Can't install gensim with python 3.4 - amazon-web-services

I cannot install gensim==3.5.0 in my elastic beanstalk environment (python 3.4). I get an error that gensim needs python >= 3.5 to run.
This was not a problem until a mid-day deployment today, that made only project code changes, nothing related to elastic beanstalk, requirements or settings.
At the same time, I'm succesfully running the same version in another identical environment. That means the same pip, same python version, same required dependencies.
I tried lowering the gensim requirement to gensim==0.13.4 which officially supports python 3.4, but I get the same error.
EDIT: I managed to make things work by installing gensim==0.10.0 and then redeploying with gensim=3.5.0. I still don't know the cause of the issue and the solution is not really a solution, so I'm still interested in insights.

Note that currently (July 2019), Python 3.4 itself will no longer be supported with fix releases. Per https://www.python.org/downloads/release/python-3410/:
Release Date: March 18, 2019
Python 3.4 has reached end-of-life. Python 3.4.10 is the final release
of 3.4.
Python 3.4.10 was released on March 18th, 2019.
Python 3.4.10 is the final release in the Python 3.4 series. As of
this release, the 3.4 branch has been retired, no further changes to
3.4 will be accepted, and no new releases will be made. This is standard Python policy; Python releases get five years of support and
are then retired.
If you're still using Python 3.4, you should consider upgrading to the
current version. Newer versions of Python have many new features,
performance improvements, and bug fixes, which should all serve to
enhance your Python programming experience.
That said, if you're truly getting an error about gensim's require Python level, truly rolling back to whatever version of gensim you were using successfully, previously, with Python 3.4 should work. (That might not need to go as far back as gensim-0.13.4.1, which is almost 2.5 years old, but if you're sure that's the version that was working for you, you could use that version.)
You should edit your question to show exactly what installation commands you've run, and exactly which message is received in response to which step, to more clearly indicate what's been tried, where the error is arising, and why a simple attempt to install-an-older-version might be getting a similar error message.

Related

How to fix new chrome driver's exception - WebDriverException: Message: invalid argument: value must be a non-negative integer

I am running a test case that I have been running for months.
Yesterday I updated Chrome to the latest release (75.0.3770.100), also I downloaded the compatible version of chrome driver (75.0.3770.90). After these updates, I tried to run the same test case and I get the following exception:
"WebDriverException: Message: invalid argument: value must be a non-negative integer"
How can I manage/fix that?
Is there any workaround?
I am using python 2.7, selenium version 3.5.0, and robot framework version 3.0.2.
I have managed to fix the issue by upgrading selenium version to 3.9.0
I had the same issue when I have updated my recent webdrivers along with the browsers but the IDE was still on old version and some plugins were outdated.
The issue was resolved once I have updated the IDE (in my case I use Pycharm).
One more thing try to make sure that your project Interpreter set correctly to Python environment something like below:
Python 2.7 (venv) ~/venv/bin/python

Choosing the appropriate version of Python runtime to use along with TensorFlow

How can I choose the appropriate version of Python runtime to use along with TensorFlow 1.0 ?
I would invite you to check the TensorFlow installation page which gives all the detail for your specific system. At the moment, TensorFlow supports Python 2.7, 3.4, 3.5, 3.6, although it seems 3.6 does not yet have GPU support. If you're starting out with Python, you should probably get 3.5 (if you need GPU) or 3.6 (if you'll only use CPU). Python 2.7 was released a long time ago.
If you don't already have Python, you can get it from the Python website. However, I recommend you get Python through Anaconda so that you get all necessary packages to run things like TensorFlow.

Python distutils refuses to recognise installed C++ compilers, `Unable to find vcvarsall.bat` under Python 3.5

I'm tearing my hair out trying to compile c though Python's distutils. Specifically I'm trying to get started with Cython following this guide: http://docs.cython.org/src/quickstart/build.html, using a setup.py.
I'm running Python 3.5.0 under Anaconda 2.4.0 (64-bit), Windows 7.
I get the Unable to find vcvarsall.bat error referred to here: Cannot find vcvarsall.bat when running a Python script and in various other questions.
None of the proposed solutions work for me. So far I have tried:
installing Visual Studio - I have all versions from 2008 to 2013, ie 9.0, 10.0, 11.0, and 12.0.
pointing environment variables (VS100COMNTOOLS etc) related to the above directly to the correct locations
installing Microsoft Visual C++ Compiler for Python (as above link). This seemed to work but I get a compilation error that seems to be an incompatibility; I think because the compiler is intended for Python 2.7.
Ravi Kumar's suggestion at the above link to update pip's setuptools. Pip gave me an error saying it couldn't update.
Installing MS Windows SDK for Windows 7 and .NET Framework 4 as https://github.com/cython/cython/wiki/CythonExtensionsOnWindows. The final step failed with an error message that the x64 compilers aren't currently installed.
The only thing I can think of is installing MinGW which I've seen suggested in some places, but that CythonExtensionsOnWindows link specifically recommended against it.
Any other suggestions would be very welcome! In particular, since I clearly have all the versions of VS I could possibly need, is there a way to force Python to recognise them, other than setting environment variables VS100COMNTOOLS or similar?
Interestingly conda does seem to recognise the installed compilers, as a separate pip install attempt that failed with this same error, worked with conda.
As it turns out, Python 3.5 needs Visual Studio 2015. Advice elsewhere on fixing this for other Python versions is outdated.
Get Visual Studio 2015 Community here: https://www.visualstudio.com/en-us/downloads/visual-studio-2015-downloads-vs.aspx, this worked for me with no further steps needed.
Many thanks to Ionel, apparently the only place on the web to find this information! http://blog.ionelmc.ro/2014/12/21/compiling-python-extensions-on-windows/

How to upgrade Portable Python 2.X after installation / deployment

I had recently installed Portable Python 2.7.6.1 from Portable Python, and I understand that the latest Python version of the 2.X family is 2.7.9.
I am wondering if it is possible for me to upgrade my installation of Portable Python, after I had deployed it on my USB? If it is possible, please show me the way.
If this question had already been answered (although I had searched and researched quite a bit), please point me the question. I am more than happy to close this question and offer my apologies.
I may be wrong - but just to provide anyone who stumbled upon this - there is no way I can upgrade the python interpreter that comes with the portable python package.
Eventually I switched to WinPython instead. It seems like there is no easy way to upgrade itself either, but for a portable environment, it can be replaced by downloading the newer version from project homepage.
Edit:
I may be too early in my response, I think Anaconda distribution has an ability to update itself by:
conda update all

Python 2.7 can't find vcvarsall.bat, MS VS C++ 2008 is installed

So I've been working with python 2.7 no problem for a while now. I've been using pip for a couple of months without issue. I recently installed virtualenv and now none of my pythons or pip can find vcvarsall.bat, even though this wasn't a problem before. I thought virtualenv seemed like a good idea, but not if it breaks everything around it. I tried to run repair on VS C++ but it didn't find any problems. Has anyone run into something like this before?
Seems that the problem is that Python 2.7 64 bit has a compilation problem. It would seem that I need to download a special package from MS to get access to 64-bit c/c++ compiler that's compatible with Python 2.7. Not a problem, except that's its 3 gigabytes. So, I just did it on my Linux VM and Windows Python 2.7 is 32-bit for me now. Not the best solution, but we're supposedly going to upgrade to Python 3 one of these years. Probably about when Python 4 comes out and the Python 3 compiler is obsolete too. C'est le vie!