{"id":6927,"date":"2020-11-09T17:49:32","date_gmt":"2020-11-09T17:49:32","guid":{"rendered":"https:\/\/dataaspirant.com\/?p=6927"},"modified":"2020-12-15T04:21:09","modified_gmt":"2020-12-15T04:21:09","slug":"anaconda-python-virtualenv","status":"publish","type":"post","link":"https:\/\/dataaspirant.com\/anaconda-python-virtualenv\/","title":{"rendered":"What\u2019s Better? Anaconda or Python Virtualenv"},"content":{"rendered":"
<\/div>

\"Anaconda

What\u2019s Better? Anaconda or Python Virtualenv<\/h1>\n

Creating deep learning or machine learning models in local systems is like a cakewalk. Things get complicated when we try to replicate the same project setup in the cloud. The two popular options we as a data science community have for managing project environments are anaconda environment and python virtualenv.<\/a><\/p>\n

Which one did you use?<\/strong> <\/p><\/blockquote>\n

Else let me put the straight question which project environment is best for deploying data science projects in the cloud?<\/p>\n

Anaconda?<\/strong>  or python virtualenv<\/strong> environment? <\/p><\/blockquote>\n

Confused to answer this question, don\u2019t blame your mind, just relax and read this article. You will get clear answers with many reasons or features to keep in mind while selecting the environment for your next data science projects<\/a>.<\/p>\n

 To get started, let quickly have a look at the table of contents of this article.<\/p>\n

Why do we need the environment?<\/h2>\n

\"Why

Why we need environment<\/p>\n

Let\u2019s start the discussion with why we need to maintain environments?<\/strong><\/p>\n

In the world of data science, at the same time, we usually work on  various projects.<\/p>\n

For example, we are building models to find the fraudulent activities of credit cards<\/a>, and at the same time, we are testing the performance of the email spam classifier model <\/a>we have already built.<\/p>\n

The data science packages and project setup is completely different for these two projects. The ideal approach is having two various machines for these two projects so we won\u2019t find any issues. <\/p>\n

But this approach is not a feasible one, and it\u2019s not a cost-effective<\/strong> way too.<\/p>\n

The other thing we can do is, using the same machine for both the projects, cost optimization wise this approach is perfect. <\/p>\n

But we will face issues with this approach as well.<\/p>\n

Let\u2019s say the packages we used got a few updations. Now we need to update the package. If we update, few functions in the last version code will throw errors.<\/p>\n

For example<\/h4>\n

Let\u2019s say we are using pandas version x<\/strong>. In this version, to get the frequency of the categorical values, we use the syntax get_frequency<\/strong>. Whereas in the new version, it\u2019s changed to frequency_values<\/strong>.<\/p>\n

Suppose we update the code and run our script without doing these modifications. Our main script will fall<\/strong>. So It\u2019s always recommended to use the same library versions in our projects.<\/p>\n

That\u2019s the reason you will find the requirements.txt<\/strong> file in Github projects<\/a>. Don\u2019t worry about the requirements.txt file. For now, we will discuss this more in our upcoming sections of this article.<\/p>\n

What is the requirements.txt file?<\/h3>\n

For now, just remembers using the requirements.txt file, <\/strong>we will keep track of each package or library we used in the project with a specific version <\/strong>details.<\/p>\n

Below is the sample requirement.txt file with contents.<\/p>\n

\"Sample

Sample requirements file<\/p>\n

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.<\/p>\n

To create the project setup, all we need is to install these packages in whatever system we intend. It could be the local laptop or desktop; else, it could be the cloud setup like google app engines or AWS EC2 engines.<\/p>\n

Now let\u2019s go back to our actual question,<\/p>\n

why do we need the environments<\/strong>?<\/p><\/blockquote>\n

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.<\/p>\n

In this way, in the same system,  we can create multiple environments and work on numerous projects.<\/p>\n

The technical way of saying this managing the python packages, in more general, it\u2019s called packages manager<\/strong>.<\/p>\n

What is Pip?<\/h2>\n

To manage various packages, we need a system that will take care of keeping track of each package and its versions. That\u2019s where we have PIP, a python package manager. Using the pip, we will get the feasibility to create the various environments. <\/p>\n

If you remember in the gender wise face detection project<\/a>, we have used the pip to install various computer vision<\/strong> python libraries.<\/p>\n

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.<\/p>\n

Pip is the ideal place for mange packages that have not come with python installations. It helps in installing any other packages, and pip community members are super active and address the comments in a much faster way.<\/p>\n

If you know any other programming language, the concept of the package manager is similar.<\/p>\n

\n