Eval 結果クエリ

Eval 結果クエリ

curl --request POST \
  --url https://api.example.com/v2/{entity}/{project}/eval_results/query \
  --header 'Content-Type: application/json' \
  --data '
{
  "evaluation_call_ids": [
    "<string>"
  ],
  "evaluation_run_ids": [
    "<string>"
  ],
  "filter_logic_operator": "or",
  "filters": [
    {
      "query": {
        "$expr": {
          "$and": [
            {
              "$literal": "<string>"
            }
          ]
        }
      },
      "evaluation_call_id": "<string>"
    }
  ],
  "include_costs": false,
  "include_predict_and_score_children": true,
  "include_raw_data_rows": false,
  "include_rows": true,
  "include_summary": false,
  "limit": 123,
  "offset": 0,
  "require_intersection": false,
  "resolve_row_refs": false,
  "sort_by": [
    {
      "field": "<string>",
      "evaluation_call_id": "<string>",
      "mode": "value"
    }
  ],
  "summary_require_intersection": true
}
'

import requests

url = "https://api.example.com/v2/{entity}/{project}/eval_results/query"

payload = {
    "evaluation_call_ids": ["<string>"],
    "evaluation_run_ids": ["<string>"],
    "filter_logic_operator": "or",
    "filters": [
        {
            "query": { "$expr": { "$and": [{ "$literal": "<string>" }] } },
            "evaluation_call_id": "<string>"
        }
    ],
    "include_costs": False,
    "include_predict_and_score_children": True,
    "include_raw_data_rows": False,
    "include_rows": True,
    "include_summary": False,
    "limit": 123,
    "offset": 0,
    "require_intersection": False,
    "resolve_row_refs": False,
    "sort_by": [
        {
            "field": "<string>",
            "evaluation_call_id": "<string>",
            "mode": "value"
        }
    ],
    "summary_require_intersection": True
}
headers = {"Content-Type": "application/json"}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {'Content-Type': 'application/json'},
  body: JSON.stringify({
    evaluation_call_ids: ['<string>'],
    evaluation_run_ids: ['<string>'],
    filter_logic_operator: 'or',
    filters: [
      {
        query: {$expr: {$and: [{$literal: '<string>'}]}},
        evaluation_call_id: '<string>'
      }
    ],
    include_costs: false,
    include_predict_and_score_children: true,
    include_raw_data_rows: false,
    include_rows: true,
    include_summary: false,
    limit: 123,
    offset: 0,
    require_intersection: false,
    resolve_row_refs: false,
    sort_by: [{field: '<string>', evaluation_call_id: '<string>', mode: 'value'}],
    summary_require_intersection: true
  })
};

fetch('https://api.example.com/v2/{entity}/{project}/eval_results/query', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://api.example.com/v2/{entity}/{project}/eval_results/query",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "POST",
  CURLOPT_POSTFIELDS => json_encode([
    'evaluation_call_ids' => [
        '<string>'
    ],
    'evaluation_run_ids' => [
        '<string>'
    ],
    'filter_logic_operator' => 'or',
    'filters' => [
        [
                'query' => [
                                '$expr' => [
                                                                '$and' => [
                                                                                                                                [
                                                                                                                                                                                                                                                                '$literal' => '<string>'
                                                                                                                                ]
                                                                ]
                                ]
                ],
                'evaluation_call_id' => '<string>'
        ]
    ],
    'include_costs' => false,
    'include_predict_and_score_children' => true,
    'include_raw_data_rows' => false,
    'include_rows' => true,
    'include_summary' => false,
    'limit' => 123,
    'offset' => 0,
    'require_intersection' => false,
    'resolve_row_refs' => false,
    'sort_by' => [
        [
                'field' => '<string>',
                'evaluation_call_id' => '<string>',
                'mode' => 'value'
        ]
    ],
    'summary_require_intersection' => true
  ]),
  CURLOPT_HTTPHEADER => [
    "Content-Type: application/json"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.example.com/v2/{entity}/{project}/eval_results/query"

