Installation of dtaidistance in python provides error - python

I am trying to install the module dtaidistance in Python to calculate the DTW distance matrix for a set of time series. But whenever I try to install the package using "pip install dtaidistance" a large error pops up and I don't know how to interpret it and what to do in order to make it function.
See the final few lines and the corresponding error message after running the install in the anaconda prompt in the screenshot below:
Error code screenshot:

The error says the cl.exe binary is not found. This probably means that you have not installed the Microsoft C compiler and linker (MSVC or Microsoft Visual Studio for Windows).
By the way, the current version of DTAIDistance should continue installing the Python implementation but the fast C-based versions will not be available.

Related

Scipy: C Callback DLL failing to Load

I have been attempting to install scipy such that I can run it from IDLE. I have installed :
the proper numpy wheel via pip
the proper scipy wheel via pip
the visual C++ redistributable
I am using python 3.6.xx 32 bit and windows 10 pro. I have uninstalled and reinstalled everything multiple times. I am unsure how to move forward or where to look, it seems to me that the low level callable library has not installed properly given I am getting the error "Import Error: DLL load failed: the specified module could not be found"

Failed to import Sklearn

I am trying to build a recommender system with Python. However, I can't proceed since I get an import error of Sklearn everytime.
I get this error:
"ImportError: DLL load failed: Can't find module."
How do i fix this? (The last sentence in the error is in Dutch btw)
A screenshot of my error is below.
By the way: I get the same import error for Seaborn
Windows 10 + Visual Studio Code - "CMD + SHIFT + P" from within visual studio code and search for "Select Interpreter" clicking on this will show a list of available python interpreters and versions installed on your system that are detected by visual studio code and your currently selected environment. (Visual Studio Code usually incorrectly defaults to 32 bit interpreter)
Ensure you have the correct python environment (where you have the required libraries & "packages" installed)
This should fix your import issue if you have the libraries installed. Use the appropriate cmd line tool to install packages. If you are using Anaconda - use Conda Install or use pip etc...
Scikit-learn uses numpy and scipy. Those modules are using C and Fortran code that needs to be compiled. Nonetheless, Windows cannot compile them right away.
I would recommend you to download Anaconda and install scikit-learn like this. (You may need to add conda to path)
Another method is to download precompiled binary from this page. Then run:
pip uninstall scikit-learn
pip install <numpy-wheel>
pip install <scipy-wheel>
pip install <scikit-wheel>

Could not import cufflinks

Problem
I am trying to install both plotly and cufflinks. However I had a problem.
The installation of both plotly and cufflinks were successful. Although, I can't import cufflinks.
Below is a picture of the problem. It seems to be a dependency error:
I tried manually downloading and installing "talib" but I keep getting failures. (Shown below).
talib\common.c(240): fatal error C1083: Cannot open include file: 'ta_libc.h': No such file or directory
error: command 'C:\\Program Files (x86)\\Microsoft Visual Studio 14.0\\VC\\BIN\\x86_amd64\\cl.exe' failed with exit status 2
Any Ideas? I tried re-installing both modules and Anaconda. Nothing So far.
Other infos:
Cufflinks version: 0.11.0
Plotly version: 2.0.10
Anaconda version: 3-4.4.0 (But I don't think it have anything to do with it)
Python version: 3.6.1
try installing this version of cufflinks, it eliminated the error for me.
pip install cufflinks==0.8.2
From this link: github.com/mrjbq7/ta-lib#troubleshooting
Troubleshooting
Sometimes installation will produce build errors like this:
func.c:256:28: fatal error: ta-lib/ta_libc.h: No such file or directory
compilation terminated.
This typically means that it can't find the underlying TA-Lib library, a dependency which needs to be installed. On Windows, this could be caused by installing the 32-bit binary distribution of the underlying TA-Lib library, but trying to use it with 64-bit Python.
Windows
Download ta-lib-0.4.0-msvc.zip and unzip to C:\ta-lib
This is a 32-bit release. If you want to use 64-bit Python, you will need to build a 64-bit version of the library.
My Fix
So, for windows, we need a 64-bit version of the library? Luckly I found a lot of modules built for 32 and 64 bits python:
http://www.lfd.uci.edu/~gohlke/pythonlibs/
search for "ta-lib" and click on the module you need (In my case cp36 64 bits).
Then, open the command prompt.
Change to where you downloaded the file: cd path/to/file
Type: pip install NameOfFile (in my case pip install TA_Lib‑0.4.10‑cp36‑cp36m‑win_amd64.whl )
Now the 64-bits Ta-Lib module should be installed in your machine. I tested the previous line of codes and it worked!
Thanks for the help :)
I have now removed all dependencies on talib. All studies are pure python based now and you should not face any of this errors.

