@arizeai/phoenix-evals 0.5.0 → 0.6.0
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/README.md +42 -0
- package/dist/esm/core/EvaluatorBase.d.ts +19 -0
- package/dist/esm/core/EvaluatorBase.d.ts.map +1 -0
- package/dist/esm/core/EvaluatorBase.js +18 -0
- package/dist/esm/core/EvaluatorBase.js.map +1 -0
- package/dist/esm/index.d.ts +1 -0
- package/dist/esm/index.d.ts.map +1 -1
- package/dist/esm/index.js +1 -0
- package/dist/esm/index.js.map +1 -1
- package/dist/esm/llm/ClassificationEvaluator.d.ts +20 -3
- package/dist/esm/llm/ClassificationEvaluator.d.ts.map +1 -1
- package/dist/esm/llm/ClassificationEvaluator.js +29 -2
- package/dist/esm/llm/ClassificationEvaluator.js.map +1 -1
- package/dist/esm/llm/LLMEvaluator.d.ts +4 -7
- package/dist/esm/llm/LLMEvaluator.d.ts.map +1 -1
- package/dist/esm/llm/LLMEvaluator.js +4 -7
- package/dist/esm/llm/LLMEvaluator.js.map +1 -1
- package/dist/esm/llm/createClassificationEvaluator.d.ts +1 -1
- package/dist/esm/llm/createClassificationEvaluator.d.ts.map +1 -1
- package/dist/esm/llm/createClassificationEvaluator.js.map +1 -1
- package/dist/esm/llm/createDocumentRelevancyEvaluator.d.ts +8 -7
- package/dist/esm/llm/createDocumentRelevancyEvaluator.d.ts.map +1 -1
- package/dist/esm/llm/createDocumentRelevancyEvaluator.js +1 -2
- package/dist/esm/llm/createDocumentRelevancyEvaluator.js.map +1 -1
- package/dist/esm/llm/createHallucinationEvaluator.d.ts +6 -6
- package/dist/esm/llm/createHallucinationEvaluator.d.ts.map +1 -1
- package/dist/esm/llm/createHallucinationEvaluator.js +1 -2
- package/dist/esm/llm/createHallucinationEvaluator.js.map +1 -1
- package/dist/esm/template/applyTemplate.d.ts +3 -3
- package/dist/esm/template/applyTemplate.d.ts.map +1 -1
- package/dist/esm/template/applyTemplate.js +15 -1
- package/dist/esm/template/applyTemplate.js.map +1 -1
- package/dist/esm/template/createTemplateVariablesProxy.d.ts +1 -1
- package/dist/esm/template/createTemplateVariablesProxy.d.ts.map +1 -1
- package/dist/esm/template/createTemplateVariablesProxy.js.map +1 -1
- package/dist/esm/template/getTemplateVariables.d.ts +2 -2
- package/dist/esm/template/getTemplateVariables.d.ts.map +1 -1
- package/dist/esm/template/getTemplateVariables.js +16 -0
- package/dist/esm/template/getTemplateVariables.js.map +1 -1
- package/dist/esm/tsconfig.esm.tsbuildinfo +1 -1
- package/dist/esm/types/data.d.ts +194 -0
- package/dist/esm/types/data.d.ts.map +1 -0
- package/dist/esm/types/data.js +2 -0
- package/dist/esm/types/data.js.map +1 -0
- package/dist/esm/types/evals.d.ts +19 -5
- package/dist/esm/types/evals.d.ts.map +1 -1
- package/dist/esm/types/index.d.ts +1 -0
- package/dist/esm/types/index.d.ts.map +1 -1
- package/dist/esm/types/index.js +1 -0
- package/dist/esm/types/index.js.map +1 -1
- package/dist/esm/types/otel.d.ts +14 -13
- package/dist/esm/types/otel.d.ts.map +1 -1
- package/dist/esm/types/templating.d.ts +8 -6
- package/dist/esm/types/templating.d.ts.map +1 -1
- package/dist/esm/utils/bindEvaluator.d.ts +219 -0
- package/dist/esm/utils/bindEvaluator.d.ts.map +1 -0
- package/dist/esm/utils/bindEvaluator.js +163 -0
- package/dist/esm/utils/bindEvaluator.js.map +1 -0
- package/dist/esm/utils/index.d.ts +2 -0
- package/dist/esm/utils/index.d.ts.map +1 -0
- package/dist/esm/utils/index.js +2 -0
- package/dist/esm/utils/index.js.map +1 -0
- package/dist/esm/utils/objectMappingUtils.d.ts +166 -0
- package/dist/esm/utils/objectMappingUtils.d.ts.map +1 -0
- package/dist/esm/utils/objectMappingUtils.js +191 -0
- package/dist/esm/utils/objectMappingUtils.js.map +1 -0
- package/dist/src/core/EvaluatorBase.d.ts +19 -0
- package/dist/src/core/EvaluatorBase.d.ts.map +1 -0
- package/dist/src/core/EvaluatorBase.js +17 -0
- package/dist/src/core/EvaluatorBase.js.map +1 -0
- package/dist/src/index.d.ts +1 -0
- package/dist/src/index.d.ts.map +1 -1
- package/dist/src/index.js +1 -0
- package/dist/src/index.js.map +1 -1
