-
-
Save huksley/bc3cb046157a99cd9d1517b32f91a99e to your computer and use it in GitHub Desktop.
/** | |
* This magically uses batchexecute protocol. It's not documented, but it works. | |
* | |
* Licensed under: MIT License | |
* | |
* Copyright (c) 2024 Ruslan Gainutdinov | |
* | |
* Permission is hereby granted, free of charge, to any person obtaining a copy | |
* of this software and associated documentation files (the "Software"), to deal | |
* in the Software without restriction, including without limitation the rights | |
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell | |
* copies of the Software, and to permit persons to whom the Software is | |
* furnished to do so, subject to the following conditions: | |
* | |
* The above copyright notice and this permission notice shall be included | |
* in all copies or substantial portions of the Software. | |
* | |
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR | |
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, | |
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE | |
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER | |
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, | |
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE | |
* SOFTWARE. | |
*/ | |
const fetchDecodedBatchExecute = (id: string) => { | |
const s = | |
'[[["Fbv4je","[\\"garturlreq\\",[[\\"en-US\\",\\"US\\",[\\"FINANCE_TOP_INDICES\\",\\"WEB_TEST_1_0_0\\"],null,null,1,1,\\"US:en\\",null,180,null,null,null,null,null,0,null,null,[1608992183,723341000]],\\"en-US\\",\\"US\\",1,[2,3,4,8],1,0,\\"655000234\\",0,0,null,0],\\"' + | |
id + | |
'\\"]",null,"generic"]]]'; | |
return fetch("https://news.google.com/_/DotsSplashUi/data/batchexecute?" + "rpcids=Fbv4je", { | |
headers: { | |
"Content-Type": "application/x-www-form-urlencoded;charset=utf-8", | |
Referrer: "https://news.google.com/" | |
}, | |
body: "f.req=" + encodeURIComponent(s), | |
method: "POST" | |
}) | |
.then(e => e.text()) | |
.then(s => { | |
const header = '[\\"garturlres\\",\\"'; | |
const footer = '\\",'; | |
if (!s.includes(header)) { | |
throw new Error("header not found: " + s); | |
} | |
const start = s.substring(s.indexOf(header) + header.length); | |
if (!start.includes(footer)) { | |
throw new Error("footer not found"); | |
} | |
const url = start.substring(0, start.indexOf(footer)); | |
return url; | |
}); | |
}; | |
/** | |
* Google News started generate encoded, internal URLs for RSS items | |
* https://news.google.com/rss/search?q=New%20York%20when%3A30d&hl=en-US&gl=US&ceid=US:en | |
* | |
* This script decodes URLs into original one, for example URL | |
* https://news.google.com/__i/rss/rd/articles/CBMiSGh0dHBzOi8vdGVjaGNydW5jaC5jb20vMjAyMi8xMC8yNy9uZXcteW9yay1wb3N0LWhhY2tlZC1vZmZlbnNpdmUtdHdlZXRzL9IBAA?oc=5 | |
* | |
* contains this | |
* https://techcrunch.com/2022/10/27/new-york-post-hacked-offensive-tweets/ | |
* | |
* In path after articles/ goes Base64 encoded binary data | |
* | |
* Format is the following: | |
* <prefix> <len bytes> <URL bytes> <len bytes> <amp URL bytes> [<suffix>] | |
* | |
* <prefix> - 0x08, 0x13, 0x22 | |
* <suffix> - 0xd2, 0x01, 0x00 (sometimes missing??) | |
* <len bytes> - formatted as 0x40 or 0x81 0x01 sometimes | |
* | |
* | |
