SciPy is a language extension, which uses NumPy to complete advanced mathematics, signal processing, optimization, statistics and many other scientific tasks.
If you have learned these two tools, the official has very detailed documents and tutorials to help you get started: I am the portal.
There is also a book NumPy and SciPy, which is very thin and only has 67 pages: I am the portal.
How to install NumPy and SciPy
The main reason for writing this article is that SciPy official website seems to have forced the installation of huge Python-based software distribution, but installing these softwares usually means deleting the previous Python, which is a bit too painful. After searching for a long time, I found SciPy's clean expansion package in the corner of official website. This article should be convenient for future generations.
There are two ways to install NumPy and SciPy:
The first method is to install a complete python-based software distribution, which generally inherits many Python expansion packages and some other utilities, such as IPython, Spyder and so on. The advantage of this method is simple, once and for all, fool-like installation, just like installing ordinary programs, installing many expansion packages at a time; The disadvantage is that it is slightly larger, and it is generally necessary to delete the previous pure version of Python, so if the configuration of something (such as Python's IDE) is based on the previous pure version of Python, it needs to be reconfigured. Of course, these are also very simple. The newly installed Python is usually in a folder of these softwares.
SciPy official website has the installation address of these softwares, and the link is here: I am the portal.
Among them, Python(x, y) seems to be famous, and some people in China have written a book on scientific computing based on this software, called Python Scientific Computing.
Play whichever you like ~
The second method is to install Python expansion package, because SciPy is based on NumPy, so you need to install NumPy first, and then SciPy. This method is a little more troublesome, but 10 minutes (not counting the download time) will do. The advantage is that the installed things are small in size, and there is no need to delete the previous Python.
NumPy download link is here: I am the portal.
Windows system, just download the corresponding exe file directly, open it and install it directly, which is simple and easy. The Linux system will download the tar package and then perform Python setup. Pybuild from CD to the corresponding directory. Python settings. Pyinstall should be enough (never tried, but it is usually the case).
SciPy download link is here: source package, executable file.
Download the executable file directly under Windows and install it directly. Under Linux, you still need to download the source package and install it in the above way (I haven't tried it, it should be like this)
Test whether the installation was successful:
NumPy, execute importnumpy when you are free, and generally install it if there are no errors.
SciPy, execute importscipy when you are free, and generally install it if there are no errors.