fi <mailto:mviljamaa at kapsi.fi>> wrote:
> >
> > Is there a clean way to include the last element when subindexing numpy arrays?
> > Since the default behaviour of numpy arrays is to omit the ?stop index?.
> >
> > So for,
> >
> > >>> A
> > array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
> > >>> A[0:5]
> > array([0, 1, 2, 3, 4])
> 
> A[5:]
> 
> --
> Robert Kern

No that returns the subarray starting from index 5 to the end.

What I want to be able to return

array([0, 1, 2, 3, 4, 5])

(i.e. last element 5 included)

but without the funky A[0:6] syntax, which looks like it should return

array([0, 1, 2, 3, 4, 5, 6])

but since bumpy arrays omit the last index, returns

array([0, 1, 2, 3, 4, 5])

which syntactically would be more reasonable to be A[0:5].

> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/017a0f6a/attachment.html>

From robert.kern at gmail.com  Wed Aug 31 08:42:07 2016
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 31 Aug 2016 13:42:07 +0100
Subject: [Numpy-discussion] Include last element when subindexing numpy
	arrays?
In-Reply-To: <931CA2D0-5618-4C87-8F85-F5E3B0112795@kapsi.fi>
References: <75413495-B5EB-4973-A0C1-2307C986D1F3@kapsi.fi>
 <CAF6FJitoMvTPaKDr6+60op_c8D3jhJMsPZBOGmxM7TbE5A20_w@mail.gmail.com>
 <931CA2D0-5618-4C87-8F85-F5E3B0112795@kapsi.fi>
Message-ID: <CAF6FJiu_n-Fc+csBy5GPCJn=x-_UfR4cQce_m0PU55ZrDf644w@mail.gmail.com>

On Wed, Aug 31, 2016 at 1:34 PM, Matti Viljamaa <mviljamaa at kapsi.fi> wrote:
>
> On 31 Aug 2016, at 15:22, Robert Kern <robert.kern at gmail.com> wrote:
>
> On Wed, Aug 31, 2016 at 12:28 PM, Matti Viljamaa <mviljamaa at kapsi.fi>
wrote:
> >
> > Is there a clean way to include the last element when subindexing numpy
arrays?
> > Since the default behaviour of numpy arrays is to omit the ?stop index?.
> >
> > So for,
> >
> > >>> A
> > array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])
> > >>> A[0:5]
> > array([0, 1, 2, 3, 4])
>
> A[5:]
>
> --
> Robert Kern
>
> No that returns the subarray starting from index 5 to the end.
>
> What I want to be able to return
>
> array([0, 1, 2, 3, 4, 5])
>
> (i.e. last element 5 included)
>
> but without the funky A[0:6] syntax, which looks like it should return
>
> array([0, 1, 2, 3, 4, 5, 6])
>
> but since bumpy arrays omit the last index, returns
>
> array([0, 1, 2, 3, 4, 5])
>
> which syntactically would be more reasonable to be A[0:5].

Ah, I see what you are asking now.

The answer is "no"; this is just the way that slicing works in Python in
general. numpy merely follows suit. It is something that you will get used
to with practice. My sense of "funkiness" and "reasonableness" is the
opposite of yours, for instance.

--
Robert Kern
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/27c57097/attachment.html>

From ndbecker2 at gmail.com  Wed Aug 31 09:04:09 2016
From: ndbecker2 at gmail.com (Neal Becker)
Date: Wed, 31 Aug 2016 09:04:09 -0400
Subject: [Numpy-discussion] State-of-the-art to use a C/C++ library from
	Python
References: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
Message-ID: <nq6kk7$voh$1@blaine.gmane.org>

Michael Bieri wrote:

> Hi all
> 
> There are several ways on how to use C/C++ code from Python with NumPy, as
> given in http://docs.scipy.org/doc/numpy/user/c-info.html . Furthermore,
> there's at least pybind11.
> 
> I'm not quite sure which approach is state-of-the-art as of 2016. How
> would you do it if you had to make a C/C++ library available in Python
> right now?
> 
> In my case, I have a C library with some scientific functions on matrices
> and vectors. You will typically call a few functions to configure the
> computation, then hand over some pointers to existing buffers containing
> vector data, then start the computation, and finally read back the data.
> The library also can use MPI to parallelize.
> 
> Best regards,
> Michael

