Why do we need virtual environment ?
The purpose of a Venv is to have a space where we can install packages which are specific to a certain project. For instance, there are lots of projects which are being deployed with Tensorflow 1.x. Version and the current updated version is Tensorflow 2.x. and Let’s consider a scenario with RASA — X framework using a single global environment. Where when we install Rasa -x framework by default it will use the Tensorflow 2.x. Version and other projects which depend on Tensorflow 1.x. version will definitely Importerror as we just updated the Tensorflow library and didn’t find the equivalent code.
So now we understand why we need a separate environment for each project, i.e., each project will be independent of the Module version from each other.
So now let’s look into venv modules and create a Virtual Environment.
In order to do this we need to use Python version 3.3+ and here am using Python 3.7.3 and anything that 3.3+ will do well.
First move to your project directory where we will do all the setups. You really don’t have to install anything for creating a virtual environment, as this comes with the standard library by default. First, you can have a look into my library packages and their version, you can use the following command
either use pip list / pip freeze
if your using ubuntu then do pip3 list / pip3 freeze
Well, here you make sure to make a note of the Tensorflow version that I have available in my system, it’s obviously Tensorflow 1.x. and not Tensorflow 2.x.
Now let’s create a virtual environment
python -m “Module Name” “Project / Virtual Environment Name”
python -m venv tf2_version
Here python -m flag which stands for module, i.e. some modules have main entry points and in order to run this has to be used.
This will create your tf2_version virtual environment and you can verify by moving into that directory in your local machine or you can verify by listing it out. In windows use the command “dir” and in Ubuntu use the command “ls”
Now we need to activate the virtual environment and use the following command for it,
“project name”\Scripts\activate.bat
tf2_version\Scripts\activate.bat
Once you activate it, you can confirm it as shown below
You can also verify it with another way by typing “where python”
You can see that the at first the current environment directory is listed.
Now you do a “pip list” and check for the list of available Modules available by default.
Now let’s say our project requires Tensorflow Version 2.x. and so let’s install it.
pip install tensorflow
Here in the comparison image you can verify that from the normal command prompt you can see the Tensorflow Version is 1.x. and in the virtual environment its 2.x.
If in any case you want to get a copy of the list of Modules that’s in this virtual environment you may download as shown below.
pip freeze > requirements.txt
Now for deactivating our virtual environment, we need to use the command deactivate as shown below
After this if you wish to delete the virtual environment, it’s just like deleting the folder or from command line you give the following command
rmdir “name of the virtual environment” /s
rmdir tf2_version /s
Here the /s make sure to delete the entire tree and the subtree in it.
Try it on your own ..!!!
I will let you guys try out this last part on your own.
Now there also another scenario where you like to create a virtual environment with the modules that we have created in the our local machine then you need to use
python -m “Module Name” “Project / Virtual Environment Name” — system-site-packages
python -m venv tf2_version — system-site-packages
If you wish to stay connected,
Or
you can just google “ narenltk / narendiran krishnan ” or just drop a mail to → narenltk@gmail.com → Happy to help..!!!