* https://news.google.com/rss/articles/CBMiqwFBVV95cUxNMTRqdUZpNl9hQldXbGo2YVVLOGFQdkFLYldlMUxUVlNEaElsYjRRODVUMkF3R1RYdWxvT1NoVzdUYS0xSHg3eVdpTjdVODQ5cVJJLWt4dk9vZFBScVp2ZmpzQXZZRy1ncDM5c2tRbXBVVHVrQnpmMGVrQXNkQVItV3h4dVQ1V1BTbjhnM3k2ZUdPdnhVOFk1NmllNTZkdGJTbW9NX0k5U3E2Tkk?oc=5 | |
* https://news.google.com/rss/articles/CBMidkFVX3lxTFB1QmFsSi1Zc3dLQkpNLThKTXExWXBGWlE0eERJQ2hLRENIOFJzRTlsRnM1NS1Hc2FlbjdIMlZ3eWNQa0JqeVYzZGs1Y0hKaUtTUko2dmJabUtVMWZob0lNSFNCa3NLQ05ROGh4cVZfVTYyUDVxc2c?oc=5 | |
* https://news.google.com/rss/articles/CBMiqwFBVV95cUxNMTRqdUZpNl9hQldXbGo2YVVLOGFQdkFLYldlMUxUVlNEaElsYjRRODVUMkF3R1RYdWxvT1NoVzdUYS0xSHg3eVdpTjdVODQ5cVJJLWt4dk9vZFBScVp2ZmpzQXZZRy1ncDM5c2tRbXBVVHVrQnpmMGVrQXNkQVItV3h4dVQ1V1BTbjhnM3k2ZUdPdnhVOFk1NmllNTZkdGJTbW9NX0k5U3E2Tkk?oc=5 | |
* | |
* FIXME: What will happen if URL more than 255 bytes?? | |
* | |
* Licensed under: MIT License | |
* | |
* Copyright (c) 2022 Ruslan Gainutdinov | |
* | |
* Permission is hereby granted, free of charge, to any person obtaining a copy | |
* of this software and associated documentation files (the "Software"), to deal | |
* in the Software without restriction, including without limitation the rights | |
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell | |
* copies of the Software, and to permit persons to whom the Software is | |
* furnished to do so, subject to the following conditions: | |
* | |
* The above copyright notice and this permission notice shall be included | |
* in all copies or substantial portions of the Software. | |
* | |
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR | |
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, | |
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE | |
* AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER | |
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, | |
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE | |
* SOFTWARE. | |
*/ | |
export const decodeGoogleNewsUrl = async (sourceUrl: string) => { | |
const url = new URL(sourceUrl); | |
const path = url.pathname.split("/"); | |
if ( | |
url.hostname === "news.google.com" && | |
path.length > 1 && | |
path[path.length - 2] === "articles" | |
) { | |
const base64 = path[path.length - 1]; | |
let str = atob(base64); | |
const prefix = Buffer.from([0x08, 0x13, 0x22]).toString("binary"); | |
if (str.startsWith(prefix)) { | |
str = str.substring(prefix.length); | |
} | |
const suffix = Buffer.from([0xd2, 0x01, 0x00]).toString("binary"); | |
if (str.endsWith(suffix)) { | |
str = str.substring(0, str.length - suffix.length); | |
} | |
// One or two bytes to skip | |
const bytes = Uint8Array.from(str, c => c.charCodeAt(0)); | |
const len = bytes.at(0)!; | |
if (len >= 0x80) { | |
str = str.substring(2, len + 2); | |
} else { | |
str = str.substring(1, len + 1); | |
} | |
if (str.startsWith("AU_yqL")) { | |
// New style encoding, introduced in July 2024. Not yet known how to decode offline. | |
const url = await fetchDecodedBatchExecute(base64); | |
return url; | |
} | |
return str; | |
} else { | |
return sourceUrl; | |
} | |
}; |
@huksley this request solution is broken
Unfortunately, the solution that @huksley proposed no longer works. It produces errors like this: [["wrb.fr","Fbv4je",null,null,null,[3],"generic"],["di",10],["af.httprm",10,"2111786207358723693",9]]
.
