Introduction
The main plan was to develop a COVID-19 tracker in Django that shows local data in the Philippines - until the Department of Health hid and disabled page where the stats can be scraped.
I did some crowdsourcing on Facebook and everyone shared a photo of the daily stats per city - which wasn't what I expected - but it was my fault, my question wasn't clear.
Coding
I already have experience in playing with API's so I thought it would be easier now since I could just use my code as a reference.
I found Chris Michael's from google and found it easy to use. The API returned:
{
"report":
{
"country":"philippines",
"flag":"https://www.worldometers.info/img/flags/small/tn_rp-flag.gif",
"cases":7579,
"deaths":501,
"recovered":862,
"active_cases":[],
"closed_cases":[]
}
}
The data isn't complete but that was fine. So in calling and using the API, I ended up with:
api.py
import requests
def get_data():
url = "https://covid19-server.chrismichael.now.sh/api/v1/ReportsByCountries/philippines"
headers = {'Accept': 'application/json'}
r = requests.get(url, headers=headers)
stat = r.json()
stat_data = {
'stat': stat['report']
}
return stat_data
Which was pretty much the same as my first API experience:
Although this was not my first time using an API in Python, this was, however, my first time integrating an API with Django. I ended up with:
views.py
...
from django.views.generic import ListView
from covid_virus_tracker.users import api
...
class HomeView(ListView):
def get(self, request):
stat_data = api.get_data()
return render(request, 'pages/home.html', stat_data)
However, the api.py was registering as a python module and I honestly had a hard time figuring out why. Until I peeked into the source code and figured out that when importing in cookiecutter-django, you've to start from the project, down to the app, then to the file, if coming from another app.
Ex: from project_name.users.models import view
Then on the pages/home.html, all I had to do was:
{{ stat.recovered}} {{ stat.cases }} {{ stat.deaths }}
However, I felt like the site needed something more. I felt like I had to let the users know what's happening all over the world, so I integrated an API that scrapes blogs with a COVID-19 keyword. I found RapidAPI, which I had an account already, and found the contextualsearch API. The API returned:
{
"type": "news",
"didUMean": "",
"totalCount": 2922,
"relatedSearch": [
"coronavirus",
"ist",
"updated",
"pm",
"daily",
"cases",
"delhi",
"india",
"published",
"times",
"sign",
"video online"
],
"value": [
{
"title": "Grounded jazz saxophonist plans to 'dream big' after lockdown",
"url": "https://www.smh.com.au/culture/music/grounded-jazz-saxophonist-plans-to-dream-big-after-lockdown-20200427-p54nis.html?ref=rss&utm_medium=rss&utm_source=rss_feed",
"description": "The coronavirus Great Pause trashed Flora Carbo's plans for a world tour with a new album. Now she's fighting snails and making plans.",
"body": "Very large text size\nARTIST IN THE TIME OF COVID-19: Flora Carbo\nJazz saxophonist Flora Carbo has recently released a new album, Voice. The next logical step? A launch then a tour two things the Fairfield artist cant do thanks to the coronavirus.\nI had a six-week European trip cancelled; I was going to play at the Amersfoort Jazz Festival [Netherlands], attend the Jazzahead! conference in Germany and collaborate with some friends.\nJazz saxophonist Flora Carbo.\nCarbo was also going to capitalise on some serious buzz with gigs at the Melbourne International Jazz Festival (now cancelled, natch), where shes previously played with Maceo Parker and the Meltdown Big Band.\nAdvertisement\nShe was all set to front Flora Carbo Trio in the lunchtime series and I was going to be playing with [trumpeter] Niran Dasika as part of his group\".\nNow, Carbo is fighting snails out of her vegie garden, re-booking gigs, completing sudokus and 100 per cent subsisting on hot chocolate\".\nIm very lucky to have be",
"keywords": "covid,time",
"language": "en",
"isSafe": true,
"datePublished": "2020-04-27T01:35:36",
"provider": {
"name": "smh"
},
"image": {
"url": "https://static.ffx.io/images/$zoom_1.32375%2C$multiply_0.7554%2C$ratio_1.776846%2C$width_1059%2C$x_0%2C$y_245/t_crop_custom/q_86%2Cf_auto/8a1b5ccca1398947c8147aae6400e5a7fecf2359",
"height": 450,
"width": 800,
"thumbnail": "https://rapidapi.contextualwebsearch.com/api/thumbnail/get?value=5941184483138831277",
"thumbnailHeight": 168,
"thumbnailWidth": 298,
"base64Encoding": null,
"name": null,
"title": null,
"imageWebSearchUrl": null
}
}
]
}
Integrating was easier now since I've done it on the api.py. I ended up with:
import requests
import environ
env = environ.Env()
def get_data():
url = "https://contextualwebsearch-websearch-v1.p.rapidapi.com/api/Search/NewsSearchAPI"
querystring = {"autoCorrect": "false", "pageNumber": "1", "pageSize": "10", "q": "covid", "safeSearch": "true"}
headers = {
'x-rapidapi-host': "contextualwebsearch-websearch-v1.p.rapidapi.com",
'x-rapidapi-key': env("x-rapidapi-key"),
}
r = requests.get(url, headers=headers, params=querystring)
data = r.json()
article_data = {
'data': data['value']
}
return article_data
Notice that one difference is having to import environ and creating an env = environ.ENV() variable. I need this to hide my API keys and this is how cookiecutter-django hides it. All I had to do was place the keys in the .envs/local/.django & .envs/production/.django and I was good to go.
The views.py is:
...
from django.views.generic import ListView
from covid_virus_tracker.users import services
...
class IndexData(ListView):
def get(self, request):
article_data = services.get_data()
return render(request, 'pages/news.html', article_data)
On the pages/news.html, all I needed to do was loop through the data to show it. Like this:
{% for item in data %}
{{ item.title }}
{{ item.datePublished|slice:":10" }} # | naturaltime from humanize didn't work so I had to slice through the date
{{ item.description }}
{{ item.provider.name }}
The site is live and can be seen at https://covid19ph-unofficial.icvn.tech/.
Top comments (14)
quick markdown/DEV tip, make sure you put the language of your code next to your code block.
Hi, what's the use of putting the language at the top?
It is what enables syntax hilighting
Isn't it highlighted already before you gave the tip?
Your article does not look highlighted to me.
That's weird. Check out all the code blocks on all the themes on the photos below:
Hey Vicente! Thanks for sharing and this looks pretty cool. I was wondering if you ever considered using a realtime stream of COVID updates to build an event-driven application instead of relying on polling an API for data updates? I wrote a blog post about it here dev.to/tweettamimi/how-i-used-covi... let me know if you're interested to collab 🙌
Hey Tamimi, what I actually have in mind is, since someone from the Philippines is taking down notes on where the budget for COVID is going, I'm planning on adding the said data on the site. But it's still on planning stage tho.
Awesome blog post! I'm actually still a novice in Django but I'm in. I could learn a thing or two from you anyways.
Sounds good! Have you done any front end development with python as a backend (i.e. using Django)? We can start with the code snippet in my blodpost as the event-driven skeleton receiving updates from any of the data sources. Subscription to topics could be as easy as subscribing to
"jhu/csse/covid19/test/cases/+/update/Philippines/#"for example to get all cases updates from the Philippines. It'll be a cool exploration of using Python/Django and messaging protocol mqttLet's speak in the chat 😄
Such a great community project!! Thanks for sharing!
Good job
I was part of the DICT team that planned to develop our local tracker but they used Java and Angular hence deciding to make my own.
Thank you.