- package/dist/src/llm/ClassificationEvaluator.d.ts +20 -3
- package/dist/src/llm/ClassificationEvaluator.d.ts.map +1 -1
- package/dist/src/llm/ClassificationEvaluator.js +13 -2
- package/dist/src/llm/ClassificationEvaluator.js.map +1 -1
- package/dist/src/llm/LLMEvaluator.d.ts +4 -7
- package/dist/src/llm/LLMEvaluator.d.ts.map +1 -1
- package/dist/src/llm/LLMEvaluator.js +16 -5
- package/dist/src/llm/LLMEvaluator.js.map +1 -1
- package/dist/src/llm/createClassificationEvaluator.d.ts +1 -1
- package/dist/src/llm/createClassificationEvaluator.d.ts.map +1 -1
- package/dist/src/llm/createClassificationEvaluator.js.map +1 -1
- package/dist/src/llm/createDocumentRelevancyEvaluator.d.ts +8 -7
- package/dist/src/llm/createDocumentRelevancyEvaluator.d.ts.map +1 -1
- package/dist/src/llm/createDocumentRelevancyEvaluator.js +2 -2
- package/dist/src/llm/createDocumentRelevancyEvaluator.js.map +1 -1
- package/dist/src/llm/createHallucinationEvaluator.d.ts +6 -6
- package/dist/src/llm/createHallucinationEvaluator.d.ts.map +1 -1
- package/dist/src/llm/createHallucinationEvaluator.js +2 -2
- package/dist/src/llm/createHallucinationEvaluator.js.map +1 -1
- package/dist/src/template/applyTemplate.d.ts +3 -3
- package/dist/src/template/applyTemplate.d.ts.map +1 -1
- package/dist/src/template/applyTemplate.js +12 -1
- package/dist/src/template/applyTemplate.js.map +1 -1
- package/dist/src/template/createTemplateVariablesProxy.d.ts +1 -1
- package/dist/src/template/createTemplateVariablesProxy.d.ts.map +1 -1
- package/dist/src/template/createTemplateVariablesProxy.js.map +1 -1
- package/dist/src/template/getTemplateVariables.d.ts +2 -2
- package/dist/src/template/getTemplateVariables.d.ts.map +1 -1
- package/dist/src/template/getTemplateVariables.js +16 -0
- package/dist/src/template/getTemplateVariables.js.map +1 -1
- package/dist/src/types/data.d.ts +194 -0
- package/dist/src/types/data.d.ts.map +1 -0
- package/dist/src/types/data.js +3 -0
- package/dist/src/types/data.js.map +1 -0
- package/dist/src/types/evals.d.ts +19 -5
- package/dist/src/types/evals.d.ts.map +1 -1
- package/dist/src/types/index.d.ts +1 -0
- package/dist/src/types/index.d.ts.map +1 -1
- package/dist/src/types/index.js +1 -0
- package/dist/src/types/index.js.map +1 -1
- package/dist/src/types/otel.d.ts +14 -13
- package/dist/src/types/otel.d.ts.map +1 -1
- package/dist/src/types/templating.d.ts +8 -6
- package/dist/src/types/templating.d.ts.map +1 -1
- package/dist/src/utils/bindEvaluator.d.ts +219 -0
- package/dist/src/utils/bindEvaluator.d.ts.map +1 -0
- package/dist/src/utils/bindEvaluator.js +166 -0
- package/dist/src/utils/bindEvaluator.js.map +1 -0
- package/dist/src/utils/index.d.ts +2 -0
- package/dist/src/utils/index.d.ts.map +1 -0
- package/dist/src/utils/index.js +18 -0
- package/dist/src/utils/index.js.map +1 -0
- package/dist/src/utils/objectMappingUtils.d.ts +166 -0
- package/dist/src/utils/objectMappingUtils.d.ts.map +1 -0
- package/dist/src/utils/objectMappingUtils.js +191 -0
- package/dist/src/utils/objectMappingUtils.js.map +1 -0
- package/dist/tsconfig.tsbuildinfo +1 -1
- package/package.json +3 -1
- package/src/core/EvaluatorBase.ts +43 -0
- package/src/index.ts +1 -0
- package/src/llm/ClassificationEvaluator.ts +41 -5
- package/src/llm/LLMEvaluator.ts +7 -16
- package/src/llm/createClassificationEvaluator.ts +1 -1
- package/src/llm/createDocumentRelevancyEvaluator.ts +17 -12
- package/src/llm/createHallucinationEvaluator.ts +12 -10
- package/src/template/applyTemplate.ts +22 -9
- package/src/template/createTemplateVariablesProxy.ts +3 -3
- package/src/template/getTemplateVariables.ts +18 -2
- package/src/types/data.ts +200 -0
- package/src/types/evals.ts +27 -7
- package/src/types/index.ts +1 -0
- package/src/types/otel.ts +15 -13
- package/src/types/templating.ts +9 -6
- package/src/utils/bindEvaluator.ts +229 -0
- package/src/utils/index.ts +1 -0
- package/src/utils/objectMappingUtils.ts +202 -0