	payload := strings.NewReader("{\n  \"evaluation_call_ids\": [\n    \"<string>\"\n  ],\n  \"evaluation_run_ids\": [\n    \"<string>\"\n  ],\n  \"filter_logic_operator\": \"or\",\n  \"filters\": [\n    {\n      \"query\": {\n        \"$expr\": {\n          \"$and\": [\n            {\n              \"$literal\": \"<string>\"\n            }\n          ]\n        }\n      },\n      \"evaluation_call_id\": \"<string>\"\n    }\n  ],\n  \"include_costs\": false,\n  \"include_predict_and_score_children\": true,\n  \"include_raw_data_rows\": false,\n  \"include_rows\": true,\n  \"include_summary\": false,\n  \"limit\": 123,\n  \"offset\": 0,\n  \"require_intersection\": false,\n  \"resolve_row_refs\": false,\n  \"sort_by\": [\n    {\n      \"field\": \"<string>\",\n      \"evaluation_call_id\": \"<string>\",\n      \"mode\": \"value\"\n    }\n  ],\n  \"summary_require_intersection\": true\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.example.com/v2/{entity}/{project}/eval_results/query")
  .header("Content-Type", "application/json")
  .body("{\n  \"evaluation_call_ids\": [\n    \"<string>\"\n  ],\n  \"evaluation_run_ids\": [\n    \"<string>\"\n  ],\n  \"filter_logic_operator\": \"or\",\n  \"filters\": [\n    {\n      \"query\": {\n        \"$expr\": {\n          \"$and\": [\n            {\n              \"$literal\": \"<string>\"\n            }\n          ]\n        }\n      },\n      \"evaluation_call_id\": \"<string>\"\n    }\n  ],\n  \"include_costs\": false,\n  \"include_predict_and_score_children\": true,\n  \"include_raw_data_rows\": false,\n  \"include_rows\": true,\n  \"include_summary\": false,\n  \"limit\": 123,\n  \"offset\": 0,\n  \"require_intersection\": false,\n  \"resolve_row_refs\": false,\n  \"sort_by\": [\n    {\n      \"field\": \"<string>\",\n      \"evaluation_call_id\": \"<string>\",\n      \"mode\": \"value\"\n    }\n  ],\n  \"summary_require_intersection\": true\n}")
  .asString();

require 'uri'
require 'net/http'

url = URI("https://api.example.com/v2/{entity}/{project}/eval_results/query")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)
request["Content-Type"] = 'application/json'
request.body = "{\n  \"evaluation_call_ids\": [\n    \"<string>\"\n  ],\n  \"evaluation_run_ids\": [\n    \"<string>\"\n  ],\n  \"filter_logic_operator\": \"or\",\n  \"filters\": [\n    {\n      \"query\": {\n        \"$expr\": {\n          \"$and\": [\n            {\n              \"$literal\": \"<string>\"\n            }\n          ]\n        }\n      },\n      \"evaluation_call_id\": \"<string>\"\n    }\n  ],\n  \"include_costs\": false,\n  \"include_predict_and_score_children\": true,\n  \"include_raw_data_rows\": false,\n  \"include_rows\": true,\n  \"include_summary\": false,\n  \"limit\": 123,\n  \"offset\": 0,\n  \"require_intersection\": false,\n  \"resolve_row_refs\": false,\n  \"sort_by\": [\n    {\n      \"field\": \"<string>\",\n      \"evaluation_call_id\": \"<string>\",\n      \"mode\": \"value\"\n    }\n  ],\n  \"summary_require_intersection\": true\n}"

response = http.request(request)
puts response.read_body

{
  "rows": [
    {
      "row_digest": "<string>",
      "evaluations": [
        {
          "evaluation_call_id": "<string>",
          "trials": [
            {
              "predict_and_score_call_id": "<string>",
              "genai_span_ref": [
                {
                  "span_id": "<string>",
                  "trace_id": "<string>"
                }
              ],
              "model_latency_seconds": 123,
              "model_output": null,
              "predict_call_id": "<string>",
              "scorer_call_ids": {},
              "scores": {},
              "total_cost": 123,
              "total_tokens": 123
            }
          ]
        }
      ],
      "raw_data_row": null
    }
  ],
  "total_rows": 123,
  "summary": {
    "evaluations": [
      {
        "evaluation_call_id": "<string>",
        "display_name": "<string>",
        "evaluation_ref": "<string>",
        "model_ref": "<string>",
        "predict_total_cost": 123,
        "predict_total_tokens": 123,
        "scorer_stats": [
          {
            "scorer_key": "<string>",
            "numeric_count": 0,
            "numeric_mean": 123,
            "pass_known_count": 0,
            "pass_rate": 123,
            "pass_signal_coverage": 123,
            "pass_true_count": 0,
            "path": "<string>",
            "trial_count": 0
          }
        ],
        "started_at": "<string>",
        "trace_id": "<string>",
        "trial_count": 0
      }
    ],
    "row_count": 0
  },
  "warnings": [
    "<string>"
  ]
}

