Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

feat: add TAP option #43

Merged
merged 4 commits into from
Dec 16, 2024
Merged

feat: add TAP option #43

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
aff7b5b
feat: add TAP format
mdonnalley Dec 13, 2024
3aa02ee
chore: regen schema
mdonnalley Dec 13, 2024
0266d73
chore: bump agents
mdonnalley Dec 16, 2024
b9d7966
chore: bump agents
mdonnalley Dec 16, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion package.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@
"@inquirer/select": "^4.0.1",
"@oclif/core": "^4",
"@oclif/multi-stage-output": "^0.7.12",
"@salesforce/agents": "^0.4.2",
"@salesforce/agents": "^0.5.1",
"@salesforce/core": "^8.8.0",
"@salesforce/kit": "^3.2.1",
"@salesforce/sf-plugins-core": "^12.1.0",
Expand Down
25 changes: 19 additions & 6 deletions schemas/agent-test-results.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -20,14 +20,27 @@
"errorMessage": {
"type": "string"
},
"testCases": {
"type": "array",
"items": {
"$ref": "#/definitions/TestCaseResult"
}
"subjectName": {
"type": "string"
},
"testSet": {
"type": "object",
"properties": {
"name": {
"type": "string"
},
"testCases": {
"type": "array",
"items": {
"$ref": "#/definitions/TestCaseResult"
}
}
},
"required": ["name", "testCases"],
"additionalProperties": false
}
},
"required": ["status", "startTime", "testCases"],
"required": ["status", "startTime", "subjectName", "testSet"],
"additionalProperties": false
},
"TestStatus": {
Expand Down
7 changes: 1 addition & 6 deletions src/flags.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,12 +11,7 @@ Messages.importMessagesDirectoryFromMetaUrl(import.meta.url);
const messages = Messages.loadMessages('@salesforce/plugin-agent', 'shared');

export const resultFormatFlag = Flags.option({
options: [
'json',
'human',
'junit',
// 'tap',
] as const,
options: ['json', 'human', 'junit', 'tap'] as const,
default: 'human',
summary: messages.getMessage('flags.result-format.summary'),
});
Expand Down
12 changes: 10 additions & 2 deletions src/handleTestResults.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6,7 +6,7 @@
*/
import { join } from 'node:path';
import { writeFile, mkdir } from 'node:fs/promises';
import { AgentTestDetailsResponse, jsonFormat, humanFormat, junitFormat } from '@salesforce/agents';
import { AgentTestDetailsResponse, jsonFormat, humanFormat, junitFormat, tapFormat } from '@salesforce/agents';
import { Ux } from '@salesforce/sf-plugins-core/Ux';

async function writeFileToDir(outputDir: string, fileName: string, content: string): Promise<void> {
Expand All @@ -24,7 +24,7 @@ export async function handleTestResults({
outputDir,
}: {
id: string;
format: 'human' | 'json' | 'junit';
format: 'human' | 'json' | 'junit' | 'tap';
results: AgentTestDetailsResponse | undefined;
jsonEnabled: boolean;
outputDir?: string;
Expand Down Expand Up @@ -59,4 +59,12 @@ export async function handleTestResults({
await writeFileToDir(outputDir, `test-result-${id}.xml`, formatted);
}
}

if (format === 'tap') {
const formatted = await tapFormat(results);
ux.log(formatted);
if (outputDir) {
await writeFileToDir(outputDir, `test-result-${id}.txt`, formatted);
}
}
}
2 changes: 1 addition & 1 deletion test/commands/agent/test/results.nut.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -47,7 +47,7 @@ describe('agent test results NUTs', () => {
).jsonOutput;

expect(output?.result.status).to.equal('COMPLETED');
expect(output?.result.testCases.length).to.equal(2);
expect(output?.result.testSet.testCases.length).to.equal(2);