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"evals.d.ts","sourceRoot":"","sources":["../../../src/types/evals.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,QAAQ,CAAC;AACvC,OAAO,EAAE,QAAQ,EAAE,MAAM,cAAc,CAAC;
|
|
1
|
+
{"version":3,"file":"evals.d.ts","sourceRoot":"","sources":["../../../src/types/evals.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,QAAQ,CAAC;AACvC,OAAO,EAAE,aAAa,EAAE,MAAM,QAAQ,CAAC;AACvC,OAAO,EAAE,cAAc,EAAE,MAAM,cAAc,CAAC;AAE9C,OAAO,EAAE,aAAa,EAAE,MAAM,IAAI,CAAC;AAEnC;;GAEG;AACH,MAAM,WAAW,aAAa,CAAC,UAAU,EAAE,SAAS;IAClD,MAAM,EAAE,UAAU,CAAC;IACnB,QAAQ,CAAC,EAAE,UAAU,CAAC;IACtB,KAAK,CAAC,EAAE,SAAS,CAAC;IAClB,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC;CACxB;AAED,MAAM,WAAW,OAAO;IACtB,KAAK,EAAE,aAAa,CAAC;CACtB;AAGD,MAAM,WAAW,iBAAkB,SAAQ,OAAO;CAAG;AAErD;;GAEG;AACH,MAAM,WAAW,gBAAgB;IAC/B;;;OAGG;IACH,KAAK,CAAC,EAAE,MAAM,CAAC;IACf;;;OAGG;IACH,KAAK,CAAC,EAAE,MAAM,CAAC;IACf;;;OAGG;IACH,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAED;;GAEG;AACH,MAAM,WAAW,oBAAoB;IACnC,KAAK,EAAE,MAAM,CAAC;IACd,WAAW,CAAC,EAAE,MAAM,CAAC;CACtB;AAED;;GAEG;AACH,MAAM,WAAW,oBAAoB;IACnC,KAAK,EAAE,MAAM,CAAC;IACd,KAAK,EAAE,MAAM,CAAC;CACf;AAED;;GAEG;AACH,MAAM,MAAM,wBAAwB,GAAG,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;AAE9D;;GAEG;AACH,MAAM,WAAW,oBAAqB,SAAQ,aAAa;IAIzD,KAAK,EAAE,aAAa,CAAC;IACrB;;;OAGG;IACH,OAAO,EAAE,wBAAwB,CAAC;IAClC;;OAEG;IACH,cAAc,EAAE,cAAc,CAAC;CAChC;AAED,MAAM,WAAW,mBAAmB,CAClC,WAAW,SAAS,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CACrE,SAAQ,aAAa;IACrB;;;OAGG;IACH,IAAI,EAAE,MAAM,CAAC;IACb;;OAEG;IACH,IAAI,EAAE,cAAc,CAAC;IACrB;;;OAGG;IACH,qBAAqB,CAAC,EAAE,qBAAqB,CAAC;IAC9C;;OAEG;IACH,YAAY,CAAC,EAAE,aAAa,CAAC,WAAW,CAAC,CAAC;CAC3C;AAED,MAAM,MAAM,sBAAsB,CAAC,UAAU,SAAS,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,IAC3E,IAAI,CAAC,mBAAmB,CAAC,UAAU,CAAC,EAAE,MAAM,CAAC,CAAC;AAEhD,MAAM,WAAW,iCAAiC,CAChD,UAAU,SAAS,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAC1C,SAAQ,oBAAoB,EAC1B,sBAAsB,CAAC,UAAU,CAAC;IACpC;;OAEG;IACH,cAAc,EAAE,cAAc,CAAC;CAChC;AAED,MAAM,MAAM,WAAW,CAAC,WAAW,SAAS,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,IAAI,CACrE,IAAI,EAAE,WAAW,KACd,OAAO,CAAC,gBAAgB,CAAC,CAAC;AAE/B;;GAEG;AACH,MAAM,MAAM,cAAc,GAAG,KAAK,GAAG,MAAM,CAAC;AAE5C;;;GAGG;AACH,MAAM,MAAM,qBAAqB,GAAG,UAAU,GAAG,UAAU,CAAC;AAE5D;;GAEG;AACH,UAAU,oBAAoB;IAC5B;;OAEG;IACH,IAAI,EAAE,MAAM,CAAC;IACb;;OAEG;IACH,IAAI,EAAE,cAAc,CAAC;IACrB;;;OAGG;IACH,qBAAqB,CAAC,EAAE,qBAAqB,CAAC;CAC/C;AAED;;;GAGG;AACH,MAAM,WAAW,kBAAkB,CAAC,WAAW,SAAS,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAC7E,SAAQ,oBAAoB;IAC5B;;OAEG;IACH,QAAQ,EAAE,WAAW,CAAC,WAAW,CAAC,CAAC;CACpC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/types/index.ts"],"names":[],"mappings":"AAAA,cAAc,SAAS,CAAC;AACxB,cAAc,cAAc,CAAC;AAC7B,cAAc,QAAQ,CAAC"}
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/types/index.ts"],"names":[],"mappings":"AAAA,cAAc,SAAS,CAAC;AACxB,cAAc,cAAc,CAAC;AAC7B,cAAc,QAAQ,CAAC;AACvB,cAAc,QAAQ,CAAC"}
|
package/dist/src/types/index.js
CHANGED
|
@@ -17,4 +17,5 @@ Object.defineProperty(exports, "__esModule", { value: true });
|
|
|
17
17
|
__exportStar(require("./evals"), exports);
|
|
18
18
|
__exportStar(require("./templating"), exports);
|
|
19
19
|
__exportStar(require("./otel"), exports);
|
|
20
|
+
__exportStar(require("./data"), exports);
|
|
20
21
|
//# sourceMappingURL=index.js.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/types/index.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;AAAA,0CAAwB;AACxB,+CAA6B;AAC7B,yCAAuB"}
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/types/index.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;AAAA,0CAAwB;AACxB,+CAA6B;AAC7B,yCAAuB;AACvB,yCAAuB"}
|
package/dist/src/types/otel.d.ts
CHANGED
|
@@ -1,17 +1,18 @@
|
|
|
1
1
|
import { Tracer } from "@opentelemetry/api";
|
|
2
|
+
export type TelemetryConfig = {
|
|
3
|
+
/**
|
|
4
|
+
* Whether OpenTelemetry is enabled on the call.