{
  "detail": [
    {
      "loc": [
        "<string>"
      ],
      "msg": "<string>",
      "type": "<string>"
    }
  ]
}

POST

{entity}

{project}

eval_results

query

Eval 結果クエリ

curl --request POST \
  --url https://api.example.com/v2/{entity}/{project}/eval_results/query \
  --header 'Content-Type: application/json' \
  --data '
{
  "evaluation_call_ids": [
    "<string>"
  ],
  "evaluation_run_ids": [
    "<string>"
  ],
  "filter_logic_operator": "or",
  "filters": [
    {
      "query": {
        "$expr": {
          "$and": [
            {
              "$literal": "<string>"
            }
          ]
        }
      },
      "evaluation_call_id": "<string>"
    }
  ],
  "include_costs": false,
  "include_predict_and_score_children": true,
  "include_raw_data_rows": false,
  "include_rows": true,
  "include_summary": false,
  "limit": 123,
  "offset": 0,
  "require_intersection": false,
  "resolve_row_refs": false,
  "sort_by": [
    {
      "field": "<string>",
      "evaluation_call_id": "<string>",
      "mode": "value"
    }
  ],
  "summary_require_intersection": true
}
'

import requests

url = "https://api.example.com/v2/{entity}/{project}/eval_results/query"

payload = {
    "evaluation_call_ids": ["<string>"],
    "evaluation_run_ids": ["<string>"],
    "filter_logic_operator": "or",
    "filters": [
        {
            "query": { "$expr": { "$and": [{ "$literal": "<string>" }] } },
            "evaluation_call_id": "<string>"
        }
    ],
    "include_costs": False,
    "include_predict_and_score_children": True,
    "include_raw_data_rows": False,
    "include_rows": True,
    "include_summary": False,
    "limit": 123,
    "offset": 0,
    "require_intersection": False,
    "resolve_row_refs": False,
    "sort_by": [
        {
            "field": "<string>",
            "evaluation_call_id": "<string>",
            "mode": "value"
        }
    ],
    "summary_require_intersection": True
}
headers = {"Content-Type": "application/json"}

response = requests.post(url, json=payload, headers=headers)

print(response.text)

const options = {
  method: 'POST',
  headers: {'Content-Type': 'application/json'},
  body: JSON.stringify({
    evaluation_call_ids: ['<string>'],
    evaluation_run_ids: ['<string>'],
    filter_logic_operator: 'or',
    filters: [
      {
        query: {$expr: {$and: [{$literal: '<string>'}]}},
        evaluation_call_id: '<string>'
      }
    ],
    include_costs: false,
    include_predict_and_score_children: true,
    include_raw_data_rows: false,
    include_rows: true,
    include_summary: false,
    limit: 123,
    offset: 0,
    require_intersection: false,
    resolve_row_refs: false,
    sort_by: [{field: '<string>', evaluation_call_id: '<string>', mode: 'value'}],
    summary_require_intersection: true
  })
};

fetch('https://api.example.com/v2/{entity}/{project}/eval_results/query', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

<?php

$curl = curl_init();

curl_setopt_array($curl, [
  CURLOPT_URL => "https://api.example.com/v2/{entity}/{project}/eval_results/query",
  CURLOPT_RETURNTRANSFER => true,
  CURLOPT_ENCODING => "",
  CURLOPT_MAXREDIRS => 10,
  CURLOPT_TIMEOUT => 30,
  CURLOPT_HTTP_VERSION => CURL_HTTP_VERSION_1_1,
  CURLOPT_CUSTOMREQUEST => "POST",
  CURLOPT_POSTFIELDS => json_encode([
    'evaluation_call_ids' => [
        '<string>'
    ],
    'evaluation_run_ids' => [
        '<string>'
    ],
    'filter_logic_operator' => 'or',
    'filters' => [
        [
                'query' => [
                                '$expr' => [
                                                                '$and' => [
                                                                                                                                [
                                                                                                                                                                                                                                                                '$literal' => '<string>'
                                                                                                                                ]
                                                                ]
                                ]
                ],
                'evaluation_call_id' => '<string>'
        ]
    ],
    'include_costs' => false,
    'include_predict_and_score_children' => true,
    'include_raw_data_rows' => false,
    'include_rows' => true,
    'include_summary' => false,
    'limit' => 123,
    'offset' => 0,
    'require_intersection' => false,
    'resolve_row_refs' => false,
    'sort_by' => [
        [
                'field' => '<string>',
                'evaluation_call_id' => '<string>',
                'mode' => 'value'
        ]
    ],
    'summary_require_intersection' => true
  ]),
  CURLOPT_HTTPHEADER => [
    "Content-Type: application/json"
  ],
]);

