Add three-stage cross-domain news research and reading briefings
Group original news into tagged fact packages, review each package with evidence and read-only research, and build complete 72-hour reading lists with independent model selections. Add paused RSS/Atom subscriptions, news-scoped Skills, versioned analysis records, and locked deployment support for SearXNG. Preserve original news and keep live migration and legacy AI cleanup separately authorized. Validate with 604 backend tests, 49 frontend tests, 19 CI control tests, lint, type checks, production build, and disposable PostgreSQL migration and three-stage smoke checks.
This commit is contained in:
eezd committed
1 parent
ae35d3d608
commit
80d266dba7
44 files changed
+2682
-574
No files matched your search
+141
-119
@@ -1,38 +1,19 @@
|
||||
"""Exercise news risk regressions against the synthetic CI database only."""
|
||||
"""Exercise both news stages against the disposable CI database and fake models."""
|
||||
import asyncio
|
||||
from datetime import UTC, datetime, timedelta
|
||||
import json
|
||||
import os
|
||||
from types import SimpleNamespace
|
||||
from unittest.mock import AsyncMock, patch
|
||||
from urllib.parse import urlparse
|
||||
|
||||
from fastapi import HTTPException
|
||||
from sqlalchemy import delete, func, select
|
||||
|
||||
from alphapulse.ai import llm
|
||||
from alphapulse.ai.skills import create_node, add_version, published_skills
|
||||
from alphapulse.api.settings import delete_provider
|
||||
from alphapulse.core.db import SessionLocal, engine
|
||||
from alphapulse.models import AiProvider, NewsItem
|
||||
from alphapulse.models.risk import RiskAnalysis, RiskEvent, RiskSettings, RiskSnapshot
|
||||
from alphapulse.news_risk import service, worker
|
||||
|
||||
|
||||
def label(identifier, title, stars=0, event_id=None):
|
||||
return service.Label(id=identifier, title=title, summary=title, category='security', entities=['Exchange'],
|
||||
stars=stars, direction='unknown', verification='reported', confidence='low', reasoning='Synthetic fixture',
|
||||
uncertainty='Synthetic fixture', evidence=title, event_id=event_id)
|
||||
|
||||
|
||||
def event(at, stars):
|
||||
return RiskEvent(title='Exchange security incident', category='security', entities=['Exchange'],
|
||||
stars=stars, peak_stars=stars, result=label(1, 'Exchange security incident', stars).model_dump(),
|
||||
published_at=at, updated_at=at)
|
||||
|
||||
|
||||
def news(external_id, at, content='Synthetic fixture'):
|
||||
return NewsItem(source_code='wscn_news', external_id='ci-risk-'+external_id, url='https://example.invalid',
|
||||
title='Exchange security incident', summary='', content_text=content, published_at=at, content_hash=external_id)
|
||||
from alphapulse.models import AiProvider, NewsItem, SkillNode, SkillVersion
|
||||
from alphapulse.models.risk import RiskSettings, RiskSnapshot, ArticleAnalysis, NewsSummary, SummaryArticle, SummaryVersion, PackageEvidence
|
||||
from alphapulse.news_risk import service, pipeline, worker, briefing
|
||||
|
||||
|
||||
async def main():
|
||||
@@ -42,111 +23,152 @@ async def main():
|
||||
now = datetime.now(UTC)
|
||||
async with SessionLocal() as session:
|
||||
assert await session.scalar(select(func.count()).select_from(RiskSettings)) == 0
|
||||
provider = AiProvider(name='ci-risk', protocol='openai', base_url='https://example.invalid',
|
||||
api_key_encrypted=b'fake', api_key_hint='fake', model='shared', max_tokens=256, context_window=200000)
|
||||
session.add(provider); await session.flush()
|
||||
provider_id = provider.id
|
||||
session.add(RiskSettings(id=1, enabled=True, provider_id=provider.id, model='synthetic-news'))
|
||||
original = event(now-timedelta(hours=3), 5)
|
||||
article = news('correction', now-timedelta(minutes=1))
|
||||
session.add_all([original, article]); await session.flush()
|
||||
original_id, article_id = original.id, article.id
