Skip to content

fix: accuracy test fixes #651

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

Merged
merged 4 commits into from
Oct 15, 2025
Merged

fix: accuracy test fixes #651

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
2cb0865
fix accuracy tests
nirinchev Oct 14, 2025
8ec9c59
remove import
nirinchev Oct 14, 2025
16121e3
revert before/afterAll solution for createIndex.test.ts
nirinchev Oct 15, 2025
584f02a
use a date that's way in the past for slow queries
nirinchev Oct 15, 2025
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
3 changes: 2 additions & 1 deletion scripts/accuracy/generateTestSummary.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -73,7 +73,8 @@ function formatToolCallsWithTooltip(toolCalls: ExpectedToolCall[] | LLMToolCall[
return toolCalls
.map((call) => {
const params = JSON.stringify(call.parameters, null, 2);
return `<span class="tool-call" title="${params.replace(/"/g, "&quot;")}">${call.toolName}</span>`;
const isOptional = "optional" in call && call.optional;
return `<span class="tool-call" title="${params.replace(/"/g, "&quot;")}">${isOptional ? "(" : ""}${call.toolName}${isOptional ? ")" : ""}</span>`;
})
.join(", ");
}
Expand Down
5 changes: 5 additions & 0 deletions tests/accuracy/createCollection.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -28,6 +28,11 @@ describeAccuracyTests([
{
prompt: "If and only if, the namespace 'mflix.documentaries' does not exist, then create it",
expectedToolCalls: [
{
toolName: "list-databases",
parameters: {},
optional: true,
},
{
toolName: "list-collections",
parameters: {
Expand Down
12 changes: 12 additions & 0 deletions tests/accuracy/dropCollection.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,18 @@ describeAccuracyTests([
{
prompt: "Remove mflix.movies namespace from my cluster.",
expectedToolCalls: [
{
toolName: "list-databases",
parameters: {},
optional: true,
},
{
toolName: "list-collections",
parameters: {
database: "mflix",
},
optional: true,
},
{
toolName: "drop-collection",
parameters: {
Expand Down
8 changes: 7 additions & 1 deletion tests/accuracy/find.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -124,6 +124,7 @@ describeAccuracyTests([
limit: Matcher.anyValue,
sort: Matcher.anyValue,
},
optional: true,
},
{
toolName: "export",
Expand All @@ -137,14 +138,19 @@ describeAccuracyTests([
arguments: Matcher.anyOf(
Matcher.emptyObjectOrUndefined,
Matcher.value({
filter: Matcher.anyValue,
filter: Matcher.emptyObjectOrUndefined,
projection: Matcher.anyValue,
limit: Matcher.anyValue,
sort: Matcher.anyValue,
})
),
},
],
jsonExportFormat: Matcher.anyOf(
Matcher.undefined,
Matcher.value("relaxed"),
Matcher.value("canonical")
),
},
},
],
Expand Down
72 changes: 21 additions & 51 deletions tests/accuracy/getPerformanceAdvisor.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,21 +35,27 @@ const mockedTools = {
},
};

const listProjectsAndClustersToolCalls = [
{
toolName: "atlas-list-projects",
parameters: {},
optional: true,
},
{
toolName: "atlas-list-clusters",
parameters: {
projectId: "mflix",
},
optional: true,
},
];

describeAccuracyTests([
// Test for Suggested Indexes operation
{
prompt: "Can you give me index suggestions for the database 'mflix' in the project 'mflix' and cluster 'mflix-cluster'?",
expectedToolCalls: [
{
toolName: "atlas-list-projects",
parameters: {},
},
{
toolName: "atlas-list-clusters",
parameters: {
projectId: "mflix",
},
},
...listProjectsAndClustersToolCalls,
{
toolName: "atlas-get-performance-advisor",
parameters: {
Expand All @@ -65,16 +71,7 @@ describeAccuracyTests([
{
prompt: "Show me drop index suggestions for the 'mflix' project and 'mflix-cluster' cluster",
expectedToolCalls: [
{
toolName: "atlas-list-projects",
parameters: {},
},
{
toolName: "atlas-list-clusters",
parameters: {
projectId: "mflix",
},
},
...listProjectsAndClustersToolCalls,
{
toolName: "atlas-get-performance-advisor",
parameters: {
Expand All @@ -88,18 +85,9 @@ describeAccuracyTests([
},
// Test for Slow Query Logs operation
{
prompt: "Show me the slow query logs for the 'mflix' project and 'mflix-cluster' cluster for the namespaces 'mflix.movies' and 'mflix.shows' since January 1st, 2025.",
prompt: "Show me the slow query logs for the 'mflix' project and 'mflix-cluster' cluster for the namespaces 'mflix.movies' and 'mflix.shows' since January 1st, 2025 (a date that is certainly in the past!).",
expectedToolCalls: [
{
toolName: "atlas-list-projects",
parameters: {},
},
{
toolName: "atlas-list-clusters",
parameters: {
projectId: "mflix",
},
},
...listProjectsAndClustersToolCalls,
{
toolName: "atlas-get-performance-advisor",
parameters: {
Expand All @@ -117,16 +105,7 @@ describeAccuracyTests([
{
prompt: "Give me schema suggestions for the 'mflix' project and 'mflix-cluster' cluster",
expectedToolCalls: [
{
toolName: "atlas-list-projects",
parameters: {},
},
{
toolName: "atlas-list-clusters",
parameters: {
projectId: "mflix",
},
},
...listProjectsAndClustersToolCalls,
{
toolName: "atlas-get-performance-advisor",
parameters: {
Expand All @@ -142,16 +121,7 @@ describeAccuracyTests([
{
prompt: "Show me all performance advisor recommendations for the 'mflix' project and 'mflix-cluster' cluster",
expectedToolCalls: [
{
toolName: "atlas-list-projects",
parameters: {},
},
{
toolName: "atlas-list-clusters",
parameters: {
projectId: "mflix",
},
},
...listProjectsAndClustersToolCalls,
{
toolName: "atlas-get-performance-advisor",
parameters: {
Expand Down
4 changes: 3 additions & 1 deletion tests/accuracy/sdk/accuracyResultStorage/resultStorage.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,7 +4,9 @@ export interface LLMToolCall {
parameters: Record<string, unknown>;
}

export type ExpectedToolCall = Omit<LLMToolCall, "toolCallId">;
export type ExpectedToolCall = Omit<LLMToolCall, "toolCallId"> & {
optional?: boolean;
};

export const AccuracyRunStatus = {
Done: "done",
Expand Down
13 changes: 8 additions & 5 deletions tests/accuracy/sdk/accuracyScorer.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -81,12 +81,15 @@ export function calculateToolCallingAccuracy(
.sort((a, b) => b.score - a.score || a.index - b.index);

const bestMatch = candidates[0];
if (!bestMatch || bestMatch.score === 0) {
return 0; // No matching tool call found, return 0
if (bestMatch) {
checkedActualToolCallIndexes.add(bestMatch.index);
currentScore = Math.min(currentScore, bestMatch.score);
} else if (expectedCall.optional) {
// Optional expected tool call not found, but it's okay, continue
continue;
} else {
return 0; // Required expected tool call not found, return 0
}

checkedActualToolCallIndexes.add(bestMatch.index);
currentScore = Math.min(currentScore, bestMatch.score);
}

return currentScore;
Expand Down
Loading