// check that cache does not have an entry
const cache = await AgentTestCache.create();
Expand Down
166 changes: 92 additions & 74 deletions test/mocks/einstein_ai-evaluations_runs_4KBSM000000003F4AQ_details.json
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,80 +3,98 @@
"startTime": "2024-11-28T12:00:00Z",
"endTime": "2024-11-28T12:05:00Z",
"errorMessage": null,
"testCases": [
{
"status": "COMPLETED",
"number": 1,
"startTime": "2024-11-28T12:00:10Z",
"endTime": "2024-11-28T12:00:20Z",
"generatedData": {
"type": "AGENT",
"actionsSequence": ["Action1", "Action2"],
"outcome": "Success",
"topic": "Mathematics",
"inputTokensCount": 50,
"outputTokensCount": 55
},
"expectationResults": [
{
"name": "topic_sequence_match",
"actualValue": "Result A",
"expectedValue": "Result A",
"score": 1.0,
"result": "Passed",
"metricLabel": "Accuracy",
"metricExplainability": "Measures the correctness of the result.",
"status": "Completed",
"startTime": "2024-11-28T12:00:12Z",
"endTime": "2024-11-28T12:00:13Z",
"errorCode": null,
"errorMessage": null
"subjectName": "Copilot_for_Salesforce",
"testSet": {
"name": "CRM_Sanity_v1",
"testCases": [
{
"status": "COMPLETED",
"number": 1,
"startTime": "2024-11-28T12:00:10Z",
"endTime": "2024-11-28T12:00:20Z",
"generatedData": {
"type": "AGENT",
"actionsSequence": ["Action1", "Action2"],
"outcome": "Success",
"topic": "Mathematics",
"inputTokensCount": 50,
"outputTokensCount": 55
},
{
"name": "action_sequence_match",
"actualValue": "Result B",
"expectedValue": "Result B",
"score": 0.9,
"result": "Passed",
"metricLabel": "Precision",
"metricExplainability": "Measures the precision of the result.",
"status": "Completed",
"startTime": "2024-11-28T12:00:14Z",
"endTime": "2024-11-28T12:00:15Z",
"errorCode": null,
"errorMessage": null
}
]
},
{
"status": "ERROR",
"number": 2,
"startTime": "2024-11-28T12:00:30Z",
"endTime": "2024-11-28T12:00:40Z",
"generatedData": {
"type": "AGENT",
"actionsSequence": ["Action3", "Action4"],
"outcome": "Failure",
"topic": "Physics",
"inputTokensCount": 60,
"outputTokensCount": 50
"expectationResults": [
{
"name": "topic_sequence_match",
"actualValue": "Result A",
"expectedValue": "Result A",
"score": 1.0,
"result": "Passed",
"metricLabel": "Accuracy",
"metricExplainability": "Measures the correctness of the result.",
"status": "Completed",
"startTime": "2024-11-28T12:00:12Z",
"endTime": "2024-11-28T12:00:13Z",
"errorCode": null,
"errorMessage": null
},
{
"name": "action_sequence_match",
"actualValue": "Result B",
"expectedValue": "Result B",
"score": 0.9,
"result": "Passed",
"metricLabel": "Precision",
"metricExplainability": "Measures the precision of the result.",
"status": "Completed",
"startTime": "2024-11-28T12:00:14Z",
"endTime": "2024-11-28T12:00:15Z",
"errorCode": null,
"errorMessage": null
}
]
},
"expectationResults": [
{
"name": "topic_sequence_match",
"actualValue": "Result C",
"expectedValue": "Result D",
"score": 0.5,
"result": "Failed",
"metricLabel": "Accuracy",
"metricExplainability": "Measures the correctness of the result.",
"status": "Completed",
"startTime": "2024-11-28T12:00:32Z",
"endTime": "2024-11-28T12:00:33Z",
"errorCode": null,
"errorMessage": null
}
]
}
]
{
"status": "ERROR",
"number": 2,
"startTime": "2024-11-28T12:00:30Z",
"endTime": "2024-11-28T12:00:40Z",
"generatedData": {
"type": "AGENT",
"actionsSequence": ["Action3", "Action4"],
"outcome": "Failure",
"topic": "Physics",
"inputTokensCount": 60,
"outputTokensCount": 50
},
"expectationResults": [
{
"name": "topic_sequence_match",
"actualValue": "Result C",
"expectedValue": "Result D",
"score": 0.5,
"result": "Failed",
"metricLabel": "Accuracy",
"metricExplainability": "Measures the correctness of the result.",
"status": "Completed",
"startTime": "2024-11-28T12:00:32Z",
"endTime": "2024-11-28T12:00:33Z",
"errorCode": null,
"errorMessage": "Expected \"Result D\" but got \"Result C\"."
},
{
"name": "topic_sequence_match",
"actualValue": "Result C",
"expectedValue": "Result D",
"score": 0.5,
"result": "Failed",
"metricLabel": "Accuracy",
"metricExplainability": "Measures the correctness of the result.",
"status": "Completed",
"startTime": "2024-11-28T12:00:32Z",
"endTime": "2024-11-28T12:00:33Z",
"errorCode": null,
"errorMessage": "Expected \"Result D\" but got \"Result C\"."
}
]
}
]
}
}
49 changes: 12 additions & 37 deletions yarn.lock
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1389,16 +1389,16 @@
resolved "https://registry.yarnpkg.com/@pkgjs/parseargs/-/parseargs-0.11.0.tgz#a77ea742fab25775145434eb1d2328cf5013ac33"
integrity sha512-+1VkjdD0QBLPodGrJUeqarH8VAIvQODIbwh9XpP5Syisf7YoQgsJKPNFoqqLQlu+VQ/tVSshMR6loPMn8U+dPg==