@huksley surprisingly the request solution is not working. Can help to fix.
yes, they have added few more parameters at end of payload, looks like some encrypted values
,1725004324,\"ATR1dL9qAQrN8uy3dkKVSj-G9RHc\"]",null,"generic"]]]&at=AEtveWhZ98E6YWfBFsXKcv6oDg_O:1725004324622&
below attached compare image, here right side contains previous payload and left side contains current payload
Any solutions?
Any solutions?
Tried to decode it, but no luck
payload = (
'[[["Fbv4je","[\"garturlreq\",[[\"en-US\",\"US\",[\"FINANCE_TOP_INDICES\",\"WEB_TEST_1_0_0\"],null,null,1,1,\"US:en\",null,360,null,null,null,null,null,0,null,null,[1677434405,738601000]],\"en-US\",\"US\",1,[2,3,4,8],1,0,\"668194412\",0,0,null,0],\"'
+ code
+ '\",1725016444,\"ATR1dL9R_yt7riBAiulU9qaZcXAJ\"]",null,"generic"]]]'
)
For 1725016444 it's a Unix timestamp. Finding solution for ATR1dL9R_yt7riBAiulU9qaZcXAJ
payload = ( '[[["Fbv4je","["garturlreq",[["en-US","US",["FINANCE_TOP_INDICES","WEB_TEST_1_0_0"],null,null,1,1,"US:en",null,360,null,null,null,null,null,0,null,null,[1677434405,738601000]],"en-US","US",1,[2,3,4,8],1,0,"668194412",0,0,null,0],"' + code + '",1725016444,"ATR1dL9R_yt7riBAiulU9qaZcXAJ"]",null,"generic"]]]' )
For 1725016444 it's a Unix timestamp. Finding solution for ATR1dL9R_yt7riBAiulU9qaZcXAJ
Will you update your python package as well if you find the solution?
payload = ( '[[["Fbv4je","["garturlreq",[["en-US","US",["FINANCE_TOP_INDICES","WEB_TEST_1_0_0"],null,null,1,1,"US:en",null,360,null,null,null,null,null,0,null,null,[1677434405,738601000]],"en-US","US",1,[2,3,4,8],1,0,"668194412",0,0,null,0],"' + code + '",1725016444,"ATR1dL9R_yt7riBAiulU9qaZcXAJ"]",null,"generic"]]]' )
For 1725016444 it's a Unix timestamp. Finding solution for ATR1dL9R_yt7riBAiulU9qaZcXAJWill you update your python package as well if you find the solution?
trying and waiting for @huksley
PYTHON SOLUTION
I don't know the exact algorithm used by Google to encode/decode the URLs, but I found a way to decode them using reverse engineering by inspecting the requests made by the browser in the redirection chain.