|
|
5
|
+
* Defaults to true for visibility into the evals calls.
|
|
6
|
+
* @default true
|
|
7
|
+
*/
|
|
8
|
+
isEnabled?: boolean;
|
|
9
|
+
/**
|
|
10
|
+
* The tracer to use for the call.
|
|
11
|
+
* If not provided, the traces will get picked up by the global tracer.
|
|
12
|
+
*/
|
|
13
|
+
tracer?: Tracer;
|
|
14
|
+
};
|
|
2
15
|
export type WithTelemetry = {
|
|
3
|
-
telemetry?:
|
|
4
|
-
/**
|
|
5
|
-
* Whether OpenTelemetry is enabled on the call.
|
|
6
|
-
* Defaults to true for visibility into the evals calls.
|
|
7
|
-
* @default true
|
|
8
|
-
*/
|
|
9
|
-
isEnabled?: boolean;
|
|
10
|
-
/**
|
|
11
|
-
* The tracer to use for the call.
|
|
12
|
-
* If not provided, the traces will get picked up by the global tracer.
|
|
13
|
-
*/
|
|
14
|
-
tracer?: Tracer;
|
|
15
|
-
};
|
|
16
|
+
telemetry?: TelemetryConfig;
|
|
16
17
|
};
|
|
17
18
|
//# sourceMappingURL=otel.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"otel.d.ts","sourceRoot":"","sources":["../../../src/types/otel.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,EAAE,MAAM,oBAAoB,CAAC;AAE5C,MAAM,MAAM,
|
|
1
|
+
{"version":3,"file":"otel.d.ts","sourceRoot":"","sources":["../../../src/types/otel.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,EAAE,MAAM,oBAAoB,CAAC;AAE5C,MAAM,MAAM,eAAe,GAAG;IAC5B;;;;OAIG;IACH,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB;;;OAGG;IACH,MAAM,CAAC,EAAE,MAAM,CAAC;CACjB,CAAC;AAEF,MAAM,MAAM,aAAa,GAAG;IAC1B,SAAS,CAAC,EAAE,eAAe,CAAC;CAC7B,CAAC"}
|
|
@@ -1,10 +1,12 @@
|
|
|
1
|
-
|
|
1
|
+
import type { ModelMessage } from "ai";
|
|
2
|
+
export type PromptTemplate = string | Array<ModelMessage>;
|
|
3
|
+
export type RenderedPrompt = string | Array<ModelMessage>;
|
|
2
4
|
export type TemplateVariables = Record<string, unknown>;
|
|
3
|
-
/**
|
|
4
|
-
* A class or object that has a prompt template
|
|
5
|
-
*/
|
|
6
5
|
export interface WithPromptTemplate {
|
|
7
|
-
readonly promptTemplate:
|
|
8
|
-
|
|
6
|
+
readonly promptTemplate: PromptTemplate;
|
|
7
|
+
/**
|
|
8
|
+
* List out the prompt template variables needed to perform evaluation
|
|
9
|
+
*/
|
|
10
|
+
readonly promptTemplateVariables: string[];
|
|
9
11
|
}
|
|
10
12
|
//# sourceMappingURL=templating.d.ts.map
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"templating.d.ts","sourceRoot":"","sources":["../../../src/types/templating.ts"],"names":[],"mappings":"AAAA,MAAM,MAAM,
|
|
1
|
+
{"version":3,"file":"templating.d.ts","sourceRoot":"","sources":["../../../src/types/templating.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,IAAI,CAAC;AAEvC,MAAM,MAAM,cAAc,GAAG,MAAM,GAAG,KAAK,CAAC,YAAY,CAAC,CAAC;AAC1D,MAAM,MAAM,cAAc,GAAG,MAAM,GAAG,KAAK,CAAC,YAAY,CAAC,CAAC;AAC1D,MAAM,MAAM,iBAAiB,GAAG,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;AAExD,MAAM,WAAW,kBAAkB;IACjC,QAAQ,CAAC,cAAc,EAAE,cAAc,CAAC;IACxC;;OAEG;IACH,QAAQ,CAAC,uBAAuB,EAAE,MAAM,EAAE,CAAC;CAC5C"}
|
|
@@ -0,0 +1,219 @@
|
|
|
1
|
+
import { EvaluatorBase } from "../core/EvaluatorBase";
|
|
2
|
+
import { ObjectMapping } from "../types/data";
|
|
3
|
+
/**
|
|
4
|
+
* Context for binding an evaluator with input mapping configuration.
|
|
5
|
+
*
|
|
6
|
+
* This type defines the structure for binding an evaluator to a specific data shape
|
|
7
|
+
* by mapping the evaluator's expected input fields to the actual data structure.
