Scrap ton VIE

Scrap ton VIE aims at providing a convenient way to find VIE job offers. A development version lives at : https://vie.johanet.fr/

Note : on 11/21/19, SSL certificates got messed up on the civiweb website. I had to skip verifying certificates, although this is probably going to be fixed. To ignore urllib3 warning from requests, use :

export PYTHONWARNINGS="ignore:Unverified HTTPS request"

Development

Front-end

Install dependencies :

npm install

Launch the development server with :

npm run watch

Back-end

npm install

npm run build

Start scraping by running python scrap.py with a proper config.ini in the same directory, e.g :

[credentials]
api_key = ABCDEFGHIJKLMNOPQSRTUVW

[civiweb]
offer_list = https://www.civiweb.com/FR/offre-liste/page/
offer_page = https://www.civiweb.com/FR/offre/

[db]
database = my_database
hostname = localhost
port = 5432
username = my_username
password = my_password

Production

Front-end

In production, use npm run build then serve the index.html with the dist folder with any web server.

Back-end

Run the scrapper as a CRON job. Inclure full paths to avoid $PATH errors, and include stderr in the log file.

crontab -e
0 1 */5 * * /usr/bin/env python3.6 /home/user/scrap/scrap.py >> /home/user/scrap/scrap.log 2>&1

Build the NodeJS app with npm run build then start the server with forever or pm2.

Play with data

Calculate VIE frequency percentage per country

WITH total AS (SELECT COUNT(*)::numeric FROM offer)
SELECT COUNT(country) AS nombre, ROUND(COUNT(country)/(SELECT * FROM total), 4) * 100 AS percentage , country FROM offer
GROUP BY country
ORDER BY nombre DESC

Find highest paying locations

SELECT country, UPPER(city), salary FROM offer
GROUP BY salary, UPPER(city), country
ORDER BY salary DESC

Calculate n-quantiles

SELECT salary, ntile(n) OVER (ORDER BY salary) AS quantile FROM offer GROUP BY salary;