Skip to content

Latest commit

 

History

History
108 lines (78 loc) · 4.84 KB

Lab.md

File metadata and controls

108 lines (78 loc) · 4.84 KB
Raw

Getting started with Python

We're going to be using Python in a few different settings this semester:

  • Juptyer Notebooks for quick prototyping / experimentation and reproducible examples in labs
  • Scripts running on your computer
  • Scripts running on a computer in the cloud

Getting a Python environment setup is notoriously hard and annoying.

There are many ways to install Python. It comes by default with most operating systems, but usually lags by a couple of years so it is out of date on features. Python distributions can also be built from source code, installed via a system package manager like apt-get or homebrew, run in Docker... It gets confusing fast and it's not uncommon to have multiple Python distributions installed on a computer and paths and dependencies get mixed up.

To try to keep everything standardized for this class, we'll use Anaconda's Miniconda Python distribution. Anaconda is a well-maintained (and loved) Python distribution with a robust package management framework. Miniconda is a slimmed-down installation that's nice and small. It allows for easy additions of additional packages.

Anaconda's framework lets you:

  • Create environments tagged to a specific version of Python (e.g., 3.8)
  • Create separate (sandboxed) environments so you don't have to worry about conflicts between projects
  • Install Python packages at an environment level, not system level. It is not advisable to install Python packages at the system level (e.g., don't do sudo pip install ... outside of virtual environments).

We're going to install the following packages today:

name: musa509week6
channels:
  - defaults
dependencies:
  - python=3.8
  - geopandas
  - pandas
  - rtree
  - psycopg2
  - sqlalchemy
  - seaborn
  - pip
  - jupyterlab
  - pip:
    - cartoframes
    - google-cloud-bigquery[pandas]
    - nb-black

Getting Python / Miniconda Installed

  1. Download the Miniconda distribution for your operating system
  2. Install Miniconda
  1. Open your computer terminal
    • Mac
      • Use the native Terminal app
      • iTerm2 which is what I like
    • Windows
    • Linux - if you're using Linux you probably already know what you like
  2. Run conda --version to check that everything installed correctly. You should get a response like conda 4.8.5

Installing Python env for this class

We'll use different environments depending on the project we're working on.

For our class today, there's a file in the repo called environment.yml. Let's take a look at it.

Setting up the environment

  1. Do one of the following:
  2. With your terminal, change directory (cd) into the directory where the repository is located. For example, if it is on your desktop: cd ~/Desktop/week-6-python-introduction. Here the ~ means the home directory.
  3. Create a new environment from the environment.yml file:
    • Run this command in your terminal: conda env create -f environment.yml
    • This should take a minute or two to run while it install all the packages (and their dependencies) into a new environment.
  4. Check that the environment is listed: conda env list
  5. Activate the environment. This means you enter the environment so you can use Python with the packages you installed.
    • Run this command: conda activate musa509week6
  6. Download this file to the repository directory: https://canvas.upenn.edu/courses/1533813/files/89441777/download?download_frd=1. This file contains credentials for using the database for class.

Python REPL

Let's look at the python command in the terminal.

Running a Jupyter server to view Lab Notebook

One of the packages installed is jupyterlab, which gives you a Jupyter Notebook server when you run the following command:

jupyter lab

A browser tab should pop up!

Note: this will only work if you activated your conda environment.

Once you finish with Jupyter, you can close the session with File -> Shut Down.

PostGIS Database in AWS

  • Show AWS RDS console
  • Show connection to database with PgAdmin4

Don't forget to deactivate the environment!

Run this command to deactivate the Python environment we created:

conda deactivate