|
|
8
|
+
*
|
|
9
|
+
* @example
|
|
10
|
+
* ```typescript
|
|
11
|
+
* // Map evaluator fields to your data structure
|
|
12
|
+
* const context: BindingContext<MyDataType> = {
|
|
13
|
+
* inputMapping: {
|
|
14
|
+
* input: "userQuery", // Maps "input" to "userQuery" field
|
|
15
|
+
* reference: "context", // Maps "reference" to "context" field
|
|
16
|
+
* output: "modelResponse", // Maps "output" to "modelResponse" field
|
|
17
|
+
* },
|
|
18
|
+
* };
|
|
19
|
+
* ```
|
|
20
|
+
*
|
|
21
|
+
* @example
|
|
22
|
+
* ```typescript
|
|
23
|
+
* // Using nested property access
|
|
24
|
+
* const context: BindingContext<ApiResponse> = {
|
|
25
|
+
* inputMapping: {
|
|
26
|
+
* input: "request.body.query",
|
|
27
|
+
* reference: "request.body.context",
|
|
28
|
+
* output: "response.data.text",
|
|
29
|
+
* },
|
|
30
|
+
* };
|
|
31
|
+
* ```
|
|
32
|
+
*
|
|
33
|
+
* @example
|
|
34
|
+
* ```typescript
|
|
35
|
+
* // Using function-based mapping for transformations
|
|
36
|
+
* const context: BindingContext<RawData> = {
|
|
37
|
+
* inputMapping: {
|
|
38
|
+
* input: "question",
|
|
39
|
+
* reference: (data) => data.context.join("\n"), // Transform array to string
|
|
40
|
+
* output: "answer",
|
|
41
|
+
* },
|
|
42
|
+
* };
|
|
43
|
+
* ```
|
|
44
|
+
*
|
|
45
|
+
* @typeParam RecordType - The type of the data record that will be evaluated
|
|
46
|
+
*/
|
|
47
|
+
export type BindingContext<RecordType extends Record<string, unknown>> = {
|
|
48
|
+
/**
|
|
49
|
+
* Mapping of evaluator input fields to data source fields.
|
|
50
|
+
*
|
|
51
|
+
* The keys represent the field names expected by the evaluator (e.g., "input", "output", "reference"),
|
|
52
|
+
* and the values specify how to extract those fields from your data structure.
|
|
53
|
+
*
|
|
54
|
+
* Supports:
|
|
55
|
+
* - Simple property names: `"fieldName"`
|
|
56
|
+
* - Dot notation: `"user.profile.name"`
|
|
57
|
+
* - Array access: `"items[0].id"`
|
|
58
|
+
* - JSONPath expressions: `"$.items[*].id"`
|
|
59
|
+
* - Function extractors: `(data) => data.customField.toUpperCase()`
|
|
60
|
+
*/
|
|
61
|
+
inputMapping: ObjectMapping<RecordType>;
|
|
62
|
+
};
|
|
63
|
+
/**
|
|
64
|
+
* Binds an evaluator to a specific data structure using input mapping.
|
|
65
|
+
*
|
|
66
|
+
* This function creates a new evaluator instance that automatically transforms
|
|
67
|
+
* your data structure to match what the evaluator expects. This is particularly
|
|
68
|
+
* useful when your data schema doesn't match the evaluator's expected input format.
|
|
69
|
+
*
|
|
70
|
+
* @param evaluator - The evaluator to bind (e.g., a hallucination evaluator)
|
|
71
|
+
* @param context - The binding context containing the input mapping configuration
|
|
72
|
+
* @returns A new evaluator instance with the input mapping applied
|
|
73
|
+
*
|
|
74
|
+
* @example
|
|
75
|
+
* **Basic usage with simple field mapping:**
|
|
76
|
+
* ```typescript
|
|
77
|
+
* import { bindEvaluator, createHallucinationEvaluator } from "@arizeai/phoenix-evals";
|
|
78
|
+
* import { openai } from "@ai-sdk/openai";
|
|
79
|
+
*
|
|
80
|
+
* type MyData = {
|
|
81
|
+
* question: string;
|
|
82
|
+
* context: string;
|
|
83
|
+
* answer: string;
|
|
84
|
+
* };
|
|
85
|
+
*
|
|
86
|
+
* const evaluator = bindEvaluator<MyData>(
|
|
87
|
+
* createHallucinationEvaluator({ model: openai("gpt-4") }),
|
|
88
|
+
* {
|
|
89
|
+
* inputMapping: {
|
|
90
|
+
* input: "question", // Evaluator expects "input", map from "question"
|
|
91
|
+
* reference: "context", // Evaluator expects "reference", map from "context"
|
|
92
|
+
* output: "answer", // Evaluator expects "output", map from "answer"
|
|
93
|
+
* },
|
|
94
|
+
* }
|
|
95
|
+
* );
|
|
96
|
+
*
|
|
97
|
+
* // Now you can evaluate with your data structure
|
|
98
|
+
* const result = await evaluator.evaluate({
|
|
99
|
+
* question: "What is AI?",
|
|
100
|
+
* context: "AI is artificial intelligence...",
|
|
101
|
+
* answer: "AI stands for artificial intelligence",
|
|
102
|
+
* });
|
|
103
|
+
* ```
|
|
104
|
+
*
|
|
105
|
+
* @example
|
|
106
|
+
* **Using nested property access:**
|
|
107
|
+
* ```typescript
|
|
108
|
+
* type ApiResponse = {
|
|
109
|
+
* request: {
|
|
110
|
+
* body: {
|
|
111
|
+
* query: string;
|
|
112
|
+
* context: string;
|
|
113
|
+
* };
|
|
114
|
+
* };
|
|
115
|
+
* response: {
|
|
116
|
+
* data: {
|
|
117
|
+
* text: string;
|
|
118
|
+
* };
|
|
119
|
+
* };
|
|
120
|
+
* };
|
|
121
|
+
*
|
|
122
|
+
