Audit Trail:

• simplify the WinPython sources publications (all packages.srcreq in one zip), that may be of use for Zero complementation
• scipy 0.18.0
• PyQt5.7.0 +sip 4.18.1
• Altair 1.0
• manual follow-up updates:
  • Spyder3dev 20160729
  • pymc3dev 20160729
• Bokeh 0.12.1 (last minute arrival)

(Build1: july 30th, a promizing PyQt5.7)

• notebook-4.2.2, promtp-toolkit-1.0.5
• xarray-0.8.1, ipykernel 4.4.1, pyzmq-15.4, Keras 1.0.7, nbformat 4.1, jupyter-core-4.1.1
• R procedure update (waiting rpy2.8.2 or manual patch)
• Spyder3dev 20160810
• pythonnet artifact of pull request 219 as of 20160811 https://ci.appveyor.com/project/TonyRoberts/pythonnet-480xs/build/job/m11tbw35v4nka809/artifacts
• ipython 5.1
• pandoc-1.7.2, pandas-datareader 0.2.1, requests-file 1.4
• self-installed Microsoft vs2015 "visualcppbuildtools_full.exe" as workaround for Python3.5 and other wheel building issues
• rebuild spyder-dev wheel with sphinx doc (no hack required with vs2015)
• put WinPython3.5 "cython" default on "vs", as "mingw" is apparently not for tomorrow

(Build2: august 17th)

• ipyparallel 5.2, spyder-3.0.0b5
• bqplot-0.8.2, holoviews-1.6.2
• (won't wait) statsmodels 0.8.0,
• (won't wait) jupyter_client 4.4, nbconvert 5, traitlets 4.3

(rc : august 24-25th )

Work items:

• WinPythonR ?
• Mingwpy effort seems to have stopped
• Matplotlib 2.0.0 + ipympl at next release (1.5.3 not showing up)
• breaking-change scikit-learn +scikit-optimize rather at next release
• rise ? (probably not: too much change ongoing with jupyterlab)
• Microsoft was suppose to do something august 4th ? https://connect.microsoft.com/VisualStudio/feedback/details/2869042/new-internal-compiler-error-only-in-vs2015-update-3 .... nothing
• simplify all the things / get clother to pip/pip+conda-forge
• Julia anywhere / WinPythonJl ? (Julia 1.0 still one year away, numba+pyjion+mingwpy may steal the need, https://www.youtube.com/watch?v=5gXMpbY1kJY)
• Tensorflow / Cuda experiments ?
• simplify build and easing auto-upgrades:
  • progress towards building Zero more easily: drop not-"packages.srcreq" packages ?
  • progress towards building-up from Zero: using packages.srcreq.zip ?
• continue re-tuning documentation notebooks in the spirit of (https://youtu.be/2-SB2Soxl30?t=899):
• a smaller WinPython_checker = "new user" /"clear value proposition"/"great onboarding"
• one detailed Notebook per thematic, = "learners"/"great tutorials"/'easy setup"
• one more detailed notebook per thematic ="engaged users"/"many taks & use-cases"/"community"
• one level of gory details = "contributors"/"help newcomers"/"improve and expand solutions"
• rework roadmap for end 2016:
  • Qt4 is detrimental to Qt5 efficient use by other Packages: we may stop it sooner
