amazon_jobs.py

from readypipe import requests, starting_task, subtask, schedule, save

JOBS_URL = 'https://www.amazon.jobs/en/search.json?base_query=&city=&country=&county=&facets%5B%5D=location&facets%5B%5D=business_category&facets%5B%5D=category&facets%5B%5D=schedule_type_id&facets%5B%5D=employee_class&facets%5B%5D=normalized_location&facets%5B%5D=job_function_id&latitude=&loc_group_id=&loc_query=&location%5B%5D=newyork&longitude=&offset=10&query_options=&radius=24km&region=&result_limit=10&sort=relevant'

@starting_task
def get_jobs():
    jobs = requests.get_json_from_content(JOBS_URL)
    data = {}
    data['num_jobs'] = jobs.get('hits', None)
    save('amazon_jobs', data)    
    facets = jobs.get('facets', None)
    if facets:
        locations = facets.get('location_facet', None)
        if locations:
            for location in locations:
                for key, value in location.items():
                    save('amazon_jobs_locations', {
                        'location': key,
                        'jobs': value
                    })
        businesses = facets.get('business_category_facet', None)
        if businesses:
            for business in businesses:
                for key, value in business.items():
                    save('amazon_jobs_businesses', {
                        'business': key,
                        'jobs': value
                    })        
        categories = facets.get('category_facet', None)
        if categories:
            for category in categories:
                for key, value in category.items():
                    save('amazon_jobs_categories', {
                        'category': key,
                        'jobs': value
                    })