|
||||
session.add(RiskAnalysis(news_id=article.id, content_hash='old-content', state='done', event_id=original.id,
|
||||
input={'title': article.title, 'published_at': article.published_at.isoformat(), 'progress_at': article.published_at.isoformat()}))
|
||||
for i in range(80):
|
||||
unrelated = event(now-timedelta(minutes=i+2), 0)
|
||||
unrelated.title = f'Company {i} ordinary earnings announcement'
|
||||
unrelated.entities = [f'Company {i}']
|
||||
unrelated.result = label(1, unrelated.title).model_dump()
|
||||
session.add(unrelated)
|
||||
provider = AiProvider(name='ci-news', protocol='openai', base_url='https://example.invalid', api_key_encrypted=b'fake',
|
||||
api_key_hint='fake', model='shared', max_tokens=2048, context_window=200000)
|
||||
session.add(provider); await session.flush(); provider_id = provider.id
|
||||
skill_ids = {}
|
||||
for role, (name, content) in pipeline.TEMPLATES.items():
|
||||
node = await create_node(session, None, 'skill', name, scope='news')
|
||||
version = await add_version(session, node, content, 'user', 'CI')
|
||||
node.published_version_id = version.id
|
||||
skill_ids[role+'_skill_id'] = node.id
|
||||
await session.flush()
|
||||
assert not await published_skills(session), 'News skills leaked into Chat or trading'
|
||||
session.add(RiskSettings(id=1, enabled=True, provider_id=provider_id, model='luna', second_provider_id=provider_id,
|
||||
second_model='reviewer', third_provider_id=provider_id, third_model='senior', start_at=now-timedelta(minutes=2), **skill_ids))
|
||||
for key, at, content in [('old', now-timedelta(days=1), 'old'), ('first', now-timedelta(minutes=1), 'original'),
|
||||
('duplicate', now-timedelta(seconds=30), 'repost')]:
|
||||
session.add(NewsItem(source_code='wscn_news', external_id='ci-news-'+key, url='https://example.invalid/'+key,
|
||||
title='Exchange security incident', summary='', content_text=content, content_hash=key, published_at=at))
|
||||
await session.commit()
|
||||
|
||||
payloads = []
|
||||
calls = []
|
||||
broken = False
|
||||
inject_update = False
|
||||
finish_during_grouping = False
|
||||
async def stream(provider, system, messages, tools, effort):
|
||||
payload = json.loads(messages[0]['text']); payloads.append(payload)
|
||||
if 'news' in payload:
|
||||
items = [label(n['id'], n['title'], 1 if n.get('revision_no') == 2 else 0,
|
||||
original_id if n.get('revision_no') == 2 else None) for n in payload['news']]
|
||||
result = service.Labels(items=items)
|
||||
nonlocal inject_update, finish_during_grouping
|
||||
data = json.loads(messages[0]['text']); calls.append((provider.model, data))
|
||||
if 'article' in data:
|
||||
assert provider.model == 'luna'
|
||||
if finish_during_grouping:
|
||||
finish_during_grouping = False
|
||||
await worker.summarize()
|
||||
candidates = data['candidates']
|
||||
result = pipeline.Grouping( reason='Financial event', events=[pipeline.Match(
|
||||
event_id=candidates[0]['id'] if candidates else None, group_key=None if candidates else 'security',
|
||||
title='Exchange security incident', category='security', entities=['Exchange'], fact='incident',
|
||||
evidence=data['article']['title'], urgent=True, tags=['Exchange'], content_kind='report')])
|
||||
elif data.get('task') == 'event':
|
||||
assert provider.model == 'reviewer'
|
||||
if broken: raise ValueError('Synthetic validation failure')
|
||||
if inject_update:
|
||||
inject_update = False
|
||||
async with SessionLocal() as other:
|
||||
changed = await other.get(NewsSummary, data['id'], with_for_update=True)
|
||||
worker.schedule(changed, datetime.now(UTC), True)
|
||||
await other.commit()
|
||||
correction = any(a['content'] == 'correction' for a in data['articles'])
|
||||
result = pipeline.EventResult(id=data['id'], title='Exchange security incident', summary='Correction' if correction else 'Incident',
|
||||