"@salesforce/agents@^0.4.2":
version "0.4.2"
resolved "https://registry.yarnpkg.com/@salesforce/agents/-/agents-0.4.2.tgz#1caaf6a521cc3385a0411840876b77147aea880a"
integrity sha512-oTHEf3Dp6T8sYZzB/6V0HaZg7k/aZfTIFsoqYeewWpEa/KP0oopGHWTO9my15n1Guoq0qif6sWqX+RwWpqGgWw==
"@salesforce/agents@^0.5.1":
version "0.5.1"
resolved "https://registry.yarnpkg.com/@salesforce/agents/-/agents-0.5.1.tgz#b6de16004505432c226c02f612c6b0b7b6227f6f"
integrity sha512-FGpCQ3PVzZunoaQVPAJG05eqafOvf2P7fx2w5aZYVg9yqwM/UnBpTBKVvkmdZDsBRTUYaExr6tvboaMc5Hsfzw==
dependencies:
"@oclif/table" "^0.3.5"
"@salesforce/core" "^8.8.0"
"@salesforce/kit" "^3.2.3"
"@salesforce/sf-plugins-core" "^12.1.0"
"@salesforce/source-deploy-retrieve" "^12.10.2"
"@salesforce/source-deploy-retrieve" "^12.10.3"
fast-xml-parser "^4"
nock "^13.5.6"

Expand Down Expand Up @@ -1544,10 +1544,10 @@
cli-progress "^3.12.0"
terminal-link "^3.0.0"

"@salesforce/source-deploy-retrieve@^12.10.2":
version "12.10.2"
resolved "https://registry.yarnpkg.com/@salesforce/source-deploy-retrieve/-/source-deploy-retrieve-12.10.2.tgz#c3737f3751f84cb4754b666edd83c014c91b87bb"
integrity sha512-V7V+nCxhJBcSEh5gJIZt44bsnYJ53HXt3Dec4gphu/Z8mu6Rr3b3yDC3Zw2EKyiDwYYxWvp0pjhqcO2xOTnNLw==
"@salesforce/source-deploy-retrieve@^12.10.3":
version "12.10.3"
resolved "https://registry.yarnpkg.com/@salesforce/source-deploy-retrieve/-/source-deploy-retrieve-12.10.3.tgz#fa16910781188877ffdfa5fde3a0318c0dfe3d07"
integrity sha512-bKIcN6VJajre2chF1xhPCjtR9gZpp8PrFFZ55UcWUMkoFAXscBPRJ7poAeorted3qMzS6wx+AuB27qYUCO+4iQ==
dependencies:
"@salesforce/core" "^8.8.0"
"@salesforce/kit" "^3.2.2"
Expand Down Expand Up @@ -7741,16 +7741,7 @@ stack-utils@^2.0.6:
dependencies:
escape-string-regexp "^2.0.0"