Using NumPy in Visual Studio

I downloaded http://pytools.codeplex.com/ (Python Tools for Visual Studio) so that I could write Python in Visual Studio.
The problem is when I try to use the most basic package "numpy" like so:
import numpy
It says "No module named 'numpy'."
How can I use NumPy and SciPy in Visual Studio?
Note: I am using Canopy Express on another machine which works perfectly; however, I don't want to install it on this machine since I already have Visual Studio installed.
I added the Python 3.3 environment to my Python "solution" by right-clicking Python environments and clicking add an environment. But, after right-clicking my environment and clicking install Python package, I typed in "numpy" and got this error when trying to install it:
...
creating build
creating build\src.win32-3.3
creating build\src.win32-3.3\numpy
creating build\src.win32-3.3\numpy\distutils
building library "npymath" sources
No module named 'numpy.distutils.msvccompiler' in numpy.distutils; trying from distutils
error: Unable to find vcvarsall.bat
----------------------------------------
Cleaning up...
Command python setup.py egg_info failed with error code 1 in c:\users\dom\appdata\local\temp\pip_build_Dom\numpy
Storing complete log in C:\Users\Dom\pip\pip.log
'numpy' failed to install. Exit code: 1
How can I install NumPy?
In Visual Studio 2013 Community Edition with a Python 2.7 environment, the following worked for me:
Open the Python Environments menu in Solution Explorer, and click on "Install Python Package..."
Then write "numpy":
Two solutions worked for me:
Going directly to my Python folder and installing the packages using pip.
Creating a virtual environment in my project and installing packages from within Visual Studio.
Consider using a more feature-rich Windows distribution of Python, like Anaconda. It comes with a boatload of packages included into the installer. This is quite handy for those who are new to Python, especially when they are using Windows.
Edit:
As of 2020, I would suggest using the python.org distribution of Python. Compiler issues on ms-windows have finally been fixed, so a lot of projects now provide binary packages (wheels) for ms-windows. Among others numpy, pandas, matplotlib, pillow, Cython and PyNaCl.
Note that wheel names ending in:
win_amd64.whl are for 64-bit versions of ms-windows, and
win32.whl are for 32-bit.
I actually ran into the same set of problems when trying to use Visual Studio for Python.
I managed to get things to work at the end - please see Using Visual Studio 2013 for Python (3.4) with NumPy and SciPy on Windows for details. Essentially you can either use the installer (if one is available) or use pip to do the installation.
Make sure C++ Common Tools are installed in Visual Studio:
See under Extra in the Visual Studio (Extra-Python-Interactive Window).
It has the ability to switch the interactive console through a different platform/Python version (32-bit, 64-bit, Debug, and Anaconda (which has the NumPy package)).
From PTVS Installation:
2. Installing packages individually
Here are some recommended packages. Generally you want to install packages using pip or through the interface in Visual Studio (which uses pip), as this will ensure that you get the latest version that will work with your version of Python, as well as any dependencies.
Some packages have complex dependencies and need to be downloaded manually, either from the project's website or from Christoph Gohlke's package collection.
I installed numpy and scipy from these links, and it works perfectly with Visual Studio.
numpy http://sourceforge.net/projects/numpy/files/NumPy/
scipy http://sourceforge.net/projects/scipy/files/scipy/
Matplotlib http://matplotlib.org/downloads.html
Pandas http://pandas.pydata.org/getpandas.html
Well, if anyone still has problem than what worked for me is:
C:/Users/DELL/AppData/Local/Programs/Python/Python39/python.exe -m pip install numpy
where the first part is simply the path of python.

Why does installing numpy using pip fail, while building directly does not?

I'm trying to install numpy under Python 3.3.0 running Mac OS 10.7.5 (Lion) and using the compilers that come with Xcode 4.5.1. I installed Python 3 and gfortran using homebrew without any hiccups, but pip3 install numpy fails. Looking at other suggestions online, I tried setting Clang as the C compiler using:
export CC=clang
export CXX=clang
export FFLAGS=-ff2c
but received the same error. The final line reports ValueError: underlying buffer has been detached, and I posted the full output.
So, instead I tried building directly from source using the instructions from SciPy.org. This appears to succeed; the build and install processes run to completion, although they are not shy about reporting warnings. The full log is again available as a Gist. I can then run numpy.test('full') from Python 3, and although there are some warnings about unclosed files that I don't understand, it reports that all tests pass.
Does anyone know why pip fails to build NumPy when I can apparently do it manually without a problem?
It looks like numpy now provides wheels on pypi, which should install correctly without needing to compile anything locally on your machine when using pip.

Categories