category='security', entities=['Exchange'], risk_stars=1 if correction else 5, importance_stars=5, direction='unknown',
|
||||
verification='refuted' if correction else 'reported', confidence='low', reasoning='Synthetic', uncertainty='Synthetic',
|
||||
facts='Synthetic', timeline='Synthetic',
|
||||
asset_impacts='Exchange', opinions='Uncertain', counter_evidence='Unknown', watchpoints='Updates', claims=[pipeline.Claim(statement='Synthetic', verdict='insufficient', evidence_refs=[])], related_packages=[], excluded_article_ids=[],
|
||||
evidence_refs=[pipeline.Evidence(article_id=a['id'], quote=a['title']) for a in data['articles']])
|
||||
elif data.get('task') == 'ranking':
|
||||
assert provider.model == 'senior'
|
||||
result = briefing.Ranked(reading=[briefing.Reading(id=p['id'], revision=p['revision'], priority=90, reason='Synthetic') for p in data['packages']], notes='Synthetic notes')
|
||||
else:
|
||||
result = service.Outlook(windows=[service.Scenario(hours=h, direction='unknown', summary='Synthetic',
|
||||
conditions='Synthetic', counter_factors='Synthetic', uncertainty='Synthetic',
|
||||
event_ids=[e['id'] for e in payload['windows'][str(h)]]) for h in service.WINDOWS])
|
||||
yield llm.Completion(tool_calls=[{'name': 'submit', 'arguments': result.model_dump()}])
|
||||
|
||||
class Binance:
|
||||
def __init__(self, **kwargs): self.http = type('Http', (), {'aclose': AsyncMock()})()
|
||||
async def ticker(self, *args): return {'lastPrice': '100', 'priceChangePercent': '-6', 'closeTime': 1}
|
||||
|
||||
trend = briefing.Trend(summary='Unknown', conditions='Unknown', counter_evidence='Unknown', watchpoints='Unknown')
|
||||
result = briefing.Synthesis(summary='Synthetic', market='Unknown', short_term=trend, medium_term=trend, references=[])
|
||||
yield llm.Completion(tool_calls=[{'name': 'submit', 'arguments': result.model_dump()}], usage={'input_tokens': 10})
|
||||
try:
|
||||
async with SessionLocal() as session:
|
||||
candidates, _ = await worker.find_candidates(session, [SimpleNamespace(
|
||||
news_id=0, input={'title': 'Exchange security update', 'summary': ''})], now)
|
||||
assert original_id in {e.id for e in candidates}
|
||||
article = await session.get(NewsItem, article_id); article.revision_no = 2
|
||||
await session.commit()
|
||||
with patch.object(llm, 'stream', stream): await worker.analyze()
|
||||
async with SessionLocal() as session:
|
||||
original = await session.get(RiskEvent, original_id)
|
||||
assert original.stars == 1 and original.peak_stars == 5
|
||||
assert original_id in {e['id'] for e in payloads[-1]['candidates']}
|
||||
assert await session.scalar(select(func.count()).select_from(RiskEvent)) == 81
|
||||
assert (await session.get(AiProvider, provider_id)).model == 'shared'
|
||||
await session.execute(delete(RiskAnalysis)); await session.execute(delete(RiskEvent))
|
||||
session.add_all([event(now-timedelta(days=3), 5) for _ in range(100)])
|
||||
fresh = event(now-timedelta(minutes=1), 4); session.add(fresh)
|
||||
await session.flush(); fresh_id = fresh.id
|
||||
await session.commit()
|
||||
with patch.object(llm, 'stream', stream), patch.object(worker, 'BinanceClient', Binance):
|
||||
await worker.summarize()
|
||||
for hours in ('24', '48'):
|
||||
assert [e['id'] for e in payloads[-1]['windows'][hours]] == [fresh_id]
|
||||
assert len(payloads[-1]['windows']['168']) == 100
|
||||
async with SessionLocal() as session:
|
||||
snapshot = await session.scalar(select(RiskSnapshot))
|
||||
snapshot.created_at = now-timedelta(minutes=11)
|
||||
provider = await session.get(AiProvider, provider_id); provider.context_window = 8192
|
||||
await session.commit()
|
||||
with patch.object(llm, 'stream', stream), patch.object(worker, 'BinanceClient', Binance):
|
||||
await worker.summarize()
|
||||
for hours in ('24', '48'):
|
||||
assert [e['id'] for e in payloads[-1]['windows'][hours]] == [fresh_id]
|
||||
assert 0 < len(payloads[-1]['windows']['168']) < 100