I prefer ndarray:
https://github.com/ndarray/ndarray



From mailinglists at xgm.de  Wed Aug 31 11:00:20 2016
From: mailinglists at xgm.de (Florian Lindner)
Date: Wed, 31 Aug 2016 17:00:20 +0200
Subject: [Numpy-discussion] Reading in a mesh file
Message-ID: <738e8474-524a-8cd0-9340-4b56d7a909da@xgm.de>

Hello,

I have mesh (more exactly: just a bunch of nodes) description with values associated to the nodes in a file, e.g. for a
3x3 mesh:

0   0   10
0   0.3 11
0   0.6 12
0.3 0   20
0.3 0.3 21
0.3 0.6 22
0.6 0   30
0.6 0.3 31
0.6 0.6 32

What is best way to read it in and get data structures like the ones I get from np.meshgrid?

Of course, I know about np.loadtxt, but I'm having trouble getting the resulting arrays (x, y, values) in the right form
and to retain association to the values.

Thanks,
Florian


From robert.kern at gmail.com  Wed Aug 31 11:06:57 2016
From: robert.kern at gmail.com (Robert Kern)
Date: Wed, 31 Aug 2016 16:06:57 +0100
Subject: [Numpy-discussion] Reading in a mesh file
In-Reply-To: <738e8474-524a-8cd0-9340-4b56d7a909da@xgm.de>
References: <738e8474-524a-8cd0-9340-4b56d7a909da@xgm.de>
Message-ID: <CAF6FJiuBdHDv2DVyzUhJtZfkh-UkcSxT5uygX=xJQs-CMxt-Bw@mail.gmail.com>

On Wed, Aug 31, 2016 at 4:00 PM, Florian Lindner <mailinglists at xgm.de>
wrote:
>
> Hello,
>
> I have mesh (more exactly: just a bunch of nodes) description with values
associated to the nodes in a file, e.g. for a
> 3x3 mesh:
>
> 0   0   10
> 0   0.3 11
> 0   0.6 12
> 0.3 0   20
> 0.3 0.3 21
> 0.3 0.6 22
> 0.6 0   30
> 0.6 0.3 31
> 0.6 0.6 32
>
> What is best way to read it in and get data structures like the ones I
get from np.meshgrid?
>
> Of course, I know about np.loadtxt, but I'm having trouble getting the
resulting arrays (x, y, values) in the right form
> and to retain association to the values.

For this particular case (known shape and ordering), this is what I would
do. Maybe throw in a .T or three depending on exactly how you want them to
be laid out.

[~/scratch]
|1> !cat mesh.txt

0   0   10
0   0.3 11
0   0.6 12
0.3 0   20
0.3 0.3 21
0.3 0.6 22
0.6 0   30
0.6 0.3 31
0.6 0.6 32

[~/scratch]
|2> nodes = np.loadtxt('mesh.txt')

[~/scratch]
|3> nodes
array([[  0. ,   0. ,  10. ],
       [  0. ,   0.3,  11. ],
       [  0. ,   0.6,  12. ],
       [  0.3,   0. ,  20. ],
       [  0.3,   0.3,  21. ],
       [  0.3,   0.6,  22. ],
       [  0.6,   0. ,  30. ],
       [  0.6,   0.3,  31. ],
       [  0.6,   0.6,  32. ]])

[~/scratch]
|4> reshaped = nodes.reshape((3, 3, -1))

[~/scratch]
|5> reshaped
array([[[  0. ,   0. ,  10. ],
        [  0. ,   0.3,  11. ],
        [  0. ,   0.6,  12. ]],

       [[  0.3,   0. ,  20. ],
        [  0.3,   0.3,  21. ],
        [  0.3,   0.6,  22. ]],

       [[  0.6,   0. ,  30. ],
        [  0.6,   0.3,  31. ],
        [  0.6,   0.6,  32. ]]])

[~/scratch]
|7> x = reshaped[..., 0]

[~/scratch]
|8> y = reshaped[..., 1]

[~/scratch]
|9> values = reshaped[..., 2]

[~/scratch]
|10> x
array([[ 0. ,  0. ,  0. ],
       [ 0.3,  0.3,  0.3],
       [ 0.6,  0.6,  0.6]])

[~/scratch]
|11> y
array([[ 0. ,  0.3,  0.6],
       [ 0. ,  0.3,  0.6],
       [ 0. ,  0.3,  0.6]])

[~/scratch]
|12> values
array([[ 10.,  11.,  12.],
       [ 20.,  21.,  22.],
       [ 30.,  31.,  32.]])