pip install beautifulsoup4 lxml
import json
from urllib.parse import quote, urlparse
import requests
from bs4 import BeautifulSoup
def get_decoding_params(gn_art_id):
response = requests.get(f"https://news.google.com/articles/{gn_art_id}")
response.raise_for_status()
soup = BeautifulSoup(response.text, "lxml")
div = soup.select_one("c-wiz > div")
return {
"signature": div.get("data-n-a-sg"),
"timestamp": div.get("data-n-a-ts"),
"gn_art_id": gn_art_id,
}
def decode_urls(articles):
articles_reqs = [
[
"Fbv4je",
f'["garturlreq",[["X","X",["X","X"],null,null,1,1,"US:en",null,1,null,null,null,null,null,0,1],"X","X",1,[1,1,1],1,1,null,0,0,null,0],"{art["gn_art_id"]}",{art["timestamp"]},"{art["signature"]}"]',
]
for art in articles
]
payload = f"f.req={quote(json.dumps([articles_reqs]))}"
headers = {"content-type": "application/x-www-form-urlencoded;charset=UTF-8"}
response = requests.post(
url="https://news.google.com/_/DotsSplashUi/data/batchexecute",
headers=headers,
data=payload,
)
response.raise_for_status()
return [json.loads(res[2])[1] for res in json.loads(response.text.split("\n\n")[1])[:-2]]
# Example usage
encoded_urls = [
"https://news.google.com/rss/articles/CBMipgFBVV95cUxPWV9fTEI4cjh1RndwanpzNVliMUh6czg2X1RjeEN0YUctUmlZb0FyeV9oT3RWM1JrMGRodGtqTk1zV3pkNEpmdGNxc2lfd0c4LVpGVENvUDFMOEJqc0FCVVExSlRrQmI3TWZ2NUc4dy1EVXF4YnBLaGZ4cTFMQXFFM2JpanhDR3hoRmthUjVjdm1najZsaFh4a3lBbDladDZtVS1FMHFn?oc=5",
"https://news.google.com/rss/articles/CBMi3AFBVV95cUxOX01TWDZZN2J5LWlmU3hudGZaRDh6a1dxUHMtalBEY1c0TlJSNlpieWxaUkxUU19MVTN3Y1BqaUZael83d1ctNXhaQUtPM0IyMFc4R3VydEtoMmFYMWpMU1Rtc3BjYmY4d3gxZHlMZG5NX0s1RmR2ZXI5YllvdzNSd2xkOFNCUTZTaEp3b0IxZEJZdVFLUDBNMC1wNGgwMGhjRG9HRFpRZU5BMFVIYjZCOWdWcHI1YzdoVHFWYnZSOEFwQ0NubGx3Rzd0SHN6OENKMXZUcHUxazA5WTIw?hl=en-US&gl=US&ceid=US%3Aen",
]
articles_params = [get_decoding_params(urlparse(url).path.split("/")[-1]) for url in encoded_urls]
decoded_urls = decode_urls(articles_params)
print(decoded_urls)
PYTHON SOLUTION
I don't know the exact algorithm used by Google to encode/decode the URLs, but I found a way to decode them using reverse engineering by inspecting the requests made by the browser in the redirection chain.
pip install beautifulsoup4 lxml
import json from urllib.parse import quote, urlparse import requests from bs4 import BeautifulSoup def get_decoding_params(gn_art_id): response = requests.get(f"https://news.google.com/articles/{gn_art_id}") response.raise_for_status() soup = BeautifulSoup(response.text, "lxml") div = soup.select_one("c-wiz > div") return { "signature": div.get("data-n-a-sg"), "timestamp": div.get("data-n-a-ts"), "gn_art_id": gn_art_id, } def decode_urls(articles): articles_reqs = [ [ "Fbv4je", f'["garturlreq",[["X","X",["X","X"],null,null,1,1,"US:en",null,1,null,null,null,null,null,0,1],"X","X",1,[1,1,1],1,1,null,0,0,null,0],"{art["gn_art_id"]}",{art["timestamp"]},"{art["signature"]}"]', ] for art in articles ] payload = f"f.req={quote(json.dumps([articles_reqs]))}" headers = {"content-type": "application/x-www-form-urlencoded;charset=UTF-8"} response = requests.post( url="https://news.google.com/_/DotsSplashUi/data/batchexecute", headers=headers, data=payload, ) response.raise_for_status() return [json.loads(res[2])[1] for res in json.loads(response.text.split("\n\n")[1])[:-2]] # Example usage encoded_urls = [ "https://news.google.com/rss/articles/CBMipgFBVV95cUxPWV9fTEI4cjh1RndwanpzNVliMUh6czg2X1RjeEN0YUctUmlZb0FyeV9oT3RWM1JrMGRodGtqTk1zV3pkNEpmdGNxc2lfd0c4LVpGVENvUDFMOEJqc0FCVVExSlRrQmI3TWZ2NUc4dy1EVXF4YnBLaGZ4cTFMQXFFM2JpanhDR3hoRmthUjVjdm1najZsaFh4a3lBbDladDZtVS1FMHFn?oc=5", "https://news.google.com/rss/articles/CBMi3AFBVV95cUxOX01TWDZZN2J5LWlmU3hudGZaRDh6a1dxUHMtalBEY1c0TlJSNlpieWxaUkxUU19MVTN3Y1BqaUZael83d1ctNXhaQUtPM0IyMFc4R3VydEtoMmFYMWpMU1Rtc3BjYmY4d3gxZHlMZG5NX0s1RmR2ZXI5YllvdzNSd2xkOFNCUTZTaEp3b0IxZEJZdVFLUDBNMC1wNGgwMGhjRG9HRFpRZU5BMFVIYjZCOWdWcHI1YzdoVHFWYnZSOEFwQ0NubGx3Rzd0SHN6OENKMXZUcHUxazA5WTIw?hl=en-US&gl=US&ceid=US%3Aen", ] articles_params = [get_decoding_params(urlparse(url).path.split("/")[-1]) for url in encoded_urls] decoded_urls = decode_urls(articles_params) print(decoded_urls)
Man, you are a genius. I never thought that way. I was checking using httptoolkit, my bad i never notice. which tool do you use? httpdebugger? Can I implement it in my module? https://github.com/SSujitX/google-news-url-decoder
def get_decoding_params(gn_art_id):
response = requests.get(f"https://news.google.com/articles/{gn_art_id}")
response.raise_for_status()
soup = BeautifulSoup(response.text, "lxml")
div = soup.select_one("c-wiz > div")
return {
"signature": div.get("data-n-a-sg"),
"timestamp": div.get("data-n-a-ts"),
"gn_art_id": gn_art_id,
}
def decode_google_news_url(source_url):
article = get_decoding_params(urlparse(source_url).path.split("/")[-1])
articles_req = [
"Fbv4je",
f'["garturlreq",[["X","X",["X","X"],null,null,1,1,"US:en",null,1,null,null,null,null,null,0,1],"X","X",1,[1,1,1],1,1,null,0,0,null,0],"{article["gn_art_id"]}",{article["timestamp"]},"{article["signature"]}"]',
]
response = requests.post(
url="https://news.google.com/_/DotsSplashUi/data/batchexecute",
headers={"content-type": "application/x-www-form-urlencoded;charset=UTF-8"},
data=f"f.req={quote(json.dumps([[articles_req]]))}",
)
response.raise_for_status()
return json.loads(json.loads(response.text.split("\n\n")[1])[:-2][0][2])[1]
# Example usage
encoded = "https://news.google.com/rss/articles/CBMiwwFBVV95cUxPdEpINnp6em8wMkZnSndsLTlmUkRlWjRyeDlGS1E1WHRmX0E2QXo0S0ZxZ2FCeUkzMnRYRm9wZEE4RGE5bzZnZGdFZUw2VWRSQ0pfcG9WQ1JyWDg3cGVZMFd2Vk4zUDhWSF8tMm45TTdsLVJLdGtLUjB6QlJSWlNfU0gwOEdzY3RtakJFTDB2bzdrUXdnZVRaWGZhVUZjWmdiNXdXX1FyODY5RnBXYTVLUFRHYUJvY25TQzhRQWNydHctR1E?oc=5&hl=en-US&gl=US&ceid=US:en"
decoded_url = decode_google_news_url(encoded)
# prints https://www.forbes.com/sites/digital-assets/2024/08/29/from-polymarket-predictions-to-press-on-nails-crypto-moves-mainstream/
print(decoded_url)
TY @jacoboisaza !! For anyone who wants a solution that takes in a single url instead of an array of URLs...
so it’s making multiple requests? seems like google really doesn’t want us to decode urls. what are the chances they block us?