* const evaluator = bindEvaluator<ApiResponse>(
|
|
123
|
+
* createHallucinationEvaluator({ model: openai("gpt-4") }),
|
|
124
|
+
* {
|
|
125
|
+
* inputMapping: {
|
|
126
|
+
* input: "request.body.query",
|
|
127
|
+
* reference: "request.body.context",
|
|
128
|
+
* output: "response.data.text",
|
|
129
|
+
* },
|
|
130
|
+
* }
|
|
131
|
+
* );
|
|
132
|
+
* ```
|
|
133
|
+
*
|
|
134
|
+
* @example
|
|
135
|
+
* **Using function-based mapping for data transformation:**
|
|
136
|
+
* ```typescript
|
|
137
|
+
* type RawData = {
|
|
138
|
+
* question: string;
|
|
139
|
+
* contexts: string[]; // Array of context strings
|
|
140
|
+
* answer: string;
|
|
141
|
+
* };
|
|
142
|
+
*
|
|
143
|
+
* const evaluator = bindEvaluator<RawData>(
|
|
144
|
+
* createHallucinationEvaluator({ model: openai("gpt-4") }),
|
|
145
|
+
* {
|
|
146
|
+
* inputMapping: {
|
|
147
|
+
* input: "question",
|
|
148
|
+
* // Transform array to single string
|
|
149
|
+
* reference: (data) => data.contexts.join("\n\n"),
|
|
150
|
+
* output: "answer",
|
|
151
|
+
* },
|
|
152
|
+
* }
|
|
153
|
+
* );
|
|
154
|
+
* ```
|
|
155
|
+
*
|
|
156
|
+
* @example
|
|
157
|
+
* **Using JSONPath for complex queries:**
|
|
158
|
+
* ```typescript
|
|
159
|
+
* type ComplexData = {
|
|
160
|
+
* conversation: {
|
|
161
|
+
* messages: Array<{ role: string; content: string }>;
|
|
162
|
+
* };
|
|
163
|
+
* metadata: {
|
|
164
|
+
* sources: string[];
|
|
165
|
+
* };
|
|
166
|
+
* };
|
|
167
|
+
*
|
|
168
|
+
* const evaluator = bindEvaluator<ComplexData>(
|
|
169
|
+
* createHallucinationEvaluator({ model: openai("gpt-4") }),
|
|
170
|
+
* {
|
|
171
|
+
* inputMapping: {
|
|
172
|
+
* // Extract last user message
|
|
173
|
+
* input: "$.conversation.messages[?(@.role=='user')].content[-1]",
|
|
174
|
+
* // Extract all sources
|
|
175
|
+
* reference: "$.metadata.sources[*]",
|
|
176
|
+
* // Extract last assistant message
|
|
177
|
+
* output: "$.conversation.messages[?(@.role=='assistant')].content[-1]",
|
|
178
|
+
* },
|
|
179
|
+
* }
|
|
180
|
+
* );
|
|
181
|
+
* ```
|
|
182
|
+
*
|
|
183
|
+
* @example
|
|
184
|
+
* **Binding multiple evaluators with different mappings:**
|
|
185
|
+
* ```typescript
|
|
186
|
+
* type EvaluationData = {
|
|
187
|
+
* userQuery: string;
|
|
188
|
+
* systemContext: string;
|
|
189
|
+
* modelOutput: string;
|
|
190
|
+
* expectedOutput?: string;
|
|
191
|
+
* };
|
|
192
|
+
*
|
|
193
|
+
* // Hallucination evaluator
|
|
194
|
+
* const hallucinationEvaluator = bindEvaluator<EvaluationData>(
|
|
195
|
+
* createHallucinationEvaluator({ model: openai("gpt-4") }),
|
|
196
|
+
* {
|
|
197
|
+
* inputMapping: {
|
|
198
|
+
* input: "userQuery",
|
|
199
|
+
* reference: "systemContext",
|
|
200
|
+
* output: "modelOutput",
|
|
201
|
+
* },
|
|
202
|
+
* }
|
|
203
|
+
* );
|
|
204
|
+
*
|
|
205
|
+
* // Document relevancy evaluator (if it exists)
|
|
206
|
+
* const relevancyEvaluator = bindEvaluator<EvaluationData>(
|
|
207
|
+
* createDocumentRelevancyEvaluator({ model: openai("gpt-4") }),
|
|
208
|
+
* {
|
|
209
|
+
* inputMapping: {
|
|
210
|
+
* query: "userQuery",
|
|
211
|
+
* document: "systemContext",
|
|
212
|
+
* output: "modelOutput",
|
|
213
|
+
* },
|
|
214
|
+
* }
|
|
215
|
+
* );
|
|
216
|
+
* ```
|
|
217
|
+
*/
|
|
218
|
+
export declare function bindEvaluator<RecordType extends Record<string, unknown>>(evaluator: EvaluatorBase<RecordType>, context: BindingContext<RecordType>): EvaluatorBase<RecordType>;
|
|
219
|
+
//# sourceMappingURL=bindEvaluator.d.ts.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"bindEvaluator.d.ts","sourceRoot":"","sources":["../../../src/utils/bindEvaluator.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,aAAa,EAAE,MAAM,uBAAuB,CAAC;AACtD,OAAO,EAAE,aAAa,EAAE,MAAM,eAAe,CAAC;AAE9C;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA2CG;AACH,MAAM,MAAM,cAAc,CAAC,UAAU,SAAS,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,IAAI;IACvE;;;;;;;;;;;;OAYG;IACH,YAAY,EAAE,aAAa,CAAC,UAAU,CAAC,CAAC;CACzC,CAAC;AAEF;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA0JG;AACH,wBAAgB,aAAa,CAAC,UAAU,SAAS,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,EACtE,SAAS,EAAE,aAAa,CAAC,UAAU,CAAC,EACpC,OAAO,EAAE,cAAc,CAAC,UAAU,CAAC,GAClC,aAAa,CAAC,UAAU,CAAC,CAM3B"}
|
|
@@ -0,0 +1,166 @@
|
|
|
1
|
+
"use strict";
|
|
2
|
+
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
+
exports.bindEvaluator = bindEvaluator;
|
|
4
|
+
/**
|
|
5
|
+
* Binds an evaluator to a specific data structure using input mapping.
