We need to install it with the help of pip. To create the virtualenv (environment), we can leverage the python virtualenv package. Then we will deep drive to understand which one is better to use in data science projects. Now let’s see how we can create virtual environments using python and conda. We learned how we can leverage the using various environments, and we also learned how pip will help us in achieving this. If you know any other programming language, the concept of the package manager is similar. It helps in installing any other packages, and pip community members are super active and address the comments in a much faster way. Pip is the ideal place for mange packages that have not come with python installations. The pip will take care of holding the versions in the cloud when we see a package and its version in the requirement.txt, then pip will provide you the specific version and help you in installing the packages. If you remember in the gender wise face detection project, we have used the pip to install various computer vision python libraries. Using the pip, we will get the feasibility to create the various environments. That’s where we have PIP, a python package manager. To manage various packages, we need a system that will take care of keeping track of each package and its versions. The technical way of saying this managing the python packages, in more general, it’s called packages manager. In this way, in the same system, we can create multiple environments and work on numerous projects. If we maintain the requirement.txt file, then for each project, we can create a different environment, and the project will run in the environment specific to the project. Now let’s go back to our actual question, why do we need the environments? It could be the local laptop or desktop else, it could be the cloud setup like google app engines or AWS EC2 engines. To create the project setup, all we need is to install these packages in whatever system we intend. In the requirements.txt file, we will store what package we have used in our project, and we also keep track of the versions we are using. What is the requirements.txt file?įor now, just remembers using the requirements.txt file, we will keep track of each package or library we used in the project with a specific version details.īelow is the sample requirement.txt file with contents. For now, we will discuss this more in our upcoming sections of this article. Don’t worry about the requirements.txt file. That’s the reason you will find the requirements.txt file in Github projects. So It’s always recommended to use the same library versions in our projects.
#MINICONDA VS ANACONDA PYTHON UPDATE#
Suppose we update the code and run our script without doing these modifications. Whereas in the new version, it’s changed to frequency_values. In this version, to get the frequency of the categorical values, we use the syntax get_frequency. If we update, few functions in the last version code will throw errors. Let’s say the packages we used got a few updations. The other thing we can do is, using the same machine for both the projects, cost optimization wise this approach is perfect.īut we will face issues with this approach as well. The ideal approach is having two various machines for these two projects so we won’t find any issues.īut this approach is not a feasible one, and it’s not a cost-effective way too. The data science packages and project setup is completely different for these two projects.
In the world of data science, at the same time, we usually work on various projects.įor example, we are building models to find the fraudulent activities of credit cards, and at the same time, we are testing the performance of the email spam classifier model we have already built. Let’s start the discussion with why we need to maintain environments?
0 kommentar(er)
