Regex URLs to determine political leaning via labelled sources

url_domain_annotate.py

# List of urls (pretend like it's populated)
urls = []

# Patterns to match in urls (note in some cases including the .org or .com to avoid matching common words or letters
# (eg for npr or slate or today)
# Add whatever other domains you want to match to the re OR (|) string
left_pattern = r"(?P<domain>nytimes|washingtonpost|npr.org|abcnews|nbcnews|huffingtonpost|slate.com|today.com)"
center_pattern = r"(?P<domain>cnn|bbc.co.uk|yahoo)"
right_pattern = r"(?P<domain>foxnews|washingtontimes|usnews|chicagotribune)"

# Keep counts in dictionary
affiliation_count = {"left": 0, "right": 0, "center": 0, "unknown": 0}

for url in urls:
    if re.search(left_pattern, url):
        # Add "left" annotation to URL object in database via PYMONGO
        affiliation_count["left"] += 1
    elif re.search(right_pattern, url):
        # Add "right" annotation to URL object in database via PYMONGO
        affiliation_count["right"] += 1
    elif re.search(center_pattern, url):
        # Add "center" annotation to URL object in database via PYMONGO
        affiliation_count["center"] += 1
    else:
        # Add "unknown" annotation to URL object in database via PYMONGO
        affiliation_count["unknown"] += 1