# This is an example of a feature-complete JobFunnel configuration YAML. # Try this out by simply running: "funnel load -s demo/settings.yaml" # Path where your master CSV, block-lists, and cache data will be stored # NOTE: we create any missing directories in these filepaths master_csv_file: demo_job_search_results/demo_search.csv cache_folder: demo_job_search_results/cache block_list_file: demo_job_search_results/demo_block_list.json duplicates_list_file: demo_job_search_results/demo_duplicates_list.json log_file: demo_job_search_results/log.log # Job search configuration search: # Locale settings, one of USA_ENGLISH, CANADA_ENGLISH, CANADA_FRENCH, # UK_ENGLISH, FRANCE_FRENCH: # This tells JobFunnel where the website we are scraping is located, and # what language the contents are in. locale: CANADA_ENGLISH # Job providers which we will search, one of INDEED, MONSTER, GLASSDOOR: # NOTE: we choose domain via locale (i.e. CANADA_ENGLISH -> www.indeed.ca) # FIXME: we need to add back GLASSDOOR when that's working again. providers: - INDEED - MONSTER # Region that we are searching for jobs within: province_or_state: "ON" # NOTE: this is generally 2 characters long. city: "Waterloo" # NOTE: this is the full city / town name. radius: 25 # km (NOTE: if we were in locale: USA_ENGLISH it's in miles) # These are the terms you would be typing into the website's search field: keywords: - Python # Don't return any listings older than this: max_listing_days: 35 # Blocked company names that will never appear in any results: company_block_list: - "Infox Consulting" # The desired level of work-remoteness (i.e. IN_PERSON, FULLY_REMOTE, ANY, # TEMPORARILY_REMOTE, PARTIALLY_REMOTE) remoteness: ANY # Logging level options are: critical, error, warning, info, debug, notset log_level: INFO # Delaying algorithm configuration delay: # Functions used for delaying algorithm: CONSTANT, LINEAR, SIGMOID algorithm: LINEAR # Maximum delay/upper bound for converging random delay max_duration: 5.0 # Minimum delay/lower bound for random delay min_duration: 1.0 # Random delay random: False # Converging random delay, only used if 'random' is set to True converging: False # # Proxy settings # proxy: # protocol: https # NOTE: you can also set to 'http' # ip: "1.1.1.1" # port: 200