Implement dual-model filing pipeline with Ollama extraction

This commit is contained in:
2026-02-28 16:31:25 -05:00
parent 0615534f4b
commit a09001501e
16 changed files with 872 additions and 51 deletions

View File

@@ -4,6 +4,7 @@ import { auth } from '@/lib/auth';
import { requireAuthenticatedSession } from '@/lib/server/auth-session';
import { asErrorMessage, jsonError } from '@/lib/server/http';
import { buildPortfolioSummary } from '@/lib/server/portfolio';
import { redactInternalFilingAnalysisFields } from '@/lib/server/api/filing-redaction';
import { getFilingByAccession, listFilingsRecords } from '@/lib/server/repos/filings';
import {
deleteHoldingByIdRecord,
@@ -332,8 +333,9 @@ export const app = new Elysia({ prefix: '/api' })
getQuote(ticker),
getPriceHistory(ticker)
]);
const redactedFilings = filings.map(redactInternalFilingAnalysisFields);
const latestFiling = filings[0] ?? null;
const latestFiling = redactedFilings[0] ?? null;
const holding = holdings.find((entry) => entry.ticker === ticker) ?? null;
const watchlistItem = watchlist.find((entry) => entry.ticker === ticker) ?? null;
@@ -341,7 +343,7 @@ export const app = new Elysia({ prefix: '/api' })
?? watchlistItem?.company_name
?? ticker;
const financials = filings
const financials = redactedFilings
.filter((entry) => entry.metrics)
.map((entry) => ({
filingDate: entry.filing_date,
@@ -353,7 +355,7 @@ export const app = new Elysia({ prefix: '/api' })
debt: entry.metrics?.debt ?? null
}));
const aiReports = filings
const aiReports = redactedFilings
.filter((entry) => entry.analysis?.text || entry.analysis?.legacyInsights)
.slice(0, 8)
.map((entry) => ({
@@ -377,7 +379,7 @@ export const app = new Elysia({ prefix: '/api' })
position: holding,
priceHistory,
financials,
filings: filings.slice(0, 20),
filings: redactedFilings.slice(0, 20),
aiReports
}
});
@@ -446,7 +448,7 @@ export const app = new Elysia({ prefix: '/api' })
limit: Number.isFinite(limit) ? limit : 50
});
return Response.json({ filings });
return Response.json({ filings: filings.map(redactInternalFilingAnalysisFields) });
}, {
query: t.Object({
ticker: t.Optional(t.String()),

View File

@@ -0,0 +1,52 @@
import { describe, expect, it } from 'bun:test';
import type { Filing } from '@/lib/types';
import { redactInternalFilingAnalysisFields } from './filing-redaction';
function filingWithExtraction(): Filing {
return {
id: 7,
ticker: 'MSFT',
filing_type: '10-K',
filing_date: '2026-02-01',
accession_number: '0000789019-26-000001',
cik: '0000789019',
company_name: 'Microsoft Corporation',
filing_url: 'https://www.sec.gov/Archives/edgar/data/789019/000078901926000001/a10k.htm',
submission_url: null,
primary_document: 'a10k.htm',
metrics: null,
analysis: {
provider: 'zhipu',
model: 'glm-4.7-flashx',
text: 'Report text',
extraction: {
summary: 'Internal extraction summary',
keyPoints: ['a'],
redFlags: ['b'],
followUpQuestions: ['c'],
portfolioSignals: ['d'],
confidence: 0.4
},
extractionMeta: {
provider: 'ollama',
model: 'qwen3:8b',
source: 'primary_document',
generatedAt: '2026-02-01T00:00:00.000Z'
}
},
created_at: '2026-02-01T00:00:00.000Z',
updated_at: '2026-02-01T00:00:00.000Z'
};
}
describe('filing response redaction', () => {
it('removes internal extraction fields while preserving public analysis fields', () => {
const redacted = redactInternalFilingAnalysisFields(filingWithExtraction());
expect(redacted.analysis?.provider).toBe('zhipu');
expect(redacted.analysis?.model).toBe('glm-4.7-flashx');
expect(redacted.analysis?.text).toBe('Report text');
expect(redacted.analysis?.extraction).toBeUndefined();
expect(redacted.analysis?.extractionMeta).toBeUndefined();
});
});

View File

@@ -0,0 +1,15 @@
import type { Filing } from '@/lib/types';
export function redactInternalFilingAnalysisFields(filing: Filing): Filing {
if (!filing.analysis) {
return filing;
}
const { extraction: _extraction, extractionMeta: _extractionMeta, ...analysis } = filing.analysis;
const hasPublicFields = Object.keys(analysis).length > 0;
return {
...filing,
analysis: hasPublicFields ? analysis : null
};
}