@wix/evalforge-evaluator 0.132.0 → 0.133.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/build/index.js +96 -57
- package/build/index.js.map +4 -4
- package/build/index.mjs +96 -57
- package/build/index.mjs.map +4 -4
- package/build/types/evaluation-loop.d.ts +23 -0
- package/package.json +5 -5
|
@@ -0,0 +1,23 @@
|
|
|
1
|
+
import type { EvalRunResult, TestScenario, Template, Assertion } from '@wix/evalforge-types';
|
|
2
|
+
import type { ScenarioItem, EvaluationData } from './fetch-evaluation-data.js';
|
|
3
|
+
export interface EvaluationLoopCallbacks {
|
|
4
|
+
runScenario: (scenario: TestScenario, template: Template | undefined, resolvedAssertions: Assertion[] | undefined) => Promise<EvalRunResult>;
|
|
5
|
+
addResult: (result: EvalRunResult) => Promise<void>;
|
|
6
|
+
onProgress: (completed: number, total: number, iterLabel: string) => void;
|
|
7
|
+
}
|
|
8
|
+
export interface EvaluationLoopResult {
|
|
9
|
+
completedExecutions: number;
|
|
10
|
+
totalExecutions: number;
|
|
11
|
+
}
|
|
12
|
+
/**
|
|
13
|
+
* Run all scenarios with iteration support.
|
|
14
|
+
*
|
|
15
|
+
* Per-iteration execution errors are caught and recorded as failed results
|
|
16
|
+
* (passRate: 0, outputText containing the error) rather than aborting the run.
|
|
17
|
+
* This enables flakiness detection: all N iterations of every scenario complete
|
|
18
|
+
* even when some iterations fail transiently.
|
|
19
|
+
*
|
|
20
|
+
* Infrastructure errors (e.g. failure to persist a result) are not caught and
|
|
21
|
+
* will propagate up to abort the run.
|
|
22
|
+
*/
|
|
23
|
+
export declare function runEvaluationLoop(scenarioItems: ScenarioItem[], evalData: EvaluationData, callbacks: EvaluationLoopCallbacks): Promise<EvaluationLoopResult>;
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@wix/evalforge-evaluator",
|
|
3
|
-
"version": "0.
|
|
3
|
+
"version": "0.133.0",
|
|
4
4
|
"description": "EvalForge Evaluator",
|
|
5
5
|
"bin": "./build/index.js",
|
|
6
6
|
"files": [
|
|
@@ -21,9 +21,9 @@
|
|
|
21
21
|
"@ai-sdk/openai": "^3.0.39",
|
|
22
22
|
"@anthropic-ai/claude-agent-sdk": "^0.2.63",
|
|
23
23
|
"@anthropic-ai/claude-code": "^2.1.63",
|
|
24
|
-
"@wix/eval-assertions": "0.
|
|
25
|
-
"@wix/evalforge-github-client": "0.
|
|
26
|
-
"@wix/evalforge-types": "0.
|
|
24
|
+
"@wix/eval-assertions": "0.38.0",
|
|
25
|
+
"@wix/evalforge-github-client": "0.42.0",
|
|
26
|
+
"@wix/evalforge-types": "0.67.0",
|
|
27
27
|
"ai": "^6.0.107",
|
|
28
28
|
"diff": "^7.0.0",
|
|
29
29
|
"tar": "^7.5.3",
|
|
@@ -62,5 +62,5 @@
|
|
|
62
62
|
"artifactId": "evalforge-evaluator"
|
|
63
63
|
}
|
|
64
64
|
},
|
|
65
|
-
"falconPackageHash": "
|
|
65
|
+
"falconPackageHash": "fffe0a562e627362231ecc39fbe237c7a33796f807b0d518734592fd"
|
|
66
66
|
}
|