  • better-Qt-less WinPythonControlPanel or find a Zero simple "other-way"
• see virtualenv as Spyder 3.0 may acheive a projet functionality with virtualenv selection Design for Projects in 3.0 spyder-ide/spyder#2460
  To Consider:
• Windows 10.2 may bring some help.
• read progress of xlwings and other Excel glueware (Anaconda 'Fusion' ??):
  • no other free glueware
  • xlwings 0.7.1 has still some critics (to look again before release, tempting)
• drop all Qtdesigner.exe and QTLinguist.exe going forward
• sqlite_web experience ...
• explore conda-forge
• test if lasagne peacock fast example can work 'not slow' on a cpu: https://github.com/stonebig/deep-learning-tutorial-pydata2016/blob/master/TUTORIAL%2002%20-%20Using%20a%20pretrained%20VGG-19%20conv-net%20to%20find%20a%20peacock.ipynb
• arrayfire: my opencl GPU card may be of use (but no easy connexion to other libs yet)
• feather-format +Arrow / sframe / ... : a multi-core dataframe fundation (the first one being from Wes+Hadley, ... best possible parents), feather needs C++11 so won't be python3.4
• javascript stack: "nod-gyp for windows" progress ?
• continue re-tuning documentation notebooks in the spirit of (https://youtu.be/2-SB2Soxl30?t=899):
  • a smaller WinPython_checker = "new user" /"clear value proposition"/"great onboarding"
  • one detailed Notebook per thematic, = "learners"/"great tutorials"/'easy setup"
  • one more detailed notebook per thematic ="engaged users"/"many taks & use-cases"/"community"
  • one level of gory details = "contributors"/"help newcomers"/"improve and expand solutions"
• keep a distant eye on Spacy KPI: Number of devs, multi-lingual
• review future entrants and leaving:
  • NLP: swaping nltk for gensim or spacy ? (https://youtu.be/QgjkgNVXSjQ?list=PLGVZCDnMOq0rzDLHi5WxWmN5vueHU5Ar7&t=1767 )
  • BigTables: sframe (or ibis+things , or dynd, or Dask.Array...) Do you have any knowledge/opnion on this ? (disclaimer I'm not a NLP user at the moment)
• scrapy 1.2 (but will be heavy with twisted)
• do an experiment virtualenv 14.0.1 the package
• having a look on Pandastable
• having a look on Apache Zeppelin & Arrow
• having a look on PowerBI Desktop
• we may not just yet remove QT_API setting plumbery for Qt5 as spyder should handle it nicely now ? /2888546de61620f5e802ca3a79cf8d9f5437da5c (sept13th, spyder2.3.7pre)
• support a 97% "requirements" based build process (when Ipython 5, no more pyreadline)
• look for a nbbrowserpdf fork as ghost.py looks dead for windows (pdf generation without pandoc)
• purchase a PC with a nvidia card (or never launch even baby tests on Theano)
  • checklist pain points to remove at each step so we could build in the cloud:

    • split in 3:
      

    •    building WinPythonZero,
      

    •    pushing wheels on Warehouse
      

    •    completing over WinPythonZero, downloading wheels from Warehouse,
      

• git back ?
• baby-step3 on AppVeyor: getting back result
• baby-step4 on AppVeyor: downloading a Christoph Gohlke package
• baby-step5 on AppVeyor: building a full winpython requires a mini-winpython ?

how to launch ,like a jupyter ,matplotib etc ? ,sorry for bad english

Hi @mzakaria130 ,

If you have download and installed a "full" version (for example WinPython-64bit-3.4.4.3Qt5.exe), you should:

• see an icon "Jupyter Notebook" in the directory where you installed WinPython
  
• if you "double click" on it, after a 10 seconds a browser window "Home" should appear with the web address " http://localhost:8888/tree"
• then:
  • click on the "docs" directory presented in this page and choose ("click" then "play") an example
  • or click on the top right "new" button, choose "python3", past this code in the available cell, click on ", click on the "play" button , and see result:

%matplotlib inline
# example from Matplotlib gallery of examples
# **http://matplotlib.org/examples/lines_bars_and_markers/line_demo_dash_control.html**
import numpy as np
import matplotlib.pyplot as plt
x = np.linspace(0, 10)
line, = plt.plot(x, np.sin(x), '--', linewidth=2)
dashes = [10, 5, 100, 5]  # 10 points on, 5 off, 100 on, 5 off
line.set_dashes(dashes)
plt.show()

You may find some tutorials on the web like https://www.youtube.com/watch?v=p7Mj-4kASmI
(google with keywords "matplotlib" "tutorial" "youtube")

Can be displayed at the spyder IDE ?

Yes, in "Console Ipython", with the same code

Can I get a code to display a list of other tools in the IDE spyder

hi @mzakaria130 . This place is not the right place for broad questions about Spyder. Please go ask them:

• on Spyder forum https://groups.google.com/forum/#!forum/spyderlib
• or on stackoverflow http://stackoverflow.com/questions/tagged/spyder

build1 (2016-07-30): Scipy 0.18.0, PyQt5.7, Altair

main features (since WinPython 2016-03):

• PyQt5.7 fits nicely with Spyder3.0 dev-20160729
• new package: Altair/Vega
• package updates: Scipy 0.18.0, Holoviews 1.6.1, Dask 0.10.2, pymc3 -20260729
• bokeh 0.12.1 (re-spinning build1 for this just-baked wheel per cgohlke)

Areas of interest for testers:

