feat(api): add findings module with vector search (MS22-DB-001, MS22-API-001) (#585)
Co-authored-by: Jason Woltje <jason@diversecanvas.com> Co-committed-by: Jason Woltje <jason@diversecanvas.com>
This commit was merged in pull request #585.
This commit is contained in:
337
apps/api/src/findings/findings.service.ts
Normal file
337
apps/api/src/findings/findings.service.ts
Normal file
@@ -0,0 +1,337 @@
|
||||
import { BadRequestException, Injectable, Logger, NotFoundException } from "@nestjs/common";
|
||||
import { Prisma } from "@prisma/client";
|
||||
import { PrismaService } from "../prisma/prisma.service";
|
||||
import { EmbeddingService } from "../knowledge/services/embedding.service";
|
||||
import type { CreateFindingDto, QueryFindingsDto, SearchFindingsDto } from "./dto";
|
||||
|
||||
const findingSelect = {
|
||||
id: true,
|
||||
workspaceId: true,
|
||||
taskId: true,
|
||||
agentId: true,
|
||||
type: true,
|
||||
title: true,
|
||||
data: true,
|
||||
summary: true,
|
||||
createdAt: true,
|
||||
updatedAt: true,
|
||||
} satisfies Prisma.FindingSelect;
|
||||
|
||||
type FindingRecord = Prisma.FindingGetPayload<{ select: typeof findingSelect }>;
|
||||
|
||||
interface RawFindingSearchResult {
|
||||
id: string;
|
||||
workspace_id: string;
|
||||
task_id: string | null;
|
||||
agent_id: string;
|
||||
type: string;
|
||||
title: string;
|
||||
data: Prisma.JsonValue;
|
||||
summary: string;
|
||||
created_at: Date;
|
||||
updated_at: Date;
|
||||
score: number;
|
||||
}
|
||||
|
||||
export interface FindingSearchResult extends FindingRecord {
|
||||
score: number;
|
||||
}
|
||||
|
||||
interface PaginatedMeta {
|
||||
total: number;
|
||||
page: number;
|
||||
limit: number;
|
||||
totalPages: number;
|
||||
}
|
||||
|
||||
export interface PaginatedFindingsResponse {
|
||||
data: FindingRecord[];
|
||||
meta: PaginatedMeta;
|
||||
}
|
||||
|
||||
export interface FindingsSearchResponse {
|
||||
data: FindingSearchResult[];
|
||||
meta: PaginatedMeta;
|
||||
query: string;
|
||||
similarityThreshold: number;
|
||||
}
|
||||
|
||||
/**
|
||||
* Service for managing structured findings with vector search support
|
||||
*/
|
||||
@Injectable()
|
||||
export class FindingsService {
|
||||
private readonly logger = new Logger(FindingsService.name);
|
||||
private readonly defaultSimilarityThreshold: number;
|
||||
|
||||
constructor(
|
||||
private readonly prisma: PrismaService,
|
||||
private readonly embeddingService: EmbeddingService
|
||||
) {
|
||||
const parsedThreshold = Number.parseFloat(process.env.FINDINGS_SIMILARITY_THRESHOLD ?? "0.5");
|
||||
|
||||
this.defaultSimilarityThreshold =
|
||||
Number.isFinite(parsedThreshold) && parsedThreshold >= 0 && parsedThreshold <= 1
|
||||
? parsedThreshold
|
||||
: 0.5;
|
||||
}
|
||||
|
||||
/**
|
||||
* Create a finding and generate its embedding from the summary when available
|
||||
*/
|
||||
async create(workspaceId: string, createFindingDto: CreateFindingDto): Promise<FindingRecord> {
|
||||
if (createFindingDto.taskId) {
|
||||
const task = await this.prisma.agentTask.findUnique({
|
||||
where: {
|
||||
id: createFindingDto.taskId,
|
||||
workspaceId,
|
||||
},
|
||||
select: { id: true },
|
||||
});
|
||||
|
||||
if (!task) {
|
||||
throw new NotFoundException(`Agent task with ID ${createFindingDto.taskId} not found`);
|
||||
}
|
||||
}
|
||||
|
||||
const createInput: Prisma.FindingUncheckedCreateInput = {
