Gets search suggestions from YouTube Music

youtube_music_suggestions.py

# AUTHOR: Noah Broyles
# DESCRIPTION: Uses the YouTube Music search suggestions API to get search suggestions from a query. API key already provided, scraped cleanly right off
# youtube's own site. ;)
#
# Last Working: Oct 25, 2021

import re
import json
import requests

from addict import Dict

# This API does use some form of key, which is automatically borrowed from YouTube ;)
INNERTUBE_API_KEY = 'AIzaSyC9XL3ZjWddXya6X74dJoCTL-WEYFDNX30'

# This is run when we are no longer getting results from the API
def refresh_api_key():
    global INNERTUBE_API_KEY
    print('Refreshing API key')
    index_page = requests.get('https://music.youtube.com', headers={"user-agent": "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:92.0) Gecko/20100101 Firefox/92.0"})
    html = index_page.content.decode()

    # The exact length of the key as of 9/28/2021 is 39. 4 years ago (https://stackoverflow.com/questions/44614612) it was also 39.
    # So it's safe to assume a length of 39.
    INNERTUBE_API_KEY = re.findall(r'"INNERTUBE_API_KEY":\s*"(.{39})",', html)[0]
    print(f"The new INNERTUBE_API_KEY is {INNERTUBE_API_KEY}")

# Returns a list of YouTubeMusic search suggestions from a query, q
def get_suggestions(q):

    API_HEADERS = {
        "origin": "https://music.youtube.com",
        "referrer": "https://music.youtube.com/",
        "sec-fetch-dest": "empty",
        "sec-fetch-mode": "same-origin",
        "sec-fetch-site": "same-origin",
        "x-youtube-client-name": "69",
        "User-Agent": "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:92.0) Gecko/20100101 Firefox/92.0,gzip(gfe)"
    }

    # I only choose to send them SOME of my data
    REQUEST_BODY = json.dumps({
        "input": q,
        "context": {
            "client": {
            "hl": "en",
            "gl": "US",
            "deviceMake": "",
            "deviceModel": "",
            #"visitorData": "CgsyR2o3ZmRqUEdjMCi5xsyKBg%3D%3D",
            "userAgent": "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:92.0) Gecko/20100101 Firefox/92.0,gzip(gfe)",
            "clientName": "WEB_REMIX",
            "clientVersion": "1.20210922.00.00",
            "osName": "X11",
            "osVersion": "",
            "originalUrl": "https://music.youtube.com/",
            "platform": "DESKTOP",
            "clientFormFactor": "UNKNOWN_FORM_FACTOR",
            #"configInfo": {
            #  "appInstallData": "CLnGzIoGEJPVrQUQ_tatBRCzhP0SENi-rQU%3D"
            #},
            "browserName": "Firefox",
            "browserVersion": "92.0",
            #"screenWidthPoints": 1316,
            #"screenHeightPoints": 605,
            #"screenPixelDensity": 1,
            #"screenDensityFloat": 1,
            #"utcOffsetMinutes": -240,
            "userInterfaceTheme": "USER_INTERFACE_THEME_DARK",
            "timeZone": "America/New_York",
            "musicAppInfo": {
                "pwaInstallabilityStatus": "PWA_INSTALLABILITY_STATUS_UNKNOWN",
                "webDisplayMode": "WEB_DISPLAY_MODE_BROWSER",
                "musicActivityMasterSwitch": "MUSIC_ACTIVITY_MASTER_SWITCH_INDETERMINATE",
                "musicLocationMasterSwitch": "MUSIC_LOCATION_MASTER_SWITCH_INDETERMINATE"
            }
            },
            "user": {
            "lockedSafetyMode": False
            },
            "request": {
            "useSsl": True,
            "internalExperimentFlags": [],
            "consistencyTokenJars": []
            }
        }
    })
    # Make the request to the API
    suggestion_data = Dict(requests.post(f'https://music.youtube.com/youtubei/v1/music/get_search_suggestions?key={INNERTUBE_API_KEY}', data=REQUEST_BODY, headers=API_HEADERS).json())
    
    # Gather our suggestions
    suggestions = [element.searchSuggestionRenderer.navigationEndpoint.searchEndpoint.query for element in suggestion_data.contents[0].searchSuggestionsSectionRenderer.contents]
        
    # The suggestions list will be empty if the API key is not correct
    if not suggestions:
        # Something about the API key was wrong
        refresh_api_key()
        return get_suggestions(q)
    else:
        return suggestions


if __name__ == '__main__':
    print(get_suggestions("yeah us"))
    
"""
['yeah usher', 'yeah usher clean', 'yeah usher lyrics', 'yeah usher remix', 'yeah usher clean radio edit', 'yeah usher instrumental', 'yeah usher slowed']
"""