$response = curl_exec($curl);
$err = curl_error($curl);

curl_close($curl);

if ($err) {
  echo "cURL Error #:" . $err;
} else {
  echo $response;
}

package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.example.com/v2/{entity}/{project}/eval_results/query"

	payload := strings.NewReader("{\n  \"evaluation_call_ids\": [\n    \"<string>\"\n  ],\n  \"evaluation_run_ids\": [\n    \"<string>\"\n  ],\n  \"filter_logic_operator\": \"or\",\n  \"filters\": [\n    {\n      \"query\": {\n        \"$expr\": {\n          \"$and\": [\n            {\n              \"$literal\": \"<string>\"\n            }\n          ]\n        }\n      },\n      \"evaluation_call_id\": \"<string>\"\n    }\n  ],\n  \"include_costs\": false,\n  \"include_predict_and_score_children\": true,\n  \"include_raw_data_rows\": false,\n  \"include_rows\": true,\n  \"include_summary\": false,\n  \"limit\": 123,\n  \"offset\": 0,\n  \"require_intersection\": false,\n  \"resolve_row_refs\": false,\n  \"sort_by\": [\n    {\n      \"field\": \"<string>\",\n      \"evaluation_call_id\": \"<string>\",\n      \"mode\": \"value\"\n    }\n  ],\n  \"summary_require_intersection\": true\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(string(body))

}

HttpResponse<String> response = Unirest.post("https://api.example.com/v2/{entity}/{project}/eval_results/query")
  .header("Content-Type", "application/json")
  .body("{\n  \"evaluation_call_ids\": [\n    \"<string>\"\n  ],\n  \"evaluation_run_ids\": [\n    \"<string>\"\n  ],\n  \"filter_logic_operator\": \"or\",\n  \"filters\": [\n    {\n      \"query\": {\n        \"$expr\": {\n          \"$and\": [\n            {\n              \"$literal\": \"<string>\"\n            }\n          ]\n        }\n      },\n      \"evaluation_call_id\": \"<string>\"\n    }\n  ],\n  \"include_costs\": false,\n  \"include_predict_and_score_children\": true,\n  \"include_raw_data_rows\": false,\n  \"include_rows\": true,\n  \"include_summary\": false,\n  \"limit\": 123,\n  \"offset\": 0,\n  \"require_intersection\": false,\n  \"resolve_row_refs\": false,\n  \"sort_by\": [\n    {\n      \"field\": \"<string>\",\n      \"evaluation_call_id\": \"<string>\",\n      \"mode\": \"value\"\n    }\n  ],\n  \"summary_require_intersection\": true\n}")
  .asString();

require 'uri'
require 'net/http'

url = URI("https://api.example.com/v2/{entity}/{project}/eval_results/query")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)
request["Content-Type"] = 'application/json'
request.body = "{\n  \"evaluation_call_ids\": [\n    \"<string>\"\n  ],\n  \"evaluation_run_ids\": [\n    \"<string>\"\n  ],\n  \"filter_logic_operator\": \"or\",\n  \"filters\": [\n    {\n      \"query\": {\n        \"$expr\": {\n          \"$and\": [\n            {\n              \"$literal\": \"<string>\"\n            }\n          ]\n        }\n      },\n      \"evaluation_call_id\": \"<string>\"\n    }\n  ],\n  \"include_costs\": false,\n  \"include_predict_and_score_children\": true,\n  \"include_raw_data_rows\": false,\n  \"include_rows\": true,\n  \"include_summary\": false,\n  \"limit\": 123,\n  \"offset\": 0,\n  \"require_intersection\": false,\n  \"resolve_row_refs\": false,\n  \"sort_by\": [\n    {\n      \"field\": \"<string>\",\n      \"evaluation_call_id\": \"<string>\",\n      \"mode\": \"value\"\n    }\n  ],\n  \"summary_require_intersection\": true\n}"