Updated on Repo
pip install googlenewsdecoder --upgrade
from googlenewsdecoder import new_decoderv1
def main():
interval_time = 5 # default interval is 1 sec, if not specified
source_url = "https://news.google.com/read/CBMi2AFBVV95cUxPd1ZCc1loODVVNHpnbFFTVHFkTG94eWh1NWhTeE9yT1RyNTRXMVV2S1VIUFM3ZlVkVjl6UHh3RkJ0bXdaTVRlcHBjMWFWTkhvZWVuM3pBMEtEdlllRDBveGdIUm9GUnJ4ajd1YWR5cWs3VFA5V2dsZnY1RDZhVDdORHRSSE9EalF2TndWdlh4bkJOWU5UMTdIV2RCc285Q2p3MFA4WnpodUNqN1RNREMwa3d5T2ZHS0JlX0MySGZLc01kWDNtUEkzemtkbWhTZXdQTmdfU1JJaXY?hl=en-US&gl=US&ceid=US%3Aen"
try:
decoded_url = new_decoderv1(source_url, interval=interval_time)
if decoded_url.get("status"):
print("Decoded URL:", decoded_url["decoded_url"])
else:
print("Error:", decoded_url["message"])
except Exception as e:
print(f"Error occurred: {e}")
# Output: decoded_urls - [{'status': True, 'decoded_url': 'https://healthdatamanagement.com/articles/empowering-the-quintuple-aim-embracing-an-essential-architecture/'}]
if __name__ == "__main__":
main()
Yes, the new challenge is to hack the requests limits...
so it’s making multiple requests? seems like google really doesn’t want us to decode urls. what are the chances they block us?
PYTHON SOLUTION
I don't know the exact algorithm used by Google to encode/decode the URLs, but I found a way to decode them using reverse engineering by inspecting the requests made by the browser in the redirection chain.
pip install beautifulsoup4 lxml
import json from urllib.parse import quote, urlparse import requests from bs4 import BeautifulSoup def get_decoding_params(gn_art_id): response = requests.get(f"https://news.google.com/articles/{gn_art_id}") response.raise_for_status() soup = BeautifulSoup(response.text, "lxml") div = soup.select_one("c-wiz > div") return { "signature": div.get("data-n-a-sg"), "timestamp": div.get("data-n-a-ts"), "gn_art_id": gn_art_id, } def decode_urls(articles): articles_reqs = [ [ "Fbv4je", f'["garturlreq",[["X","X",["X","X"],null,null,1,1,"US:en",null,1,null,null,null,null,null,0,1],"X","X",1,[1,1,1],1,1,null,0,0,null,0],"{art["gn_art_id"]}",{art["timestamp"]},"{art["signature"]}"]', ] for art in articles ] payload = f"f.req={quote(json.dumps([articles_reqs]))}" headers = {"content-type": "application/x-www-form-urlencoded;charset=UTF-8"} response = requests.post( url="https://news.google.com/_/DotsSplashUi/data/batchexecute", headers=headers, data=payload, ) response.raise_for_status() return [json.loads(res[2])[1] for res in json.loads(response.text.split("\n\n")[1])[:-2]] # Example usage encoded_urls = [ "https://news.google.com/rss/articles/CBMipgFBVV95cUxPWV9fTEI4cjh1RndwanpzNVliMUh6czg2X1RjeEN0YUctUmlZb0FyeV9oT3RWM1JrMGRodGtqTk1zV3pkNEpmdGNxc2lfd0c4LVpGVENvUDFMOEJqc0FCVVExSlRrQmI3TWZ2NUc4dy1EVXF4YnBLaGZ4cTFMQXFFM2JpanhDR3hoRmthUjVjdm1najZsaFh4a3lBbDladDZtVS1FMHFn?oc=5", "https://news.google.com/rss/articles/CBMi3AFBVV95cUxOX01TWDZZN2J5LWlmU3hudGZaRDh6a1dxUHMtalBEY1c0TlJSNlpieWxaUkxUU19MVTN3Y1BqaUZael83d1ctNXhaQUtPM0IyMFc4R3VydEtoMmFYMWpMU1Rtc3BjYmY4d3gxZHlMZG5NX0s1RmR2ZXI5YllvdzNSd2xkOFNCUTZTaEp3b0IxZEJZdVFLUDBNMC1wNGgwMGhjRG9HRFpRZU5BMFVIYjZCOWdWcHI1YzdoVHFWYnZSOEFwQ0NubGx3Rzd0SHN6OENKMXZUcHUxazA5WTIw?hl=en-US&gl=US&ceid=US%3Aen", ] articles_params = [get_decoding_params(urlparse(url).path.split("/")[-1]) for url in encoded_urls] decoded_urls = decode_urls(articles_params) print(decoded_urls)