--
Robert Kern
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/84f69188/attachment.html>

From othalan at othalan.net  Wed Aug 31 13:08:58 2016
From: othalan at othalan.net (David Morris)
Date: Wed, 31 Aug 2016 20:08:58 +0300
Subject: [Numpy-discussion] State-of-the-art to use a C/C++ library from
	Python
In-Reply-To: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
References: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
Message-ID: <CAHk-ryCWWmMqwbxXzSOAVw2SvXNdH3zr50SgVCAmNXh7yKe8Vw@mail.gmail.com>

On Wed, Aug 31, 2016 at 2:28 PM, Michael Bieri <mibieri at gmail.com> wrote:

> Hi all
>
> There are several ways on how to use C/C++ code from Python with NumPy, as
> given in http://docs.scipy.org/doc/numpy/user/c-info.html . Furthermore,
> there's at least pybind11.
>
> I'm not quite sure which approach is state-of-the-art as of 2016. How
> would you do it if you had to make a C/C++ library available in Python
> right now?
>
> In my case, I have a C library with some scientific functions on matrices
> and vectors. You will typically call a few functions to configure the
> computation, then hand over some pointers to existing buffers containing
> vector data, then start the computation, and finally read back the data.
> The library also can use MPI to parallelize.
>

I have been delighted with Cython for this purpose.  Great integration with
NumPy (you can access numpy arrays directly as C arrays), very python like
syntax and amazing performance.

Good luck,

David
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/4ccd5338/attachment.html>

From sylvain.corlay at gmail.com  Wed Aug 31 13:14:24 2016
From: sylvain.corlay at gmail.com (Sylvain Corlay)
Date: Wed, 31 Aug 2016 19:14:24 +0200
Subject: [Numpy-discussion] State-of-the-art to use a C/C++ library from
	Python
In-Reply-To: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
References: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
Message-ID: <CAK=Phk68QjNPj2MpkijXN4JGqvHQGgv8Q744LqPdv+hBDZJHJw@mail.gmail.com>

+1 on pybind11.

Sylvain

On Wed, Aug 31, 2016 at 1:28 PM, Michael Bieri <mibieri at gmail.com> wrote:

> Hi all
>
> There are several ways on how to use C/C++ code from Python with NumPy, as
> given in http://docs.scipy.org/doc/numpy/user/c-info.html . Furthermore,
> there's at least pybind11.
>
> I'm not quite sure which approach is state-of-the-art as of 2016. How
> would you do it if you had to make a C/C++ library available in Python
> right now?
>
> In my case, I have a C library with some scientific functions on matrices
> and vectors. You will typically call a few functions to configure the
> computation, then hand over some pointers to existing buffers containing
> vector data, then start the computation, and finally read back the data.
> The library also can use MPI to parallelize.
>
> Best regards,
> Michael
>
>
>
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/881c2df7/attachment.html>

From nevion at gmail.com  Wed Aug 31 13:20:28 2016
From: nevion at gmail.com (Jason Newton)
Date: Wed, 31 Aug 2016 13:20:28 -0400
Subject: [Numpy-discussion] State-of-the-art to use a C/C++ library from
	Python
In-Reply-To: <CAHk-ryCWWmMqwbxXzSOAVw2SvXNdH3zr50SgVCAmNXh7yKe8Vw@mail.gmail.com>
References: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
 <CAHk-ryCWWmMqwbxXzSOAVw2SvXNdH3zr50SgVCAmNXh7yKe8Vw@mail.gmail.com>
Message-ID: <CAGou9MiCwQTEuSr3Q-vYVjUwuGUBHZxKFDzNTrDMfyw3sPEChA@mail.gmail.com>

I just wanted to follow up on the C++ side of OP email - Cython has quite a
few difficulties working with C++ code at the moment.  It's really more of
a C solution most of the time and you must split things up into a mostly C
call interface (that is the C code Cython can call) and limit
exposure/complications with templates and  complex C++11+ constructs.  This
may change in the longer term but in the near, that is the state.