|
||||
async with SessionLocal() as session:
|
||||
await session.execute(delete(RiskEvent))
|
||||
provider = await session.get(AiProvider, provider_id); provider.context_window = 8192
|
||||
article = await session.get(NewsItem, article_id); article.title = 'Batch article'
|
||||
article.content_hash = 'current'; article.content_text = 'x'*1000; article.revision_no = 3
|
||||
session.add_all([news(str(i), now-timedelta(minutes=i+2), 'x'*1000) for i in range(9)])
|
||||
session.add(news('oversized', now-timedelta(seconds=1), 'x'*30000))
|
||||
await session.commit()
|
||||
with patch.object(llm, 'stream', stream):
|
||||
for _ in range(15): await worker.analyze()
|
||||
async with SessionLocal() as session:
|
||||
counts = dict((await session.execute(select(RiskAnalysis.state, func.count()).group_by(RiskAnalysis.state))).all())
|
||||
assert counts == {'done': 10, 'failed': 1}, counts
|
||||
failed = await session.scalar(select(RiskAnalysis).where(RiskAnalysis.state == 'failed'))
|
||||
assert len(failed.input['content']) == 30000 and '上下文' in failed.error
|
||||
assert all(len(p['news']) < 10 for p in payloads if 'news' in p)
|
||||
try:
|
||||
await delete_provider(provider_id, session)
|
||||
except HTTPException as error:
|
||||
assert error.status_code == 409
|
||||
else:
|
||||
raise AssertionError('Enabled risk provider deletion was accepted')
|
||||
settings = await session.get(RiskSettings, 1); settings.enabled = False
|
||||
await session.commit()
|
||||
await delete_provider(provider_id, session)
|
||||
assert (await session.get(RiskSettings, 1)).provider_id is None
|
||||
print('News risk: correction recall, independent windows, context packing, isolated oversized article and provider deletion passed')
|
||||
await worker.analyze()
|
||||
finish_during_grouping = True
|
||||
await worker.analyze()
|
||||
async with SessionLocal() as session:
|
||||
event = await session.scalar(select(NewsSummary))
|
||||
assert event.state == 'pending' and event.revision == 2 and event.published_revision == 1, 'Grouping lost an update after concurrent summary completion'
|
||||
await worker.summarize()
|
||||
async with SessionLocal() as session:
|
||||
assert await session.scalar(select(func.count()).select_from(ArticleAnalysis)) == 2
|
||||
event = await session.scalar(select(NewsSummary)); event_id = event.id
|
||||
assert event.stars == 5 and event.revision == 2
|
||||
assert await session.scalar(select(func.count()).select_from(SummaryArticle)) == 2
|
||||
article = await session.scalar(select(NewsItem).where(NewsItem.external_id == 'ci-news-first'))
|
||||
article.content_hash = 'corrected'; article.content_text = 'correction'; article.revision_no = 2
|
||||
article.source_updated_at = datetime.now(UTC)
|
||||
await session.commit()
|
||||
await worker.analyze(); await worker.summarize()
|
||||
async with SessionLocal() as session:
|
||||
event = await session.get(NewsSummary, event_id)
|
||||
assert event.stars == 1 and event.peak_stars == 5 and event.result['verification'] == 'refuted'
|
||||
version = await session.scalar(select(SummaryVersion).order_by(SummaryVersion.id.desc()).limit(1))
|
||||
assert len(version.input['articles']) == 2, 'Superseded article revision entered the current summary'
|
||||
event.revision += 1; event.state = 'pending'; event.due_at = datetime.now(UTC)
|
||||
await session.commit()
|
||||
inject_update = True
|
||||
await worker.summarize()
|
||||
async with SessionLocal() as session:
|
||||
event = await session.get(NewsSummary, event_id)
|
||||
assert event.state == 'pending' and event.revision > event.published_revision, 'Concurrent update was lost'
|
||||
broken = True
|
||||
await worker.summarize()
|
||||
async with SessionLocal() as session:
|
||||
event = await session.get(NewsSummary, event_id)
|
||||
assert event.state == 'failed' and event.stars == 1 and event.result['summary'] == 'Correction'
|
||||