Man, you are a genius. I never thought that way. I was checking using httptoolkit, my bad i never notice. which tool do you use? httpdebugger? Can I implement it in my module? https://github.com/SSujitX/google-news-url-decoder
it's open source so take it, mix it, and share it
I used Chrome DevTools to track the redirections and inspect the payloads and responses. After many attempts at cleaning and simplifying the process, I discovered this minimal approach that works well in Python.
Is there a PHP solution?
I converted all the requests using GPT, licked the code, but nothing works :( I tested the test.php fome with those function through a browser.
I'm already thinking - maybe I shouldn't run this code from the browser?
But I don't see any point in it.
I converted this to PHP and got it working but goggle throttles it then blocks you. So I signed up for bing news api and pull the news articles once a day then insert them into a db and query the db when the user pulls up the page.
Can someone convert the new solution to javascript? I will update the gist afterwards, thank you!
When I call _/DotsSplashUi/data/batchexecute?rpcids=Fbv4je the answer does not include "garturlres". Anyone else having this problem? Was there already an update in July managing this?
Has anyone figured out how to play nice with the rate limits?
This js
version works for me (I've just converted mccabe-david python example). + I use a socks-proxy-agent package and https://webshare.io/ proxies to prevent 429 response.
import axios from 'axios';
import { SocksProxyAgent } from 'socks-proxy-agent';
async function getGoogleNewsRssFinalUrl(url){
const parsedUrl = new URL(url);
const gnArticleId = parsedUrl.pathname.split('/').pop();
const httpsAgent = new SocksProxyAgent(`socks5://user:pass@192.168.0.1:5987`);
const { data: gnData } = await axios.get(`https://news.google.com/articles/${gnArticleId}`, {
httpAgent: httpsAgent,
httpsAgent: httpsAgent,
});
const $ = cheerio.load(gnData);
const div = $('c-wiz > div').first();
const article = {
signature: div.attr('data-n-a-sg'),
timestamp: div.attr('data-n-a-ts'),
gn_art_id: gnArticleId,
};
const articlesReq = [
'Fbv4je',
`["garturlreq",[["X","X",["X","X"],null,null,1,1,"US:en",null,1,null,null,null,null,null,0,1],"X","X",1,[1,1,1],1,1,null,0,0,null,0],"${article.gn_art_id}",${article.timestamp},"${article.signature}"]`,
];
const response = await axios.post(
'https://news.google.com/_/DotsSplashUi/data/batchexecute',
new URLSearchParams({ 'f.req': JSON.stringify([[articlesReq]]) }).toString(),
{
headers: { 'Content-Type': 'application/x-www-form-urlencoded;charset=UTF-8' },
httpAgent: httpsAgent,
httpsAgent: httpsAgent,
},
);
return JSON.parse(JSON.parse(response.data.split('\n\n')[1].slice(0))[0][2])[1];
}
PYTHON SOLUTION
I don't know the exact algorithm used by Google to encode/decode the URLs, but I found a way to decode them using reverse engineering by inspecting the requests made by the browser in the redirection chain.
