SPC drift profile

The SpcDriftProfile serves as the core component for monitoring model drift in production.

Properties

Property	Type	Description	Example
scouter_version	str	The version of scouter that was used to create your SPC drift profile.	spc_profile.scouter_version → "1.0.0"
features	dict[str, SpcFeatureDriftProfile]	A mapping of feature names to their respective drift profiles.	spc_profile.features['feature_name'] → *Instance of SpcFeatureDriftProfile*
config	SpcDriftConfig	The drift config defined at the time of profile creationg.	spc_profile.config → *Instance of SpcDriftConfig*

Methods

model_dump_json()

Serializes the SpcDriftProfile instance to a JSON string.

• Parameters: None
• Returns: A JSON string representation of the instance.
• Return Type: str

---

model_dump()

Return dictionary representation of the drift profile.

• Parameters: None
• Returns: dict[str, Any] representation of the instance.
• Return Type: dict[str, Any]

---

save_to_json()

Save drift profile to json file.

• Parameters:
  • path (Optional[Path]): Optional path to save the drift profile. If None, outputs to drift_profile.json.
• Returns: None
• Return Type: None

---

update_config_args()

Inplace operation that updates config args.

• Parameters:
  • space (Optional[str]): Name of the model space.
  • name (Optional[str]): Name of the model.
  • version (Optional[str]): Version of the model.
  • targets (Optional[str]): Target(s) of the model / Dependant variable(s).
  • sample (Optional[bool]): Whether to use sampling or not.
  • sample_size (Optional[bool]): Size of the samples you want to use.
  • alert_config (Optional[SpcAlertConfig]): Instance of SpcAlertConfig
• Returns: None
• Return Type: None

---

model_validate_json() (static method)

Validate a JSON string representation of a saved profile.

• Parameters:
  • json_string (str): SpcDriftProfile in JSON string format.
• Returns: SpcDriftProfile
• Return Type: SpcDriftProfile

---

from_file() (static method)

Load a SpcDriftProfile from file.

• Parameters:
  • path (Path): Path to the file.
• Returns: SpcDriftProfile
• Return Type: SpcDriftProfile

---

model_validate() (static method)

Validate a dict representation of a saved profile.

• Parameters:
  • data (Dict[str, Any]): dict representation of your SpcDriftProfile.
• Returns: SpcDriftProfile
• Return Type: SpcDriftProfile

Features

SpcFeatureDriftProfile

The SpcFeatureDriftProfile is assigned to each feature when creating a SpcDriftProfile. The SpcFeatureDriftProfile will contain information about your features’s center and zone control limits.

Properties

Property	Type	Description	Example
id	str	The name of the feature.	profile.name → "feature_1"
center	float	Mean of the sample for the feature.	profile.center → .6
one_ucl	float	Zone 1 upper control limit.	profile.one_ucl → .1
one_lcl	float	Zone 1 lower control limit.	profile.one_lcl → .2
two_ucl	float	Zone 2 upper control limit.	profile.two_ucl → .3
two_lcl	float	Zone 2 lower control limit.	profile.two_lcl → .4
three_ucl	float	Zone 3 upper control limit.	profile.three_ucl → .5
three_lcl	float	Zone 3 lower control limit.	profile.three_lcl → .6
timestamp	str	Time of creation.	profile.timestamp → "2025-03-13T14:30:00Z"