Get All NASA Astronomy Pictures of the Day from 2019 with Python and nasapy

NASA’s Astronomy Picture of the Day 2020–02–27

NASA provides a daily feed of beautiful astronomical images like the one on the left through its NASA Astronomy Picture of the Day service, one of the most popular federal agency owned websites. Using the NASA API and nasapy with Python, we can download an entire year’s worth of astronomy imagery automatically rather than clicking through over 300 webpages and downloading the images individually.

Before starting, import the libraries that we will be using, including nasapy. The following code is assumed to be run in an Jupyter Notebook and thus uses the notebook function IPython.display to view some of the images that we will download.

import nasapy
import os
import pandas as pd
import datetime
import urllib.request
from IPython.display import Image

The NASA APOD (Astronomy Picture of the Day) API endpoint requires a registered API key to be included in the request. To receive an API key, an account can be created at the NASA API website. A DEMO_KEY can also be passed instead of an actual key, though the request limit is greatly restricted. We initialize the Nasa class from nasapy with the key received from creating an account.

key = os.environ.get('NASA_KEY')
n = nasapy.Nasa(key=key)

Before downloading the images available in 2019, we need to get the dates in 2019! Because typing 365 days worth of strings into a list would be quite tedious, we instead leverage pandas' date_range function in combination with a few other pandas methods to get a list of all dates in 2019.

dates_2019 = pd.date_range(end = '2019-12-31', periods=365).to_pydatetime().tolist()
print(len(dates_2019))

The NASA APOD API endpoint only gives us the URLs of the available image for a given date; thus, we first loop through all the dates in the dates_2019 list and get the available image URL. The API endpoint also provides a URL link for a high-definition image if one is available.

urls = []

for date in dates_2019:
    apod = n.picture_of_the_day(date, hd=True)
    
    if apod['media_type'] == 'image':
        if 'hdurl' not in apod.keys():
            apod['hdurl'] = None

        urls.append({'date': apod['date'], 
                     'title': apod['title'], 
                     'sd': apod['url'], 
                     'hd': apod['hdurl']})

After getting all the available URLs, we can print the length to see how many days of images we were able to extract.

print(len(urls))

We can now begin downloading the images! To separate the standard and high-definition images, we first create two directories using the makedirs function from the os standard library. The urlretrieve function from urllib.request is used to download the images into the relevant directories. The downloaded images are each named according to their title, the image quality, and the date.

sd_dir = 'apod/sd'
hd_dir = 'apod/hd'

os.makedirs(sd_dir)
os.makedirs(hd_dir)

for img in urls:   
    urllib.request.urlretrieve(img['sd'], 
                               os.path.join(sd_dir, 
                                            img['title'].lower().replace(' ', '_').replace('/', '-') + 
                                            '_sd_' + img['date'].replace(' ', '_')) + '.jpg')
    
    if img['hd'] is not None:
        urllib.request.urlretrieve(img['hd'], 
                                   os.path.join(hd_dir, 
                                                img['title'].lower().replace(' ', '_').replace('/', '-') + 
                                                '_hd_' + img['date'].replace(' ', '_')) + '.jpg')

After the images are all downloaded into their respective directories, we can get a list of all the high-definition images using the os function listdir.

hd_images = os.listdir(hd_dir)

Combining the relative path directory to the high-definition image location and the first image in the list using os.path.join, we can view the first image downloaded into the directory!

Image(os.path.join(hd_dir, hd_images[0]), width=480, height=480)

Beautiful! Space and astronomy genuinely are gorgeous. As we downloaded so many different images, it might be hard to decide on which one to view, so we leverage the randint function to load a random image from the directory.

from random import randint
Image(os.path.join(hd_dir, hd_images[randint(0, len(hd_images))]), width=480, height=480)