• PyQt 5.7 (PyQt5.5 for Python3.4)
• Bokeh 0.12.1

Next focus:

• awaited bug fix of IPython5, ipykernel, Matplotlib1.5.2

  MD5	SHA-1	SHA-256	Binary	Size
  73dc2a42cc7a0719db6a474f58222c52	0e732257546cd1e853a4e9744cf8f8a7c157ce84	535c5009d5e6befb8baa99513625d29966636342f67b15f42cfddc2c11f5b4e5	WinPython-64bit-3.5.2.2Qt5b1.exe	269 869 901 Bytes
  4813823065501bf3b9453c329551966e	b27a32ea6ec60ac9e1dcbfd6b7be0287d429e8c7	60868e3f57f3814a92895c4099329fafea5e550cb2d0a5ee427f8a31c90e01af	WinPython-32bit-3.5.2.2Qt5b1.exe	232 187 795 Bytes
  16958b9547094c4a84b2a2a9ae44338c	f5baad7cd5fcb7d9d7aa8c4be969efe987b6fd2f	359209f87ab0a3a6460c2c316a93f57c43cff7d5cd70baf3c88b32e5e392be64	WinPython-64bit-3.4.4.4Qt5b1.exe	299 608 086 Bytes
  1e6e6e42199b00f047f234d11368c682	c9b464ed840ef30bf39916efdeb7fe263b23bf6b	a09ef0b2f1bc4fffc18f0e2d12a4092f5587ce8fcbb2c61a7a695c03a791e09a	WinPython-32bit-3.4.4.4Qt5b1.exe	254 904 758 Bytes

History of changes for WinPython 3.5.2.2Qt5

The following changes were made to WinPython distribution since version 3.5.2.1Qt5.

Python packages

New packages:

• altair 1.0.0 (High-level declarative visualization library for Python)
• vega 0.4.1 (An IPython/ Jupyter widget for Vega and Vega-Lite)
• webencodings 0.5 (Character encoding aliases for legacy web content)

Upgraded packages:

• alabaster 0.7.8 → 0.7.9 (A configurable sidebar-enabled Sphinx theme)
• astroid 1.4.7 → 1.4.8 (Rebuild a new abstract syntax tree from Python's ast (required for pylint))
• babel 2.3.2 → 2.3.4 (Internationalization utilities)
• beautifulsoup4 4.4.1 → 4.5.0 (Screen-scraping library)
• blosc 1.3.3 → 1.4.1 (Blosc data compressor)
• bokeh 0.11.1 → 0.12.1 (Statistical and novel interactive HTML plots for Python)
• dask 0.10.1 → 0.10.2 (Minimal task scheduling abstraction)
• distributed 1.11.2 → 1.11.3 (Distributed computing)
• holoviews 1.6.0 → 1.6.1 (Composable, declarative data structures for building complex visualizations easily.)
• html5lib 0.9999999 → 0.999999999 (HTML parser based on the WHATWG HTML specification)
• lmfit 0.9.3 → 0.9.5 (Least-Squares Minimization with Bounds and Constraints)
• lxml 3.6.0 → 3.6.1 (Powerful and Pythonic XML processing library combining libxml2/libxslt with the ElementTree API.)
• nbsphinx 0.2.8 → 0.2.9 (Jupyter Notebook Tools for Sphinx)
• partd 0.3.4 → 0.3.5 (Appendable key-value storage)
• pylint 1.6.1 → 1.6.4 (Logilab code analysis module: analyzes Python source code looking for bugs and signs of poor quality)
• pymongo 3.2.2 → 3.3.0 (Python driver for MongoDB http://www.mongodb.org)
• pyqt5 5.6 → 5.7 (Python bindings for the Qt cross platform GUI toolkit)
• pytz 2016.4 → 2016.6.1 (World Timezone Definitions for Python)
• requests-toolbelt 0.6.2 → 0.7.0 (Requests is an Apache2 Licensed HTTP library, written in Python, for human beings.)
• scipy 0.18.0rc2 → 0.18.0 (SciPy: Scientific Library for Python (advanced math, signal processing, optimization, statistics, ...))
• setuptools 24.0.2 → 25.1.1 (Download, build, install, upgrade, and uninstall Python packages - easily)
• sip 4.18 → 4.18.1 (Python extension module generator for C and C++ libraries)
• sqlparse 0.1.19 → 0.2.0 (Non-validating SQL parser)
• tornado 4.4 → 4.4.1 (Scalable, non-blocking web server and tools (required for IPython notebook))
• zarr 1.0.0 → 1.1.0 (A minimal implementation of chunked, compressed, N-dimensional arrays for Python.)