pip install beautifulsoup4 lxml
import json from urllib.parse import quote, urlparse import requests from bs4 import BeautifulSoup def get_decoding_params(gn_art_id): response = requests.get(f"https://news.google.com/articles/{gn_art_id}") response.raise_for_status() soup = BeautifulSoup(response.text, "lxml") div = soup.select_one("c-wiz > div") return { "signature": div.get("data-n-a-sg"), "timestamp": div.get("data-n-a-ts"), "gn_art_id": gn_art_id, } def decode_urls(articles): articles_reqs = [ [ "Fbv4je", f'["garturlreq",[["X","X",["X","X"],null,null,1,1,"US:en",null,1,null,null,null,null,null,0,1],"X","X",1,[1,1,1],1,1,null,0,0,null,0],"{art["gn_art_id"]}",{art["timestamp"]},"{art["signature"]}"]', ] for art in articles ] payload = f"f.req={quote(json.dumps([articles_reqs]))}" headers = {"content-type": "application/x-www-form-urlencoded;charset=UTF-8"} response = requests.post( url="https://news.google.com/_/DotsSplashUi/data/batchexecute", headers=headers, data=payload, ) response.raise_for_status() return [json.loads(res[2])[1] for res in json.loads(response.text.split("\n\n")[1])[:-2]] # Example usage encoded_urls = [ "https://news.google.com/rss/articles/CBMipgFBVV95cUxPWV9fTEI4cjh1RndwanpzNVliMUh6czg2X1RjeEN0YUctUmlZb0FyeV9oT3RWM1JrMGRodGtqTk1zV3pkNEpmdGNxc2lfd0c4LVpGVENvUDFMOEJqc0FCVVExSlRrQmI3TWZ2NUc4dy1EVXF4YnBLaGZ4cTFMQXFFM2JpanhDR3hoRmthUjVjdm1najZsaFh4a3lBbDladDZtVS1FMHFn?oc=5", "https://news.google.com/rss/articles/CBMi3AFBVV95cUxOX01TWDZZN2J5LWlmU3hudGZaRDh6a1dxUHMtalBEY1c0TlJSNlpieWxaUkxUU19MVTN3Y1BqaUZael83d1ctNXhaQUtPM0IyMFc4R3VydEtoMmFYMWpMU1Rtc3BjYmY4d3gxZHlMZG5NX0s1RmR2ZXI5YllvdzNSd2xkOFNCUTZTaEp3b0IxZEJZdVFLUDBNMC1wNGgwMGhjRG9HRFpRZU5BMFVIYjZCOWdWcHI1YzdoVHFWYnZSOEFwQ0NubGx3Rzd0SHN6OENKMXZUcHUxazA5WTIw?hl=en-US&gl=US&ceid=US%3Aen", ] articles_params = [get_decoding_params(urlparse(url).path.split("/")[-1]) for url in encoded_urls] decoded_urls = decode_urls(articles_params) print(decoded_urls)
Hi guys.
Remember that although I found a way to decode the URL with two requests, I also discovered that the second request can be done in batch to optimize latency, rate limits, and the cost of a potential proxy.
Downloading hundreds of pages that are around 6MB each is a bit too much, so I'm hoping to find a request that fetches the signature starting with ATR in future
Downloading hundreds of pages that are around 6MB each is a bit too much, so I'm hoping to find a request that fetches the signature starting with ATR in future
I was referring to the possibility of making batch request to https://news.google.com/_/DotsSplashUi/data/batchexecute
yea. this is starting to feel iffy to me. I'm just gonna use the news.google.com/rss/articles url without all the hackery. maybe there will be a way to decode the id like before again
@jacoboisaza I'm sorry, I didn't understand the code properly before. This is an awesome hack, thank you!
Hi, I have tried @jacoboisaza @SSujitX solutions. But i faced 429 Client Error: Too Many Requests for url. Anyone facing the same error? how to fix it?
def busted :(