I used to use Boost.Python but I'm getting my feet wet with Pybind (which
is basically the same api but works more as you expect it to with it's
signature/type plumbing  (including std::shared_ptr islanding), with some
other C++11 based improvements, and is header only + submodule friendly!).
I also remembered ndarray thanks to Neal's post but I haven't figured out
how to leverage it better than pybind, at the moment.  I'd be interested to
see ndarray gain support for pybind interoperability...

-Jason

On Wed, Aug 31, 2016 at 1:08 PM, David Morris <othalan at othalan.net> wrote:

> On Wed, Aug 31, 2016 at 2:28 PM, Michael Bieri <mibieri at gmail.com> wrote:
>
>> Hi all
>>
>> There are several ways on how to use C/C++ code from Python with NumPy,
>> as given in http://docs.scipy.org/doc/numpy/user/c-info.html .
>> Furthermore, there's at least pybind11.
>>
>> I'm not quite sure which approach is state-of-the-art as of 2016. How
>> would you do it if you had to make a C/C++ library available in Python
>> right now?
>>
>> In my case, I have a C library with some scientific functions on matrices
>> and vectors. You will typically call a few functions to configure the
>> computation, then hand over some pointers to existing buffers containing
>> vector data, then start the computation, and finally read back the data.
>> The library also can use MPI to parallelize.
>>
>
> I have been delighted with Cython for this purpose.  Great integration
> with NumPy (you can access numpy arrays directly as C arrays), very python
> like syntax and amazing performance.
>
> Good luck,
>
> David
>
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/c26dce9e/attachment.html>

From insertinterestingnamehere at gmail.com  Wed Aug 31 14:17:56 2016
From: insertinterestingnamehere at gmail.com (Ian Henriksen)
Date: Wed, 31 Aug 2016 18:17:56 +0000
Subject: [Numpy-discussion] State-of-the-art to use a C/C++ library from
	Python
In-Reply-To: <CAGou9MiCwQTEuSr3Q-vYVjUwuGUBHZxKFDzNTrDMfyw3sPEChA@mail.gmail.com>
References: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
 <CAHk-ryCWWmMqwbxXzSOAVw2SvXNdH3zr50SgVCAmNXh7yKe8Vw@mail.gmail.com>
 <CAGou9MiCwQTEuSr3Q-vYVjUwuGUBHZxKFDzNTrDMfyw3sPEChA@mail.gmail.com>
Message-ID: <CAPWbfpR47cdOmL6BjEOn2ZTvGD+m667dhjXTFTO1+-bSpCfC_Q@mail.gmail.com>

We use Cython very heavily in DyND's Python bindings. It has worked well
for us
even when working with some very modern C++. That said, a lot depends on
exactly which C++ features you want to expose as a part of the interface.
Interfaces that require things like non-type template parameters or variadic
templates will often require a some extra C++ code to work them in to
something
that Cython can understand. In my experience, those particular limitations
haven't
been that hard to work with.
Best,
Ian Henriksen

On Wed, Aug 31, 2016 at 12:20 PM Jason Newton <nevion at gmail.com> wrote:

> I just wanted to follow up on the C++ side of OP email - Cython has quite
> a few difficulties working with C++ code at the moment.  It's really more
> of a C solution most of the time and you must split things up into a mostly
> C call interface (that is the C code Cython can call) and limit
> exposure/complications with templates and  complex C++11+ constructs.  This
> may change in the longer term but in the near, that is the state.
>
> I used to use Boost.Python but I'm getting my feet wet with Pybind (which
> is basically the same api but works more as you expect it to with it's
> signature/type plumbing  (including std::shared_ptr islanding), with some
> other C++11 based improvements, and is header only + submodule friendly!).
> I also remembered ndarray thanks to Neal's post but I haven't figured out
> how to leverage it better than pybind, at the moment.  I'd be interested to
> see ndarray gain support for pybind interoperability...
>
> -Jason
>
> On Wed, Aug 31, 2016 at 1:08 PM, David Morris <othalan at othalan.net> wrote:
>
>> On Wed, Aug 31, 2016 at 2:28 PM, Michael Bieri <mibieri at gmail.com> wrote:
>>
>>> Hi all
>>>
>>> There are several ways on how to use C/C++ code from Python with NumPy,
>>> as given in http://docs.scipy.org/doc/numpy/user/c-info.html .
>>> Furthermore, there's at least pybind11.
>>>
>>> I'm not quite sure which approach is state-of-the-art as of 2016. How
>>> would you do it if you had to make a C/C++ library available in Python
>>> right now?
>>>
>>> In my case, I have a C library with some scientific functions on
>>> matrices and vectors. You will typically call a few functions to configure
>>> the computation, then hand over some pointers to existing buffers
>>> containing vector data, then start the computation, and finally read back
>>> the data. The library also can use MPI to parallelize.
>>>
>>
>> I have been delighted with Cython for this purpose.  Great integration
>> with NumPy (you can access numpy arrays directly as C arrays), very python
>> like syntax and amazing performance.
>>
>> Good luck,
>>
>> David
>>
>> _______________________________________________
>> NumPy-Discussion mailing list
>> NumPy-Discussion at scipy.org
>> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>>
>>
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/32d18d44/attachment.html>

From nevion at gmail.com  Wed Aug 31 16:57:06 2016
From: nevion at gmail.com (Jason Newton)
Date: Wed, 31 Aug 2016 16:57:06 -0400
Subject: [Numpy-discussion] State-of-the-art to use a C/C++ library from
	Python
In-Reply-To: <CAPWbfpR47cdOmL6BjEOn2ZTvGD+m667dhjXTFTO1+-bSpCfC_Q@mail.gmail.com>
References: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
 <CAHk-ryCWWmMqwbxXzSOAVw2SvXNdH3zr50SgVCAmNXh7yKe8Vw@mail.gmail.com>
 <CAGou9MiCwQTEuSr3Q-vYVjUwuGUBHZxKFDzNTrDMfyw3sPEChA@mail.gmail.com>
 <CAPWbfpR47cdOmL6BjEOn2ZTvGD+m667dhjXTFTO1+-bSpCfC_Q@mail.gmail.com>
Message-ID: <CAGou9MipAeUF-wqKxM2PL+yjOH3BgZw+xBnrh7CtT9nBHh8ZmA@mail.gmail.com>

Hey Ian - I hope I gave Cython a fair comment, but I have to add the
disclaimer that your capability to understand/implement those
solutions/workarounds in that project is greatly enhanced from your knowing
the innards of Cython from being core developer on the Cython project. This
doesn't detract from DyDN's accomplishments (if nothing it means Cython
users should look there for how to use C++ with Cython and the workarounds
used + shortcomings) but I would not expect not everyone would want to jump
through those hoops to get things working without a firm understanding of
Cython's edges, and all this potential for special/hack adaption code is
still something to keep in mind when comparing to something more straight
forward and easier to understand coming from a more pure C/C++ side, where
things are a bit more dangerous and fairly more verbose but make play with
the language and environment first-class (like Boost.Python/pybind).  Since
this thread is a survey over state and options it's my intent just to make
sure readers have something bare in mind for current pros/cons of the
approaches.

-Jason

On Wed, Aug 31, 2016 at 2:17 PM, Ian Henriksen <
insertinterestingnamehere at gmail.com> wrote:

> We use Cython very heavily in DyND's Python bindings. It has worked well
> for us
> even when working with some very modern C++. That said, a lot depends on
> exactly which C++ features you want to expose as a part of the interface.
> Interfaces that require things like non-type template parameters or
> variadic
> templates will often require a some extra C++ code to work them in to
> something
> that Cython can understand. In my experience, those particular limitations
> haven't
> been that hard to work with.
> Best,
> Ian Henriksen
>
>
> On Wed, Aug 31, 2016 at 12:20 PM Jason Newton <nevion at gmail.com> wrote:
>
>> I just wanted to follow up on the C++ side of OP email - Cython has quite
>> a few difficulties working with C++ code at the moment.  It's really more
>> of a C solution most of the time and you must split things up into a mostly
>> C call interface (that is the C code Cython can call) and limit
>> exposure/complications with templates and  complex C++11+ constructs.  This
>> may change in the longer term but in the near, that is the state.
>>
>> I used to use Boost.Python but I'm getting my feet wet with Pybind (which
>> is basically the same api but works more as you expect it to with it's
>> signature/type plumbing  (including std::shared_ptr islanding), with some
>> other C++11 based improvements, and is header only + submodule friendly!).
>> I also remembered ndarray thanks to Neal's post but I haven't figured out
>> how to leverage it better than pybind, at the moment.  I'd be interested to
>> see ndarray gain support for pybind interoperability...
>>
>> -Jason
>>
>> On Wed, Aug 31, 2016 at 1:08 PM, David Morris <othalan at othalan.net>
>> wrote:
>>
>>> On Wed, Aug 31, 2016 at 2:28 PM, Michael Bieri <mibieri at gmail.com>
>>> wrote:
>>>
>>>> Hi all
>>>>
>>>> There are several ways on how to use C/C++ code from Python with NumPy,
>>>> as given in http://docs.scipy.org/doc/numpy/user/c-info.html .
>>>> Furthermore, there's at least pybind11.
>>>>
>>>> I'm not quite sure which approach is state-of-the-art as of 2016. How
>>>> would you do it if you had to make a C/C++ library available in Python
>>>> right now?
>>>>
>>>> In my case, I have a C library with some scientific functions on
>>>> matrices and vectors. You will typically call a few functions to configure
>>>> the computation, then hand over some pointers to existing buffers
>>>> containing vector data, then start the computation, and finally read back
>>>> the data. The library also can use MPI to parallelize.
>>>>
>>>
>>> I have been delighted with Cython for this purpose.  Great integration
>>> with NumPy (you can access numpy arrays directly as C arrays), very python
>>> like syntax and amazing performance.
>>>
>>> Good luck,
>>>
>>> David
>>>
>>> _______________________________________________
>>> NumPy-Discussion mailing list
>>> NumPy-Discussion at scipy.org
>>> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>>>
>>>
>> _______________________________________________
>> NumPy-Discussion mailing list
>> NumPy-Discussion at scipy.org
>> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>>
>
> _______________________________________________
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
> https://mail.scipy.org/mailman/listinfo/numpy-discussion
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/a5839ccf/attachment.html>

From stefanv at berkeley.edu  Wed Aug 31 18:04:33 2016
From: stefanv at berkeley.edu (Stefan van der Walt)
Date: Wed, 31 Aug 2016 15:04:33 -0700
Subject: [Numpy-discussion] State-of-the-art to use a C/C++ library from
 Python
In-Reply-To: <CAGou9MipAeUF-wqKxM2PL+yjOH3BgZw+xBnrh7CtT9nBHh8ZmA@mail.gmail.com>
References: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
 <CAHk-ryCWWmMqwbxXzSOAVw2SvXNdH3zr50SgVCAmNXh7yKe8Vw@mail.gmail.com>
 <CAGou9MiCwQTEuSr3Q-vYVjUwuGUBHZxKFDzNTrDMfyw3sPEChA@mail.gmail.com>
 <CAPWbfpR47cdOmL6BjEOn2ZTvGD+m667dhjXTFTO1+-bSpCfC_Q@mail.gmail.com>
 <CAGou9MipAeUF-wqKxM2PL+yjOH3BgZw+xBnrh7CtT9nBHh8ZmA@mail.gmail.com>
Message-ID: <1472681073.2686064.712068009.73919B3D@webmail.messagingengine.com>

On Wed, Aug 31, 2016, at 13:57, Jason Newton wrote:
> Hey Ian - I hope I gave Cython a fair comment, but I have to add the
> disclaimer that your capability to understand/implement those
> solutions/workarounds in that project is greatly enhanced from your
> knowing the innards of Cython from being core developer on the Cython
> project. This doesn't detract from DyDN's accomplishments (if nothing
> it means Cython users should look there for how to use C++ with Cython
> and the workarounds used + shortcomings) but I would not expect not
> everyone would want to jump through those hoops to get things working
> without a firm understanding of Cython's edges, and all this potential
> for special/hack adaption code is still something to keep in mind when
> comparing to something more straight forward and easier to understand
> coming from a more pure C/C++ side, where things are a bit more
> dangerous and fairly more verbose but make play with the language and
> environment first-class (like Boost.Python/pybind).  Since this thread
> is a survey over state and options it's my intent just to make sure
> readers have something bare in mind for current pros/cons of the
> approaches.