|
|
6
|
+
*
|
|
7
|
+
* This function creates a new evaluator instance that automatically transforms
|
|
8
|
+
* your data structure to match what the evaluator expects. This is particularly
|
|
9
|
+
* useful when your data schema doesn't match the evaluator's expected input format.
|
|
10
|
+
*
|
|
11
|
+
* @param evaluator - The evaluator to bind (e.g., a hallucination evaluator)
|
|
12
|
+
* @param context - The binding context containing the input mapping configuration
|
|
13
|
+
* @returns A new evaluator instance with the input mapping applied
|
|
14
|
+
*
|
|
15
|
+
* @example
|
|
16
|
+
* **Basic usage with simple field mapping:**
|
|
17
|
+
* ```typescript
|
|
18
|
+
* import { bindEvaluator, createHallucinationEvaluator } from "@arizeai/phoenix-evals";
|
|
19
|
+
* import { openai } from "@ai-sdk/openai";
|
|
20
|
+
*
|
|
21
|
+
* type MyData = {
|
|
22
|
+
* question: string;
|
|
23
|
+
* context: string;
|
|
24
|
+
* answer: string;
|
|
25
|
+
* };
|
|
26
|
+
*
|
|
27
|
+
* const evaluator = bindEvaluator<MyData>(
|
|
28
|
+
* createHallucinationEvaluator({ model: openai("gpt-4") }),
|
|
29
|
+
* {
|
|
30
|
+
* inputMapping: {
|
|
31
|
+
* input: "question", // Evaluator expects "input", map from "question"
|
|
32
|
+
* reference: "context", // Evaluator expects "reference", map from "context"
|
|
33
|
+
* output: "answer", // Evaluator expects "output", map from "answer"
|
|
34
|
+
* },
|
|
35
|
+
* }
|
|
36
|
+
* );
|
|
37
|
+
*
|
|
38
|
+
* // Now you can evaluate with your data structure
|
|
39
|
+
* const result = await evaluator.evaluate({
|
|
40
|
+
* question: "What is AI?",
|
|
41
|
+
* context: "AI is artificial intelligence...",
|
|
42
|
+
* answer: "AI stands for artificial intelligence",
|
|
43
|
+
* });
|
|
44
|
+
* ```
|
|
45
|
+
*
|
|
46
|
+
* @example
|
|
47
|
+
* **Using nested property access:**
|
|
48
|
+
* ```typescript
|
|
49
|
+
* type ApiResponse = {
|
|
50
|
+
* request: {
|
|
51
|
+
* body: {
|
|
52
|
+
* query: string;
|
|
53
|
+
* context: string;
|
|
54
|
+
* };
|
|
55
|
+
* };
|
|
56
|
+
* response: {
|
|
57
|
+
* data: {
|
|
58
|
+
* text: string;
|
|
59
|
+
* };
|
|
60
|
+
* };
|
|
61
|
+
* };
|
|
62
|
+
*
|
|
63
|
+
* const evaluator = bindEvaluator<ApiResponse>(
|
|
64
|
+
* createHallucinationEvaluator({ model: openai("gpt-4") }),
|
|
65
|
+
* {
|
|
66
|
+
* inputMapping: {
|
|
67
|
+
* input: "request.body.query",
|
|
68
|
+
* reference: "request.body.context",
|
|
69
|
+
* output: "response.data.text",
|
|
70
|
+
* },
|
|
71
|
+
* }
|
|
72
|
+
* );
|
|
73
|
+
* ```
|
|
74
|
+
*
|
|
75
|
+
* @example
|
|
76
|
+
* **Using function-based mapping for data transformation:**
|
|
77
|
+
* ```typescript
|
|
78
|
+
* type RawData = {
|
|
79
|
+
* question: string;
|
|
80
|
+
* contexts: string[]; // Array of context strings
|
|
81
|
+
* answer: string;
|
|
82
|
+
* };
|
|
83
|
+
*
|
|
84
|
+
* const evaluator = bindEvaluator<RawData>(
|
|
85
|
+
* createHallucinationEvaluator({ model: openai("gpt-4") }),
|
|
86
|
+
* {
|
|
87
|
+
* inputMapping: {
|
|
88
|
+
* input: "question",
|
|
89
|
+
* // Transform array to single string
|
|
90
|
+
* reference: (data) => data.contexts.join("\n\n"),
|
|
91
|
+
* output: "answer",
|
|
92
|
+
* },
|
|
93
|
+
* }
|
|
94
|
+
* );
|
|
95
|
+
* ```
|
|
96
|
+
*
|
|
97
|
+
* @example
|
|
98
|
+
* **Using JSONPath for complex queries:**
|
|
99
|
+
* ```typescript
|
|
100
|
+
* type ComplexData = {
|
|
101
|
+
* conversation: {
|
|
102
|
+
* messages: Array<{ role: string; content: string }>;
|
|
103
|
+
* };
|
|
104
|
+
* metadata: {
|
|
105
|
+
* sources: string[];
|
|
106
|
+
* };
|
|
107
|
+
* };
|
|
108
|
+
*
|
|
109
|
+
* const evaluator = bindEvaluator<ComplexData>(
|
|
110
|
+