Hi stonebig,

I noticed in WinPython3.5.2.1Qt5, tools like the Qt Designer are gone. I read it has something to do with the PyQt5 wheel. Are these tools going to be back with the upcoming version?

Thanks!

hi @J3ronimo

Apparently, they are gone forever from PyQt official package by Phil Thompson.

My understanding of Phil is that:

• his PyQt wheel will only contain PyQt basic part,
• the rest should be complementary wheels,
• he will provide some: sip, PyQtChart, PyQtDataVisualization, PyQtPurchasing, QScintilla
• others trustable wheel providers are also free to provide:
  • alternate PyQt wheel with more "features" inside,
  • or complementary wheels (https://www.riverbankcomputing.com/pipermail/pyqt/2016-August/037852.html)

The basic part contains a browser, so I won't complain, and it may be the only complement some packages like Flexx will need in PyQt.

Oh okay. It was very handy though to have it working out of the box alongside with PyQt, especially the Designer. Would be nice if someone came up with a working approach for a complementary wheel soon, at least for the Designer, and then maybe this could be added back into WinPython.

build2 (2016-08-17): IPython-5.1, Pandoc 1.17.2

main features (since build1):

• new package: pypandoc
• package updates: ipython-5.1. prompt_toolkit-1.0.6, pandas_datareader, xarray
• tools updates: Pandoc 1.17.2

Areas of interest for testers:

• does the new Pandoc works ok ? interest of pypandoc ?
• "mingw" is no more the default compiler for python3.5, because of Python 3.5 wheels mingwpy/mingwpy#25 (comment) . Shifting back to mingw is a matter of clicking on the "scripts\make_cython_use_mingw.bat"

Next focus:

• awaited bug fix of jupyter_client 4.4, nbconvert 5, traitlets 4.3, statsmodels2.0

  MD5	SHA-1	SHA-256	Binary	Size
  6e82df63c967ce042bcc62546c6772b8	071db15a1b580b8c1a822d98d6fbd729c154758e	4d21767fdb871eb913cde1b61884d4aada4710aacfc3b754abdd3456e51044b0	WinPython-64bit-2.7.12.3Zerob2.exe	22 231 543 Bytes
  d3c2724144b97e7a4446b5c1b43b3b1a	0e1e1027a29e0cb72bd0c3b5ccf62c5e7ea01489	0c47e7e7ba78736d9edfb16b236c2435af57ba5dddd83b8980cb8566d7981d30	WinPython-32bit-2.7.12.3Zerob2.exe	21 510 150 Bytes
  7c0844cc3e61d5ac5cdb0671edbe4282	246d977f456ae840cbc69a450290c00b651c9cee	49f5d0eda9c8148a2847fef5c8c756d4421a97bc8083101608b4d2858efbef7b	WinPython-64bit-3.4.4.4Zerob2.exe	24 609 007 Bytes
  91773ce2887f03ffe1418a97aa0cf881	1314cb3a82c8b8a46dd95539ab25b8aa6cf3b176	191f42c4c0675b9b9bb100f345c9f47d4a37ab78e4de1bb28b176d81d27daece	WinPython-32bit-3.4.4.4Zerob2.exe	23 892 453 Bytes
  812eb6f5157e6b7de45ddf9dbc5dd838	14625be34565ba2027d4d0953b85efbf8ce3c2c5	f74a5925aa23ea9d7e43f9926141508263d6097b471023d57ffbde0140a18e9b	WinPython-64bit-3.4.4.4Qt5b2.exe	301 477 830 Bytes
  89a80b732946b7e88a13a59754612303	16e76432393ae81aed4541a1aca74f347d4b095c	5e02bb87e2618b83f3197b4c158c3cd712e7633d005a345b7103c41ca1b71174	WinPython-32bit-3.4.4.4Qt5b2.exe	256 786 397 Bytes
  a5c2b9c5799ef3bbbc072b3e8fa75f93	8e253d65473a59b9ef464c1eaa9962bf9e5f84e8	150c7cf10f2265e6bf1889725022539add56f1658e05694af29818db141804c3	WinPython-64bit-3.5.2.2Zerob2.exe	23 042 058 Bytes
  3f9ffa6749e1f1341e0e01c572933433	6c1245b52a28b589c6083719e2854a58fec3beb8	a74482b9b884716fbeeec4ad4ba57a8044cf16d008378b4363827cb9668dfcaf	WinPython-32bit-3.5.2.2Zerob2.exe	22 294 014 Bytes
  c2ae6e517e265e54b0286c74c3c10d97	3b3033b73eebd96cfcc56e5ef1b81803441c228b	76eebdecc87fab23d3fff7baca8a636a175fc66e64466b4b2605136271d98d90	WinPython-64bit-3.5.2.2Qt5b2.exe	271 693 801 Bytes
  88331b89f32ef5fb6037d3f3e3e37962	7d86afe0b54ca56286e2c8d62a3bcbba8e878e11	2cc6def1a6bbcae886402472bba72b96f5446b8f5adb748c4094e5b14dc37217	WinPython-32bit-3.5.2.2Qt5b2.exe	233 832 670 Bytes