There are many teaching resources available for Cython, after which
exposure to sharp edges may be greatly reduced.  See, e.g.,

https://github.com/stefanv/teaching/blob/master/2014_assp_split_cython/slides/split2014_cython.pdf

and accompanying problems and exercises at

https://github.com/stefanv/teaching/tree/master/2014_assp_split_cython

St?fan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/e1155b2d/attachment.html>

From insertinterestingnamehere at gmail.com  Wed Aug 31 18:53:05 2016
From: insertinterestingnamehere at gmail.com (Ian Henriksen)
Date: Wed, 31 Aug 2016 22:53:05 +0000
Subject: [Numpy-discussion] State-of-the-art to use a C/C++ library from
	Python
In-Reply-To: <CAGou9MipAeUF-wqKxM2PL+yjOH3BgZw+xBnrh7CtT9nBHh8ZmA@mail.gmail.com>
References: <CAJQ3EXaMjUnqc=2wUoP+mNV-5i6fGh8L=vzf7fP+AnVp421Mig@mail.gmail.com>
 <CAHk-ryCWWmMqwbxXzSOAVw2SvXNdH3zr50SgVCAmNXh7yKe8Vw@mail.gmail.com>
 <CAGou9MiCwQTEuSr3Q-vYVjUwuGUBHZxKFDzNTrDMfyw3sPEChA@mail.gmail.com>
 <CAPWbfpR47cdOmL6BjEOn2ZTvGD+m667dhjXTFTO1+-bSpCfC_Q@mail.gmail.com>
 <CAGou9MipAeUF-wqKxM2PL+yjOH3BgZw+xBnrh7CtT9nBHh8ZmA@mail.gmail.com>
Message-ID: <CAPWbfpTr1Kr_JzMu+AcKo1tiRywjDLNKjaX0_-89wLU_Y5eJ+g@mail.gmail.com>

On Wed, Aug 31, 2016 at 3:57 PM Jason Newton <nevion at gmail.com> wrote:

> Hey Ian - I hope I gave Cython a fair comment, but I have to add the
> disclaimer that your capability to understand/implement those
> solutions/workarounds in that project is greatly enhanced from your knowing
> the innards of Cython from being core developer on the Cython project. This
> doesn't detract from DyDN's accomplishments (if nothing it means Cython
> users should look there for how to use C++ with Cython and the workarounds
> used + shortcomings) but I would not expect not everyone would want to jump
> through those hoops to get things working without a firm understanding of
> Cython's edges, and all this potential for special/hack adaption code is
> still something to keep in mind when comparing to something more straight
> forward and easier to understand coming from a more pure C/C++ side, where
> things are a bit more dangerous and fairly more verbose but make play with
> the language and environment first-class (like Boost.Python/pybind).  Since
> this thread is a survey over state and options it's my intent just to make
> sure readers have something bare in mind for current pros/cons of the
> approaches.
>
> -Jason
>

No offense taken at all. I'm actually not a Cython developer, just a
frequent
contributor. That said, knowing the compiler internals certainly helps when
finding
workarounds and building intermediate interfaces. My main point was just
that, in
my experience, Cython has worked well for many things beyond plain C
interfaces
and that workarounds (hackery entirely aside) for any missing features are
usually
manageable. Given that my perspective is a bit different in that regard, it
seemed
worth chiming in on the discussion. I suppose the moral of the story is
that there's
still not a clear cut "best" way of building wrappers and that your mileage
may vary
depending on what features you need.
Thanks,
Ian Henriksen
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20160831/9365963f/attachment.html>