|
||||
workspaceId,
|
||||
agentId: createFindingDto.agentId,
|
||||
type: createFindingDto.type,
|
||||
title: createFindingDto.title,
|
||||
data: createFindingDto.data as Prisma.InputJsonValue,
|
||||
summary: createFindingDto.summary,
|
||||
};
|
||||
|
||||
if (createFindingDto.taskId) {
|
||||
createInput.taskId = createFindingDto.taskId;
|
||||
}
|
||||
|
||||
const finding = await this.prisma.finding.create({
|
||||
data: createInput,
|
||||
select: findingSelect,
|
||||
});
|
||||
|
||||
await this.generateAndStoreEmbedding(finding.id, workspaceId, finding.summary);
|
||||
|
||||
if (this.embeddingService.isConfigured()) {
|
||||
return this.findOne(finding.id, workspaceId);
|
||||
}
|
||||
|
||||
return finding;
|
||||
}
|
||||
|
||||
/**
|
||||
* Get paginated findings with optional filters
|
||||
*/
|
||||
async findAll(workspaceId: string, query: QueryFindingsDto): Promise<PaginatedFindingsResponse> {
|
||||
const page = query.page ?? 1;
|
||||
const limit = query.limit ?? 50;
|
||||
const skip = (page - 1) * limit;
|
||||
|
||||
const where: Prisma.FindingWhereInput = {
|
||||
workspaceId,
|
||||
};
|
||||
|
||||
if (query.agentId) {
|
||||
where.agentId = query.agentId;
|
||||
}
|
||||
|
||||
if (query.type) {
|
||||
where.type = query.type;
|
||||
}
|
||||
|
||||
if (query.taskId) {
|
||||
where.taskId = query.taskId;
|
||||
}
|
||||
|
||||
const [data, total] = await Promise.all([
|
||||
this.prisma.finding.findMany({
|
||||
where,
|
||||
select: findingSelect,
|
||||
orderBy: {
|
||||
createdAt: "desc",
|
||||
},
|
||||
skip,
|
||||
take: limit,
|
||||
}),
|
||||
this.prisma.finding.count({ where }),
|
||||
]);
|
||||
|
||||
return {
|
||||
data,
|
||||
meta: {
|
||||
total,
|
||||
page,
|
||||
limit,
|
||||
totalPages: Math.ceil(total / limit),
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
/**
|
||||
* Get a single finding by ID
|
||||
*/
|
||||
async findOne(id: string, workspaceId: string): Promise<FindingRecord> {
|
||||
const finding = await this.prisma.finding.findUnique({
|
||||
where: {
|
||||
id,
|
||||
workspaceId,
|
||||
},
|
||||
select: findingSelect,
|
||||
});
|
||||
|
||||
if (!finding) {
|
||||
throw new NotFoundException(`Finding with ID ${id} not found`);
|
||||
}
|
||||
|
||||
return finding;
|
||||
}
|
||||
|
||||
/**
|
||||
* Semantic search findings using vector similarity
|
||||
*/
|
||||
async search(workspaceId: string, searchDto: SearchFindingsDto): Promise<FindingsSearchResponse> {
|
||||
if (!this.embeddingService.isConfigured()) {
|
||||
throw new BadRequestException(
|
||||
"Finding vector search requires OPENAI_API_KEY to be configured"
|
||||
);
|
||||
}
|
||||
|
||||
const page = searchDto.page ?? 1;
|
||||
const limit = searchDto.limit ?? 20;
|
||||
const offset = (page - 1) * limit;
|
||||
const similarityThreshold = searchDto.similarityThreshold ?? this.defaultSimilarityThreshold;
|
||||
const distanceThreshold = 1 - similarityThreshold;
|
||||
|
||||
const queryEmbedding = await this.embeddingService.generateEmbedding(searchDto.query);
|
||||
const embeddingString = `[${queryEmbedding.join(",")}]`;
|
||||
|
||||
const agentFilter = searchDto.agentId
|
||||
? Prisma.sql`AND f.agent_id = ${searchDto.agentId}`
|
||||
: Prisma.sql``;
|
||||
const typeFilter = searchDto.type ? Prisma.sql`AND f.type = ${searchDto.type}` : Prisma.sql``;
|
||||
const taskFilter = searchDto.taskId
|
||||