History of changes for WinPython 3.5.2.2Qt5

The following changes were made to WinPython distribution since version 3.5.2.1Qt5.

Tools

Upgraded packages:

• Pandoc 1.13.1 → 1.17.2 (a universal document converter)

Python packages

New packages:

• altair 1.0.0 (High-level declarative visualization library for Python)
• pypandoc 1.2.0 (Thin wrapper for pandoc.)
• requests-file 1.4 (Requests is an Apache2 Licensed HTTP library, written in Python, for human beings.)
• vega 0.4.2 (An IPython/ Jupyter widget for Vega and Vega-Lite)
• webencodings 0.5 (Character encoding aliases for legacy web content)

Upgraded packages:

• alabaster 0.7.8 → 0.7.9 (A configurable sidebar-enabled Sphinx theme)
• astroid 1.4.7 → 1.4.8 (Rebuild a new abstract syntax tree from Python's ast (required for pylint))
• babel 2.3.2 → 2.3.4 (Internationalization utilities)
• beautifulsoup4 4.4.1 → 4.5.1 (Screen-scraping library)
• blosc 1.3.3 → 1.4.1 (Blosc data compressor)
• bokeh 0.11.1 → 0.12.1 (Statistical and novel interactive HTML plots for Python)
• certifi 2016.2.28 → 2016.8.8 (Python package for providing Mozilla's CA Bundle.)
• dask 0.10.1 → 0.10.2 (Minimal task scheduling abstraction)
• distributed 1.11.2 → 1.11.3 (Distributed computing)
• holoviews 1.6.0 → 1.6.1 (Composable, declarative data structures for building complex visualizations easily.)
• html5lib 0.9999999 → 0.999999999 (HTML parser based on the WHATWG HTML specification)
• ipykernel 4.3.1 → 4.4.1 (IPython Kernel for Jupyter)
• ipython 5.0.0 → 5.1.0 (Enhanced Python shell)
• jupyter-core 4.1.0 → 4.1.1 (Jupyter metapackage. Install all the Jupyter components in one go.)
• keras 1.0.6 → 1.0.7 (Theano-based Deep Learning library)
• lmfit 0.9.3 → 0.9.5 (Least-Squares Minimization with Bounds and Constraints)
• lxml 3.6.0 → 3.6.1 (Powerful and Pythonic XML processing library combining libxml2/libxslt with the ElementTree API.)
• mccabe 0.4.0 → 0.5.1 (McCabe checker, plugin for flake8)
• monotonic 1.1 → 1.2 (An implementation of time.monotonic() for Python 2 & < 3.3)
• nbformat 4.0.1 → 4.1.0 (The Jupyter Notebook format)
• nbsphinx 0.2.8 → 0.2.9 (Jupyter Notebook Tools for Sphinx)
• notebook 4.2.1 → 4.2.2 (# Jupyter Notebook)
• pandas-datareader 0.2.0 → 0.2.1 (Data readers extracted from the pandas codebase,should be compatible with recent pandas versions)
• partd 0.3.4 → 0.3.5 (Appendable key-value storage)
• pickleshare 0.7.2 → 0.7.4 (Tiny 'shelve'-like database with concurrency support)
• prompt-toolkit 1.0.3 → 1.0.6 (Library for building powerful interactive command lines in Python)
• pylint 1.6.1 → 1.6.4 (Logilab code analysis module: analyzes Python source code looking for bugs and signs of poor quality)
• pymongo 3.2.2 → 3.3.0 (Python driver for MongoDB http://www.mongodb.org)
• pyparsing 2.1.5 → 2.1.8 (A Python Parsing Module)
• pyqt5 5.6 → 5.7 (Python bindings for the Qt cross platform GUI toolkit)
• pytest 2.9.1 → 2.9.2 (pytest: simple powerful testing with Python)
• pytz 2016.4 → 2016.6.1 (World Timezone Definitions for Python)
• pyzmq 15.3.0 → 15.4.0 (Lightweight and super-fast messaging based on ZeroMQ library (required for IPython Qt console))
• qtpy 1.1.1 → 1.1.2 (Provides an abstraction layer on top of the various Qt bindings (PyQt5, PyQt4 and PySide) and additional custom QWidgets.)
• requests 2.10.0 → 2.11.0 (Requests is an Apache2 Licensed HTTP library, written in Python, for human beings.)
• requests-toolbelt 0.6.2 → 0.7.0 (Requests is an Apache2 Licensed HTTP library, written in Python, for human beings.)
• scipy 0.18.0rc2 → 0.18.0 (SciPy: Scientific Library for Python (advanced math, signal processing, optimization, statistics, ...))
• setuptools 24.0.2 → 25.1.6 (Download, build, install, upgrade, and uninstall Python packages - easily)
• sip 4.18 → 4.18.1 (Python extension module generator for C and C++ libraries)
• sqlparse 0.1.19 → 0.2.1 (Non-validating SQL parser)
• tornado 4.4 → 4.4.1 (Scalable, non-blocking web server and tools (required for IPython notebook))
• twine 1.7.4 → 1.8.1 (Collection of utilities for interacting with PyPI)
• winpython 1.6.20160625 → 1.6.20160816 (WinPython distribution tools, including WPPM (package manager))
• xarray 0.7.2 → 0.8.1 (N-D labeled arrays and datasets in Python)
• zarr 1.0.0 → 1.1.0 (A minimal implementation of chunked, compressed, N-dimensional arrays for Python.)

