Open
Description
Description:
I'm following the Safe and reliable AI via guardrails course on DeepLearning.AI and encountered an issue when trying to start the Guardrails server on Windows 10.
Environment:
- OS: Windows 10
- Python version: 3.10.16
- Guardrails version:
guardrails-ai==0.6.6
- Installed via:
pip
using arequirements.txt
Steps to Reproduce:
-
Created a clean virtual environment
-
Installed packages using
requirements.txt
(successfully):bs4 guardrails-ai==0.6.6 numpy pandas matplotlib openai ipykernel ipywidgets
-
Ran (successfully):
guardrails configure
-
Installed hub guardrails (successfully), as outlined in the "Instructions to install Guardrails server" section of Lesson 3's notebook in the course:
guardrails hub install --no-install-local-models hub://guardrails/provenance_llm; guardrails hub install hub://guardrails/detect_pii; guardrails hub install --no-install-local-models hub://tryolabs/restricttotopic; guardrails hub install --no-install-local-models hub://guardrails/competitor_check;
-
Started the server (resulted in an error):
guardrails start --config config.py
Observed Error:
Failed to install guardrails-api>="^0.0.0a0"
Exit code: 1
stderr: ERROR: Invalid requirement: 'guardrails-api>="^0.0.0a0"': Expected end or semicolon (after name and no valid version specifier)
guardrails-api>="^0.0.0a0"
^
stdout:
Question:
- Is the
guardrails-api
package supposed to be installed manually? - It's not mentioned in the documentation or official website. How should I proceed?
- I'd like to follow the official instructions as closely as possible to avoid conflicts. Any advice?