broken = False
|
||||
async with SessionLocal() as session:
|
||||
failed = RiskSnapshot(state='failed', model='reviewer', quotes={}, error='Synthetic old failure',
|
||||
created_at=datetime.now(UTC)-timedelta(minutes=20))
|
||||
session.add(failed); await session.commit(); failed_id = failed.id
|
||||
await worker.outlook()
|
||||
async with SessionLocal() as session:
|
||||
snapshot = await session.scalar(select(RiskSnapshot).order_by(RiskSnapshot.id.desc()).limit(1))
|
||||
assert snapshot.state == 'done' and snapshot.result['inputs'][0]['id'] == event_id
|
||||
assert (await session.get(AiProvider, provider_id)).model == 'shared'
|
||||
from fastapi import FastAPI
|
||||
import httpx
|
||||
from alphapulse.api import risk as risk_api, news as news_api
|
||||
app = FastAPI()
|
||||
app.include_router(risk_api.router); app.include_router(news_api.router)
|
||||
app.dependency_overrides[risk_api.require_admin] = lambda: None
|
||||
with patch.object(risk_api.redis, 'get', AsyncMock(return_value=None)):
|
||||
async with httpx.AsyncClient(transport=httpx.ASGITransport(app), base_url='http://test') as client:
|
||||
for path in ('/api/risk/settings', '/api/risk/overview', '/api/risk/summaries',
|
||||
f'/api/risk/summaries/{event_id}', '/api/news'):
|
||||
response = await client.get(path)
|
||||
assert response.status_code == 200, (path, response.text)
|
||||
assert response.json()['code'] == 0
|
||||
response = await client.get(f'/api/risk/summaries/{event_id}')
|
||||
assert response.json()['data']['event']['article_count'] == 2
|
||||
assert response.json()['data']['versions']
|
||||
response = await client.get('/api/risk/summaries', params={'hours': 25})
|
||||
assert response.status_code == 422
|
||||
response = await client.post(f'/api/risk/tasks/{failed_id}/retry', params={'stage': 'outlook'})
|
||||
assert response.status_code == 202, response.text
|
||||
await worker.outlook()
|
||||
async with SessionLocal() as session:
|
||||
assert await session.scalar(select(func.count()).select_from(RiskSnapshot)) == 3, 'Old failed outlook retry was ignored'
|
||||
assert not (await session.get(RiskSnapshot, failed_id)).quotes.get('retry_requested')
|
||||
async with SessionLocal() as session:
|
||||
try: await delete_provider(provider_id, session)
|
||||
except HTTPException as error: assert error.status_code == 409
|
||||
else: raise AssertionError('Enabled model provider deletion accepted')
|
||||
settings = await session.get(RiskSettings, 1); settings.enabled = False
|
||||
await session.commit(); await delete_provider(provider_id, session)
|
||||
assert settings.provider_id is None and settings.second_provider_id is None and settings.third_provider_id is None
|
||||
print('Three-stage packages: scope, cutoff, grouping, reposts, revisions, corrections, peak risk, retained success, outlook and provider isolation passed')
|
||||
finally:
|
||||
async with SessionLocal() as session:
|
||||
for model in (RiskAnalysis, RiskEvent, RiskSnapshot, RiskSettings):
|
||||
for model in (PackageEvidence, SummaryVersion, SummaryArticle, NewsSummary, ArticleAnalysis, RiskSnapshot, RiskSettings):
|
||||
await session.execute(delete(model))
|
||||
await session.execute(delete(NewsItem).where(NewsItem.external_id.startswith('ci-risk-')))
|
||||
await session.execute(delete(NewsItem).where(NewsItem.external_id.startswith('ci-news-')))
|
||||
await session.execute(delete(AiProvider).where(AiProvider.id == provider_id))
|
||||
await session.execute(SkillNode.__table__.update().where(SkillNode.id.in_(skill_ids.values())).values(published_version_id=None))
|
||||
await session.execute(delete(SkillVersion).where(SkillVersion.skill_id.in_(skill_ids.values())))
|
||||
await session.execute(delete(SkillNode).where(SkillNode.id.in_(skill_ids.values())))
|
||||
await session.commit()
|
||||
await engine.dispose()
|
||||
|
||||
|
||||
asyncio.run(main())
|
||||
Reference in new issue
Block a user