bug: Namespace error message is very confusing, the regex in the error message is not the regex that validates the namespace name

[ogenstad]

Component

infrahubctl

Infrahub SDK version

1.14.0

Current Behavior

This was originally reported in opsmill/infrahub#7265, as it's a bug related to the SDK it has been reopened here instead.

Address:            http://localhost:8000 
 Connection Status:  ✅                    
 Python Version:     3.12.2                
 SDK Version:        1.14.0                
 Infrahub Version:   1.4.8                 
 User:               Admin           

Issue:

version: "1.0"
nodes:
  - name: Device
    namespace: NGC 
    human_friendly_id: ['hostname__value']
    attributes:
      - name: hostname
        kind: Text
        unique: true
      - name: model
        kind: Text

According to the REGEX in the error message below, the namespace name above should be valid.

In the code, the regex that is actually used to validate the namespace is different from the one in the error message. The documentation (that I could find) does not describe what a valid schema name looks like.

This schema name is invalid 
/infrahubctl schema load --branch demo2 schemas/test.yml 
Unable to load the schema:
  Node: NGCDEVICE | namespace (NGC) | String should match pattern '^[A-Z]+$' (string_pattern_mismatch)

Valid namespace names are currently are '^[A-Z][a-z]+$'

1. Block caps should be allowed for namespace names.

2. At a minimum, the error message should display the actual regex used to validate the namespace name, not an incompatible regex that will never be a valid schema name.

Expected Behavior

The regex in the error message should be the one used to validate the namespace name, not a random incompatible one.

Steps to Reproduce

See schema above and try to import it.

Additional Information

This seems to be coming from the rich library when we print out the error here: https://github.com/opsmill/infrahub-sdk-python/blob/v1.14.0/infrahub_sdk/ctl/schema.py#L76

Looks like rich is treating '^[A-Z][a-z0-9]+$' as invalid markup and strips out some parts.

The correct way would be to instead have:

console.print(f"  Node: {node.get('namespace', None)}{node.get('name', None)} | {error_message}", markup=False)