* createHallucinationEvaluator({ model: openai("gpt-4") }),
|
|
111
|
+
* {
|
|
112
|
+
* inputMapping: {
|
|
113
|
+
* // Extract last user message
|
|
114
|
+
* input: "$.conversation.messages[?(@.role=='user')].content[-1]",
|
|
115
|
+
* // Extract all sources
|
|
116
|
+
* reference: "$.metadata.sources[*]",
|
|
117
|
+
* // Extract last assistant message
|
|
118
|
+
* output: "$.conversation.messages[?(@.role=='assistant')].content[-1]",
|
|
119
|
+
* },
|
|
120
|
+
* }
|
|
121
|
+
* );
|
|
122
|
+
* ```
|
|
123
|
+
*
|
|
124
|
+
* @example
|
|
125
|
+
* **Binding multiple evaluators with different mappings:**
|
|
126
|
+
* ```typescript
|
|
127
|
+
* type EvaluationData = {
|
|
128
|
+
* userQuery: string;
|
|
129
|
+
* systemContext: string;
|
|
130
|
+
* modelOutput: string;
|
|
131
|
+
* expectedOutput?: string;
|
|
132
|
+
* };
|
|
133
|
+
*
|
|
134
|
+
* // Hallucination evaluator
|
|
135
|
+
* const hallucinationEvaluator = bindEvaluator<EvaluationData>(
|
|
136
|
+
* createHallucinationEvaluator({ model: openai("gpt-4") }),
|
|
137
|
+
* {
|
|
138
|
+
* inputMapping: {
|
|
139
|
+
* input: "userQuery",
|
|
140
|
+
* reference: "systemContext",
|
|
141
|
+
* output: "modelOutput",
|
|
142
|
+
* },
|
|
143
|
+
* }
|
|
144
|
+
* );
|
|
145
|
+
*
|
|
146
|
+
* // Document relevancy evaluator (if it exists)
|
|
147
|
+
* const relevancyEvaluator = bindEvaluator<EvaluationData>(
|
|
148
|
+
* createDocumentRelevancyEvaluator({ model: openai("gpt-4") }),
|
|
149
|
+
* {
|
|
150
|
+
* inputMapping: {
|
|
151
|
+
* query: "userQuery",
|
|
152
|
+
* document: "systemContext",
|
|
153
|
+
* output: "modelOutput",
|
|
154
|
+
* },
|
|
155
|
+
* }
|
|
156
|
+
* );
|
|
157
|
+
* ```
|
|
158
|
+
*/
|
|
159
|
+
function bindEvaluator(evaluator, context) {
|
|
160
|
+
let boundEvaluator = evaluator;
|
|
161
|
+
if (context.inputMapping) {
|
|
162
|
+
boundEvaluator = boundEvaluator.bindInputMapping(context.inputMapping);
|
|
163
|
+
}
|
|
164
|
+
return boundEvaluator;
|
|
165
|
+
}
|
|
166
|
+
//# sourceMappingURL=bindEvaluator.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"bindEvaluator.js","sourceRoot":"","sources":["../../../src/utils/bindEvaluator.ts"],"names":[],"mappings":";;AA2NA,sCASC;AApKD;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;GA0JG;AACH,SAAgB,aAAa,CAC3B,SAAoC,EACpC,OAAmC;IAEnC,IAAI,cAAc,GAA8B,SAAS,CAAC;IAC1D,IAAI,OAAO,CAAC,YAAY,EAAE,CAAC;QACzB,cAAc,GAAG,cAAc,CAAC,gBAAgB,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC;IACzE,CAAC;IACD,OAAO,cAAc,CAAC;AACxB,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../../src/utils/index.ts"],"names":[],"mappings":"AAAA,cAAc,iBAAiB,CAAC"}
|
|
@@ -0,0 +1,18 @@
|
|
|
1
|
+
"use strict";
|
|
2
|
+
var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
|
|
3
|
+
if (k2 === undefined) k2 = k;
|
|
4
|
+
var desc = Object.getOwnPropertyDescriptor(m, k);
|
|
5
|
+
if (!desc || ("get" in desc ? !m.__esModule : desc.writable || desc.configurable)) {
|
|
6
|
+
desc = { enumerable: true, get: function() { return m[k]; } };
|
|
7
|
+
}
|
|
8
|
+
Object.defineProperty(o, k2, desc);
|
|
9
|
+
}) : (function(o, m, k, k2) {
|
|
10
|
+
if (k2 === undefined) k2 = k;
|
|
11
|
+
o[k2] = m[k];
|
|
12
|
+
}));
|
|
13
|
+
var __exportStar = (this && this.__exportStar) || function(m, exports) {
|
|
14
|
+
for (var p in m) if (p !== "default" && !Object.prototype.hasOwnProperty.call(exports, p)) __createBinding(exports, m, p);
|
|
15
|
+
};
|
|
16
|
+
Object.defineProperty(exports, "__esModule", { value: true });
|
|
17
|
+
__exportStar(require("./bindEvaluator"), exports);
|
|
18
|
+
//# sourceMappingURL=index.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/utils/index.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;AAAA,kDAAgC"}
|