? Prisma.sql`AND f.task_id = ${searchDto.taskId}::uuid`
|
||||
: Prisma.sql``;
|
||||
|
||||
const searchResults = await this.prisma.$queryRaw<RawFindingSearchResult[]>`
|
||||
SELECT
|
||||
f.id,
|
||||
f.workspace_id,
|
||||
f.task_id,
|
||||
f.agent_id,
|
||||
f.type,
|
||||
f.title,
|
||||
f.data,
|
||||
f.summary,
|
||||
f.created_at,
|
||||
f.updated_at,
|
||||
(1 - (f.embedding <=> ${embeddingString}::vector)) AS score
|
||||
FROM findings f
|
||||
WHERE f.workspace_id = ${workspaceId}::uuid
|
||||
AND f.embedding IS NOT NULL
|
||||
${agentFilter}
|
||||
${typeFilter}
|
||||
${taskFilter}
|
||||
AND (f.embedding <=> ${embeddingString}::vector) <= ${distanceThreshold}
|
||||
ORDER BY f.embedding <=> ${embeddingString}::vector
|
||||
LIMIT ${limit}
|
||||
OFFSET ${offset}
|
||||
`;
|
||||
|
||||
const countResult = await this.prisma.$queryRaw<[{ count: bigint }]>`
|
||||
SELECT COUNT(*) as count
|
||||
FROM findings f
|
||||
WHERE f.workspace_id = ${workspaceId}::uuid
|
||||
AND f.embedding IS NOT NULL
|
||||
${agentFilter}
|
||||
${typeFilter}
|
||||
${taskFilter}
|
||||
AND (f.embedding <=> ${embeddingString}::vector) <= ${distanceThreshold}
|
||||
`;
|
||||
|
||||
const total = Number(countResult[0].count);
|
||||
|
||||
const data: FindingSearchResult[] = searchResults.map((row) => ({
|
||||
id: row.id,
|
||||
workspaceId: row.workspace_id,
|
||||
taskId: row.task_id,
|
||||
agentId: row.agent_id,
|
||||
type: row.type,
|
||||
title: row.title,
|
||||
data: row.data,
|
||||
summary: row.summary,
|
||||
createdAt: row.created_at,
|
||||
updatedAt: row.updated_at,
|
||||
score: row.score,
|
||||
}));
|
||||
|
||||
return {
|
||||
data,
|
||||
meta: {
|
||||
total,
|
||||
page,
|
||||
limit,
|
||||
totalPages: Math.ceil(total / limit),
|
||||
},
|
||||
query: searchDto.query,
|
||||
similarityThreshold,
|
||||
};
|
||||
}
|
||||
|
||||
/**
|
||||
* Delete a finding
|
||||
*/
|
||||
async remove(id: string, workspaceId: string): Promise<{ message: string }> {
|
||||
const existingFinding = await this.prisma.finding.findUnique({
|
||||
where: {
|
||||
id,
|
||||
workspaceId,
|
||||
},
|
||||
select: { id: true },
|
||||
});
|
||||
|
||||
if (!existingFinding) {
|
||||
throw new NotFoundException(`Finding with ID ${id} not found`);
|
||||
}
|
||||
|
||||
await this.prisma.finding.delete({
|
||||
where: {
|
||||
id,
|
||||
workspaceId,
|
||||
},
|
||||
});
|
||||
|
||||
return { message: "Finding deleted successfully" };
|
||||
}
|
||||
|
||||
/**
|
||||
* Generate and persist embedding for a finding summary
|
||||
*/
|
||||
private async generateAndStoreEmbedding(
|
||||
findingId: string,
|
||||
workspaceId: string,
|
||||
summary: string
|
||||
): Promise<void> {
|
||||
if (!this.embeddingService.isConfigured()) {
|
||||
return;
|
||||
}
|
||||
|
||||
try {
|
||||
const embedding = await this.embeddingService.generateEmbedding(summary);
|
||||
const embeddingString = `[${embedding.join(",")}]`;
|
||||
|
||||
await this.prisma.$executeRaw`
|
||||
UPDATE findings
|
||||
SET embedding = ${embeddingString}::vector,
|
||||
updated_at = NOW()
|
||||
WHERE id = ${findingId}::uuid
|
||||
AND workspace_id = ${workspaceId}::uuid
|
||||
`;
|
||||
} catch (error) {
|
||||
const message = error instanceof Error ? error.message : String(error);
|
||||
this.logger.warn(`Failed to generate embedding for finding ${findingId}: ${message}`);
|
||||
}
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user