"string-width-cjs@npm:string-width@^4.2.0":
version "4.2.3"
resolved "https://registry.yarnpkg.com/string-width/-/string-width-4.2.3.tgz#269c7117d27b05ad2e536830a8ec895ef9c6d010"
integrity sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g==
dependencies:
emoji-regex "^8.0.0"
is-fullwidth-code-point "^3.0.0"
strip-ansi "^6.0.1"

string-width@^4.0.0, string-width@^4.1.0, string-width@^4.2.0, string-width@^4.2.3:
"string-width-cjs@npm:string-width@^4.2.0", string-width@^4.0.0, string-width@^4.1.0, string-width@^4.2.0, string-width@^4.2.3:
version "4.2.3"
resolved "https://registry.yarnpkg.com/string-width/-/string-width-4.2.3.tgz#269c7117d27b05ad2e536830a8ec895ef9c6d010"
integrity sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g==
Expand Down Expand Up @@ -7872,14 +7863,7 @@ string_decoder@~1.1.1:
dependencies:
safe-buffer "~5.1.0"

"strip-ansi-cjs@npm:strip-ansi@^6.0.1":
version "6.0.1"
resolved "https://registry.yarnpkg.com/strip-ansi/-/strip-ansi-6.0.1.tgz#9e26c63d30f53443e9489495b2105d37b67a85d9"
integrity sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==
dependencies:
ansi-regex "^5.0.1"

[email protected], strip-ansi@^6.0.0, strip-ansi@^6.0.1:
"strip-ansi-cjs@npm:strip-ansi@^6.0.1", [email protected], strip-ansi@^6.0.0, strip-ansi@^6.0.1:
version "6.0.1"
resolved "https://registry.yarnpkg.com/strip-ansi/-/strip-ansi-6.0.1.tgz#9e26c63d30f53443e9489495b2105d37b67a85d9"
integrity sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==
Expand Down Expand Up @@ -8546,7 +8530,7 @@ workerpool@^6.5.1:
resolved "https://registry.yarnpkg.com/workerpool/-/workerpool-6.5.1.tgz#060f73b39d0caf97c6db64da004cd01b4c099544"
integrity sha512-Fs4dNYcsdpYSAfVxhnl1L5zTksjvOJxtC5hzMNl+1t9B8hTJTdKDyZ5ju7ztgPy+ft9tBFXoOlDNiOT9WUXZlA==

"wrap-ansi-cjs@npm:wrap-ansi@^7.0.0":
"wrap-ansi-cjs@npm:wrap-ansi@^7.0.0", wrap-ansi@^7.0.0:
version "7.0.0"
resolved "https://registry.yarnpkg.com/wrap-ansi/-/wrap-ansi-7.0.0.tgz#67e145cff510a6a6984bdf1152911d69d2eb9e43"
integrity sha512-YVGIj2kamLSTxw6NsZjoBxfSwsn0ycdesmc4p+Q21c5zPuZ1pl+NfxVdxPtdHvmNVOQ6XSYG4AUtyt/Fi7D16Q==
Expand All @@ -8564,15 +8548,6 @@ wrap-ansi@^6.2.0:
string-width "^4.1.0"
strip-ansi "^6.0.0"

wrap-ansi@^7.0.0:
version "7.0.0"
resolved "https://registry.yarnpkg.com/wrap-ansi/-/wrap-ansi-7.0.0.tgz#67e145cff510a6a6984bdf1152911d69d2eb9e43"
integrity sha512-YVGIj2kamLSTxw6NsZjoBxfSwsn0ycdesmc4p+Q21c5zPuZ1pl+NfxVdxPtdHvmNVOQ6XSYG4AUtyt/Fi7D16Q==
dependencies:
ansi-styles "^4.0.0"
string-width "^4.1.0"
strip-ansi "^6.0.0"

wrap-ansi@^8.1.0:
version "8.1.0"
resolved "https://registry.yarnpkg.com/wrap-ansi/-/wrap-ansi-8.1.0.tgz#56dc22368ee570face1b49819975d9b9a5ead214"
Expand Down
Loading