Modellwerk
drift check
Compares two submitted sample sets feature by feature and returns deterministic drift metrics. Continuous features are binned on the combined value range and receive PSI, Jensen-Shannon divergence and a Kolmogorov-Smirnov statistic; categorical features receive PSI and Jensen-Shannon divergence over exact category labels. The response names every feature, its sample counts, the metric values, the threshold used, and the reason a feature was flagged.
POST https://modell.halowerk.com/v1/drift-checkEingaben
| Feld | Typ | Pflicht | Beschreibung |
|---|---|---|---|
| reference_samples | object | Ja | Baseline samples keyed by feature name. Each value is an array of finite numbers for continuous features or scalar category labels for categorical features. |
| current_samples | object | Ja | Current samples keyed by the same feature names as reference_samples. |
| threshold | number | Nein | Drift threshold applied to PSI for the final feature flag. |
| bins | integer | Nein | Number of equal-width bins for continuous features. |
| feature_types | object | Nein | Optional explicit feature types keyed by feature name. |
Vollständiges Eingabeschema anzeigen
{
"type": "object",
"required": [
"reference_samples",
"current_samples"
],
"additionalProperties": false,
"properties": {
"reference_samples": {
"type": "object",
"minProperties": 1,
"maxProperties": 50,
"description": "Baseline samples keyed by feature name. Each value is an array of finite numbers for continuous features or scalar category labels for categorical features.",
"additionalProperties": {
"type": "array",
"minItems": 2,
"maxItems": 10000,
"items": {
"type": [
"number",
"string",
"boolean"
]
}
}
},
"current_samples": {
"type": "object",
"minProperties": 1,
"maxProperties": 50,
"description": "Current samples keyed by the same feature names as reference_samples.",
"additionalProperties": {
"type": "array",
"minItems": 2,
"maxItems": 10000,
"items": {
"type": [
"number",
"string",
"boolean"
]
}
}
},
"threshold": {
"type": "number",
"minimum": 0.01,
"maximum": 1,
"default": 0.2,
"description": "Drift threshold applied to PSI for the final feature flag."
},
"bins": {
"type": "integer",
"minimum": 2,
"maximum": 50,
"default": 10,
"description": "Number of equal-width bins for continuous features."
},
"feature_types": {
"type": "object",
"maxProperties": 50,
"description": "Optional explicit feature types keyed by feature name.",
"additionalProperties": {
"type": "string",
"enum": [
"continuous",
"categorical"
]
}
}
}
}Veröffentlichtes Eingabebeispiel
{
"reference_samples": {
"risk_score": [
0.12,
0.18,
0.2,
0.21,
0.25,
0.28,
0.31,
0.34
],
"segment": [
"retail",
"retail",
"smb",
"smb",
"enterprise",
"retail"
]
},
"current_samples": {
"risk_score": [
0.14,
0.19,
0.23,
0.32,
0.4,
0.44,
0.48,
0.52
],
"segment": [
"retail",
"smb",
"smb",
"enterprise",
"enterprise",
"enterprise"
]
},
"threshold": 0.2,
"bins": 5,
"feature_types": {
"risk_score": "continuous",
"segment": "categorical"
}
}Ergebnis & Dokumentation
Veröffentlichtes Ergebnisschema anzeigen
{
"$ref": "https://modell.halowerk.com/v1/drift-check/schema#/outputSchema"
}Veröffentlichtes Ergebnisbeispiel öffnen
Maschinenlesbare Dokumentation dieses Hubs
Quelle dieser Seite: der veröffentlichte HALOWERK-Katalog. Stand: 2026-11-01, 15:19 UTC.
Weitere Endpunkte dieses Hubs
- Modellwerk0,003 USDC
chat completions
POSThttps://modell.halowerk.com/v1/chat-completionsRuns openai/gpt-oss-20b via GroqCloud. Messages are sent to Groq Inc., USA; payment metadata is not forwarded. At most 8 messages, 8192 UTF-8 bytes overall and 1024 completion tokens including reasoning. A length finish_reason exposes trunc
- Modellwerk0,002 USDC
context plan
POSThttps://modell.halowerk.com/v1/context-planPlans the division of a long input for a given model. It reserves room for the system prompt, the expected answer and a safety margin, then works out how many chunks are needed and where to cut, preferring paragraph boundaries over line and
- Modellwerk0,002 USDC
cost estimate
POSThttps://modell.halowerk.com/v1/cost-estimateWorks out what a call would cost. Give either the text itself or a token count, plus the expected output length, and get input, output, cache-read and cache-write cost per model with the total. Token counts derived from text are estimates f