Is there a way to download the program in a ZIP fie?
After all it should be portable, so why the installer?

Thank You.

The installer is the historical way. by itself it does nothing but:

• splashing the licence and asking you where to unzip,
• allowing a "batch" auto-unzip.

So you can "install", "zip" the result, post the "zip" on internet, then "re-unzip" from internet "zip" file, with exact same effect.

rc (2016-08-24/25):

main features (since build2):

• new package:
• package updates: prompt_toolkit-1.0.7, holoviews-1.6.2, bqplot-0.8.2, spyder-3.0.0b5, ipyparallel-5.2.0, dask-0.11.0, xarray-0.8.2
• tools updates:

Areas of interest for testers:

• small change in widgets install (--py --sys-prefix), hopefully a good move ?

• statsmodels 0.8 was still week away

  MD5	SHA-1	SHA-256	Binary	Size
  64ace04d35e8d2eaa4c67469152e61fd	563b1f2e0e01d6ef60c4f0c6c1ce6448a94c69ff	0664b241e5dd8ef02d18f9ef1c07ef3860d605e164c50564dec5e078779fc1fd	WinPython-32bit-3.4.4.4Zerorc.exe	23 881 113 Bytes
  67c6ecb9e215a83551743426d7a08015	bcfb98567bfd0c68f4488c93f89c3e0a6e252f12	abecc24189db6ae5b10426c0eaba5c7f819a9d5362942808d0d960311a6480fe	WinPython-64bit-3.4.4.4Zerorc.exe	24 600 054 Bytes
  31cc8cbd1c94f69332b137cfa4f12c5f	bde744a8b442108bf200eb37799e3bacd6501f21	eebb24cd2a0996853818d8bf87c19f8411de94c5b2d335fc7c978cbe8287bb37	WinPython-64bit-3.5.2.2Zerorc.exe	23 032 756 Bytes
  6e738e456196b71b2b3c3d847d916f60	aecf18ba1fbcd955ddd40d962bd5b94ce6bc4418	1f6fa6457c5ab41f4eaaab5cd443c20a9fbfb0a3caef80ad89c4d92582d216b5	WinPython-32bit-3.5.2.2Zerorc.exe	22 280 272 Bytes
  0574aa934268fa6b4fcd450373ccf28e	7077e5185917d320084f8d63a13f524c2900a877	280a877d452e05f71bd6a72685b99b99afb257863b2b5a8e174007eb58288225	WinPython-64bit-2.7.12.3Zerorc.exe	22 221 744 Bytes
  fa4d385602f9e200b27970e5d863e656	1e0b915bb312a67df3bab39ba7078fc571f234be	032e367187eb85945c01e1711950eed65f077bfc254c0dea5279cba580bd45e7	WinPython-32bit-2.7.12.3Zerorc.exe	21 498 414 Bytes
  f682dbc3eeb99843c8b848ade3b220cb	c39e10adc8f2f6c42c06ef538edb4779bcd85f34	7b38419233b775354ae09d53f253b7fa5407cf99ad15c20ea3646c1e51759c57	WinPython-64bit-3.4.4.4Qt5rc.exe	301 448 368 Bytes
  d372671a80bad0bd2e652e569c45e4ae	ead43341e0ae01860a90e6847333f5f6fe15a14c	b65c453fc09a5b5f585fd25aef23bb668ace6533e9f409e3d94125b4791c7d7c	WinPython-32bit-3.4.4.4Qt5rc.exe	256 752 166 Bytes
  5a96da96e8a38700076b3200c6e60dfe	280cce807857e84b550246b155592456899364da	ae9dd71ef002b3cf85dc9c05cfebd78acc714cfe1a5c1e69cbdf78f4558559ba	WinPython-64bit-3.5.2.2Qt5rc.exe	271 660 710 Bytes
  e4e10b6ae96bf11874cad56bdd302990	da6c45b49bf197fdc871188c7b606676682f0b2d	f24b7f1861d945abcbc9984df69fb94a28fdfcd78ee3dfc612c029a29e24e908	WinPython-32bit-3.5.2.2Qt5rc.exe	233 796 411 Bytes
  903cf2e4b3b19c04e06365d7a1886437	12672218ee8ad67f16708231db0541a6db509a8a	a230587bc770267c62724e15da2e655ccc8b260c9ffb9d3d7a36a6dc12243df5	WinPython-64bit-3.5.2.2rc.exe	258 697 814 Bytes
  51f40fb744d70c5a59fe3997043f55b3	8d11dbf4e0808fd2bfcd284327690edcf9a15dce	abb4d323dbd72781d7fa2a650385331182d5e64e8748d56b33a2ab16254a4752	WinPython-32bit-3.5.2.2rc.exe	224 074 210 Bytes

History of changes for WinPython 3.4.4.4Qt5

The following changes were made to WinPython distribution since version 3.4.4.3Qt5.

Tools

Upgraded packages:

• Pandoc 1.13.1 → 1.17.2 (a universal document converter)

Python packages

New packages:

• altair 1.0.0 (High-level declarative visualization library for Python)
• pypandoc 1.2.0 (Thin wrapper for pandoc.)
• requests-file 1.4 (Requests is an Apache2 Licensed HTTP library, written in Python, for human beings.)
• vega 0.4.2 (An IPython/ Jupyter widget for Vega and Vega-Lite)
• webencodings 0.5 (Character encoding aliases for legacy web content)

Upgraded packages:

