fix: Fix LinkedIn get_user() and create_post() method. #397
Conversation
| @@ -97,10 +96,14 @@ async def create_post(client: RestliClient): | |||
| raise ValueError( | |||
| f"Error: failed to create post. Status code: {response.status_code}. Response: {response.entity}" | |||
| ) | |||
|
|
|||
| response.entity["link"] = ( | |||
| f"https://www.linkedin.com/feed/update/{response.entity['id']}" # craft the link manually | |||
There was a problem hiding this comment.
the link to post created
| response = client.get(resource_path="/userinfo", access_token=ACCESS_TOKEN) | ||
| def get_user(client: RestliClient) -> dict: | ||
|
|
||
| MAX_RETRIES = 3 |
There was a problem hiding this comment.
3 retries to prevent a weird 401 error from the Linkedin API side.
tybalex
requested review from
cjellick,
thedadams,
njhale,
sanjay920,
StrongMonkey,
iwilltry42 and
g-linville
| MAX_RETRIES = 3 | ||
| DELAY = 1 # seconds between retries | ||
|
|
||
| for attempt in range(1, MAX_RETRIES + 1): |
There was a problem hiding this comment.
It seems strange to me that we would need to retry this, Especially for a 401 error.
There was a problem hiding this comment.
Yeah, I’m with you on this. Here’s the issue in detail:
Sometimes, the first time the tool makes a get_user() request, it gets an error response like this:
{
"status": 401,
"serviceerrorcode": 65601,
"code": "revoked_access_token"
}
However, the second time it works just fine.
My theory: The OAuth process in the LinkedIn API server might be slow sometimes. After the access token is granted, the get_user() request might hit the server before the token is marked as valid in their system.
Another possible fix: I tried adding a time.sleep(2) at the beginning of the function, and it worked. But a retry mechanism would likely make more sense.
What does this PR do:
get_user()function.