Add timeout to requests calls

[pixeeai]

Many developers will be surprised to learn that requests library calls do not include timeouts by default. This means that an attempted request could hang indefinitely if no connection is established or if no data is received from the server.

The requests documentation suggests that most calls should explicitly include a timeout parameter. This codemod adds a default timeout value in order to set an upper bound on connection times and ensure that requests connect or fail in a timely manner. This value also ensures the connection will timeout if the server does not respond with data within a reasonable amount of time.

While timeout values will be application dependent, we believe that this codemod adds a reasonable default that serves as an appropriate ceiling for most situations.

Our changes look like the following:

 import requests
 
- requests.get("http://example.com")
+ requests.get("http://example.com", timeout=60)

More reading

• https://docs.python-requests.org/en/master/user/quickstart/#timeouts

🧚🤖 Powered by Pixeebot

Feedback | Community | Docs | Codemod ID: pixee:python/add-requests-timeouts

[zainhoda]

This is a great idea. However, the timeouts needs to vary based on endpoint. Some endpoints that call LLMs will be expected to take quite a long time (potentially up to 5 minutes in cases like the VLLM calls, which call a locally-running LLM).

[drdavella]

Hi @zainhoda I'm a Pixeebot maintainer. Thank you for the feedback, we appreciate it very much.

You are correct that the timeout will be dependent on the application and endpoint being used. Our intent with this change is to provide a reasonable starting point that developers can then modify as necessary. Eventually we'd like to make this value configurable and/or provide an interactive way for users to update the value provided.