• alabaster 0.7.8 → 0.7.9 (A configurable sidebar-enabled Sphinx theme)
• astroid 1.4.7 → 1.4.8 (Rebuild a new abstract syntax tree from Python's ast (required for pylint))
• babel 2.3.2 → 2.3.4 (Internationalization utilities)
• beautifulsoup4 4.4.1 → 4.5.1 (Screen-scraping library)
• blosc 1.3.3 → 1.4.1 (Blosc data compressor)
• bokeh 0.11.1 → 0.12.1 (Statistical and novel interactive HTML plots for Python)
• bqplot 0.7.1 → 0.8.2 (Interactive plotting for the Jupyter notebook, using d3.js and ipywidgets.)
• certifi 2016.2.28 → 2016.8.8 (Python package for providing Mozilla's CA Bundle.)
• dask 0.10.1 → 0.11.0 (Minimal task scheduling abstraction)
• distributed 1.11.2 → 1.12.0 (Distributed computing)
• holoviews 1.6.0 → 1.6.2 (Composable, declarative data structures for building complex visualizations easily.)
• html5lib 0.9999999 → 0.999999999 (HTML parser based on the WHATWG HTML specification)
• ipykernel 4.3.1 → 4.4.1 (IPython Kernel for Jupyter)
• ipyparallel 5.1.1 → 5.2.0 (Interactive Parallel Computing with IPython)
• ipython 5.0.0 → 5.1.0 (Enhanced Python shell)
• jupyter-core 4.1.0 → 4.1.1 (Jupyter metapackage. Install all the Jupyter components in one go.)
• keras 1.0.6 → 1.0.7 (Theano-based Deep Learning library)
• lmfit 0.9.3 → 0.9.5 (Least-Squares Minimization with Bounds and Constraints)
• lxml 3.6.0 → 3.6.4 (Powerful and Pythonic XML processing library combining libxml2/libxslt with the ElementTree API.)
• mccabe 0.4.0 → 0.5.1 (McCabe checker, plugin for flake8)
• monotonic 1.1 → 1.2 (An implementation of time.monotonic() for Python 2 & < 3.3)
• nbformat 4.0.1 → 4.1.0 (The Jupyter Notebook format)
• nbsphinx 0.2.8 → 0.2.9 (Jupyter Notebook Tools for Sphinx)
• notebook 4.2.1 → 4.2.2 (# Jupyter Notebook)
• pandas-datareader 0.2.0 → 0.2.1 (Data readers extracted from the pandas codebase,should be compatible with recent pandas versions)
• partd 0.3.4 → 0.3.6 (Appendable key-value storage)
• pickleshare 0.7.2 → 0.7.4 (Tiny 'shelve'-like database with concurrency support)
• pillow 3.3.0 → 3.3.1 (Python Imaging Library (fork))
• prompt-toolkit 1.0.3 → 1.0.7 (Library for building powerful interactive command lines in Python)
• pylint 1.6.1 → 1.6.4 (Logilab code analysis module: analyzes Python source code looking for bugs and signs of poor quality)
• pymongo 3.2.2 → 3.3.0 (Python driver for MongoDB http://www.mongodb.org)
• pyparsing 2.1.5 → 2.1.8 (A Python Parsing Module)
• pytest 2.9.1 → 2.9.2 (pytest: simple powerful testing with Python)
• pytz 2016.4 → 2016.6.1 (World Timezone Definitions for Python)
• pyzmq 15.3.0 → 15.4.0 (Lightweight and super-fast messaging based on ZeroMQ library (required for IPython Qt console))
• qtpy 1.1.1 → 1.1.2 (Provides an abstraction layer on top of the various Qt bindings (PyQt5, PyQt4 and PySide) and additional custom QWidgets.)
• requests 2.10.0 → 2.11.1 (Requests is an Apache2 Licensed HTTP library, written in Python, for human beings.)
• requests-toolbelt 0.6.2 → 0.7.0 (Requests is an Apache2 Licensed HTTP library, written in Python, for human beings.)
• scipy 0.18.0rc2 → 0.18.0 (SciPy: Scientific Library for Python (advanced math, signal processing, optimization, statistics, ...))
• setuptools 24.0.2 → 25.1.6 (Download, build, install, upgrade, and uninstall Python packages - easily)
• sphinx 1.4.5 → 1.4.6 (Tool for generating documentation which uses reStructuredText as its markup language)
• spyder 3.0.0.dev0 → 3.0.0b5 (Scientific PYthon Development EnviRonment: designed for interactive computing and data visualisation with a simple and intuitive user interface)
• sqlparse 0.1.19 → 0.2.1 (Non-validating SQL parser)
• tornado 4.4 → 4.4.1 (Scalable, non-blocking web server and tools (required for IPython notebook))
• twine 1.7.4 → 1.8.1 (Collection of utilities for interacting with PyPI)
• winpython 1.6.20160625 → 1.6.20160816 (WinPython distribution tools, including WPPM (package manager))
• xarray 0.7.2 → 0.8.2 (N-D labeled arrays and datasets in Python)
• zarr 1.0.0 → 1.1.0 (A minimal implementation of chunked, compressed, N-dimensional arrays for Python.)

Any chance for ZIP only files?

Thank You.

zip only of which version exactly ?

Instead of EXE installer.
I'd be happy to have a ZIP which all needed is to extract it.

I know the EXE is only that, yet still.

which version of WinPython exactly ? the WinPython3.4-64 bit ?

I always use the latest one, but I guess my request would be for any release.
Just in addition to the installer base EXE file add an option to download ZIP file and extract it manually.

Thank You.

@RoyiAvital if you want to extract the exe without running it, 7-Zip can do it.