response = http.request(request)
puts response.read_body

{
  "rows": [
    {
      "row_digest": "<string>",
      "evaluations": [
        {
          "evaluation_call_id": "<string>",
          "trials": [
            {
              "predict_and_score_call_id": "<string>",
              "genai_span_ref": [
                {
                  "span_id": "<string>",
                  "trace_id": "<string>"
                }
              ],
              "model_latency_seconds": 123,
              "model_output": null,
              "predict_call_id": "<string>",
              "scorer_call_ids": {},
              "scores": {},
              "total_cost": 123,
              "total_tokens": 123
            }
          ]
        }
      ],
      "raw_data_row": null
    }
  ],
  "total_rows": 123,
  "summary": {
    "evaluations": [
      {
        "evaluation_call_id": "<string>",
        "display_name": "<string>",
        "evaluation_ref": "<string>",
        "model_ref": "<string>",
        "predict_total_cost": 123,
        "predict_total_tokens": 123,
        "scorer_stats": [
          {
            "scorer_key": "<string>",
            "numeric_count": 0,
            "numeric_mean": 123,
            "pass_known_count": 0,
            "pass_rate": 123,
            "pass_signal_coverage": 123,
            "pass_true_count": 0,
            "path": "<string>",
            "trial_count": 0
          }
        ],
        "started_at": "<string>",
        "trace_id": "<string>",
        "trial_count": 0
      }
    ],
    "row_count": 0
  },
  "warnings": [
    "<string>"
  ]
}

{
  "detail": [
    {
      "loc": [
        "<string>"
      ],
      "msg": "<string>",
      "type": "<string>"
    }
  ]
}

パスパラメータ

entity

string

必須

project

string

必須

ボディ

application/json

evaluation_call_ids

string[] | null

含める評価ルート Call ID。

evaluation_run_ids

string[] | null

Evaluation Runs API の評価 Call ID のエイリアス。

filter_logic_operator

enum<string>

デフォルト:or

評価間でフィルターをどのように結合するかを指定します: 'and'（すべて一致 - 行はすべての eval に一致する必要があります）または 'or'（いずれか一致 - 行はいずれかの eval に一致する必要があります）。デフォルトは 'or'（いずれか一致）です。

利用可能なオプション:

and,

or

filters

EvalResultsFilter · object[] | null

グループ化された行に適用されるフィルターです。複数のフィルターは AND 条件で結合されます。

Show child attributes

include_costs

boolean

デフォルト:false

true の場合、predict-and-score の子 Call にコスト情報を付与し、summary で予測専用の predict_total_cost をレポートできるようにします。明示的に有効化した場合のみ適用され、それ以外の呼び出し元ではコスト計算をスキップします。

include_predict_and_score_children

boolean

デフォルト:true

true の場合（デフォルト）、各 predict_and_score Call の子 Call（predict/score）を取得して、predict_call_id、scorer_call_ids、さらにより正確なレイテンシ/トークンデータを取得します。false の場合、これらのフィールドは predict_and_score Call 自体から導出されます（predict_call_id と scorer_call_ids は null/空になります）。

include_raw_data_rows

boolean

デフォルト:false

true の場合、各結果行の raw_data_row を設定します。インライン行は dict 値として返され、データセット参照行は、resolve_row_refs も true の場合を除き、ref 文字列として返されます。

include_rows

boolean

デフォルト:true

true の場合、グループ化された行/試行データを rows に含め、要求された行レベル view の total_rows を計算します。

include_summary

boolean

デフォルト:false

true の場合、集約された Scorer/評価の summary データを summary に含めます。

limit

integer | null

グループ化と積集合の後に適用されるオプションの行レベルのページサイズ。

offset

integer

デフォルト:0

グループ化と積集合の後に適用されるオプションの行レベルのページオフセット。

require_intersection

boolean

デフォルト:false

true の場合、要求されたすべての評価に存在する行のみを含めます。

resolve_row_refs

boolean

デフォルト:false

true の場合（include_raw_data_rows=True が必要）、表のルックアップによってデータセット行の参照文字列を実際の行データに解決します。false の場合、データセット行 ref はそのまま返されます。

sort_by

EvalResultsSortBy · object[] | null

結果行の並べ替え指定です。サポートされるフィールド接頭辞: scores., inputs., outputs.。null の場合、行は row_digest の昇順で並べ替えられます。

Show child attributes

summary_require_intersection

boolean | null

summary セクションに対するオプションの積集合動作です。null の場合は、require_intersection の値が使用されます。

レスポンス

正常なレスポンス

rows

EvalResultsRow · object[]

必須

Show child attributes

total_rows

integer

必須

summary

EvalResultsSummaryRes · object | null

Show child attributes

warnings

string[]

致命的ではない警告（例: データセット行 ref の解決失敗）。

データセット取得

評価 run の一覧

W&B Models

W&B Weave

Serverless Inference

Serverless Training

パスパラメータ

ボディ

レスポンス