root/Jyotisha
1
0
Fork 0
Code Issues 1 Pull Requests Actions Packages Projects Releases Wiki Activity

Harden Jyotish runtime evidence gates

Browse Source
This commit is contained in:
732642856
2026-07-05 18:27:33 +08:00
parent 8520921f40
commit 4764482c73
76 changed files with 5628 additions and 253 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/test_api_server_security.py
+255 -2
View File
@@ -387,7 +387,10 @@ def test_high_rigor_workflow_plan_only_exposes_official_hard_override_contract()
assert result['execution_plan'][-1] == 'return_official_primary_supplemental_fallback_conflict_contract'
def test_high_rigor_vedastro_official_summary_passes_through_contract_fields() -> None:
def test_high_rigor_vedastro_official_summary_passes_through_contract_fields(monkeypatch) -> None:
monkeypatch.delenv('VEDASTRO_FREE_TIER_QUEUE', raising=False)
monkeypatch.delenv('VEDASTRO_FREE_TIER_QUEUE_ENABLED', raising=False)
monkeypatch.delenv('VEDASTRO_ENABLE_FREE_TIER_QUEUE', raising=False)
handler = _handler()
chart = {
@@ -428,6 +431,17 @@ def test_high_rigor_vedastro_official_summary_passes_through_contract_fields() -
assert result['fallback_used'] == ['local_chart_fallback']
assert result['blocked_items'] == ['official_event_radar_partial']
assert result['conflicts'] == [{'type': 'official_local_dasha_conflict'}]
runtime_truth = result['runtime_truth']
assert runtime_truth['catalog_boundary'] == 'catalog_recognized_not_full_runtime_execution'
assert runtime_truth['official_execution_layers']['chart_core'] == 'ok'
assert runtime_truth['official_execution_layers']['event_radar'] == 'partial'
assert runtime_truth['official_execution_layers']['catalog_status'] == 'partial'
assert runtime_truth['fallback_active'] is True
assert runtime_truth['blocked_items'] == ['official_event_radar_partial']
assert runtime_truth['free_tier_strategy']['using_free_tier'] is True
assert runtime_truth['free_tier_strategy']['queue_enabled'] is False
assert runtime_truth['free_tier_strategy']['cache_hit'] is False
assert runtime_truth['free_tier_strategy']['guard_status'] == 'degraded_or_partial'
def test_high_rigor_vedastro_official_summary_exposes_top_reader_contract_from_full_snapshot() -> None:
@@ -481,6 +495,7 @@ def test_high_rigor_vedastro_official_summary_exposes_top_reader_contract_from_f
'summary': {'catalog_method_count': 641},
}
},
'raw_response': {'official': 'raw'},
},
},
'ai_prompt_pack': {
@@ -534,6 +549,49 @@ def test_high_rigor_vedastro_official_summary_exposes_top_reader_contract_from_f
assert result['verdict'] == 'high_probability_window'
assert result['dominant_label'] == 'career_status'
assert result['main_conflicts'] == [{'type': 'official_local_dasha_conflict'}]
assert result['runtime_truth']['primary_route'] == 'career'
assert result['runtime_truth']['routes_available'] == ['career', 'relationship', 'finance']
assert result['raw_response'] == {'official': 'raw'}
def test_high_rigor_vedastro_official_summary_uses_module_snapshot_cache_truth_when_range_scan_snapshot_missing() -> None:
handler = _handler()
chart = {
'birth': {'ayanamsa_display': 'Lahiri', 'ayanamsa_name': 'lahiri', 'node_mode': 'mean'},
'modules': {
'vedastro_range_scan_result': {
'status': 'ok',
'source_metadata': {},
},
'vedastro_official_full_snapshot': {
'status': 'official_snapshot_budget_exhausted',
'strict_workflow_primary_route': 'career',
'strict_workflow_routes_available': ['career'],
'strict_workflow_contracts': {
'career': {
'official_primary_evidence': {'chart_core': {'status': 'ok'}},
},
},
'source_metadata': {
'semantic_cache': {
'cache_hit': True,
},
},
},
},
'ai_prompt_pack': {
'evidence_snapshot': {
'vedastro_official_snapshot': {
'status': 'official_snapshot_budget_exhausted',
},
},
},
}
result = handler._high_rigor_vedastro_official_summary(chart)
assert result['status'] == 'official_snapshot_budget_exhausted'
assert result['runtime_truth']['free_tier_strategy']['cache_hit'] is True
def test_api_prompt_pack_official_snapshot_carries_strict_workflow_contracts() -> None:
@@ -708,6 +766,7 @@ def test_consultation_workflow_surfaces_top_reader_contract_in_official_summary(
'lon': 114.2,
'tz': 8,
'theme': ['career'],
'blind': True,
})
contract = result['vedastro_official']['strict_workflow_contracts']['career']
@@ -2127,6 +2186,7 @@ def test_consultation_workflow_uses_unified_orchestrator_contract(monkeypatch) -
'lon': 114.2,
'tz': 8,
'theme': ['career'],
'blind': True,
})
assert result['success'] is True
@@ -2144,8 +2204,24 @@ def test_consultation_workflow_uses_unified_orchestrator_contract(monkeypatch) -
'run_rectification_gate',
'run_thematic_report',
]
assert result['runtime_planner']['skipped_steps'] == ['run_historical_event_backtest']
assert 'run_historical_event_backtest' in result['runtime_planner']['skipped_steps']
assert result['source_priority']['mode'] == 'vedastro_official_snapshot_first'
assert result['runtime_evidence_log']['surface'] == 'api_web'
assert result['runtime_evidence_log']['route']['question_type'] == 'career'
assert result['runtime_evidence_log']['vedastro_cloud_state'] in {
'official_verified',
'official_blocked',
'local_fallback',
}
assert result['runtime_evidence_log']['quality_gate']['technique_audit_table_required'] is True
assert result['runtime_evidence_log']['quality_gate']['technique_audit_table'][0]['technique'] == 'VedAstro Cloud State'
assert result['machine_evidence_packet']['status'] == 'partial'
assert result['real_case_calibration']['status'] == 'partial_scored'
assert result['real_case_calibration']['batch_id'] == 'real_case_studies_batch1'
assert result['runtime_evidence_log']['evidence_packet_contract']['status'] == 'partial'
assert result['runtime_evidence_log']['real_case_calibration']['status'] == 'partial_scored'
assert result['runtime_evidence_log']['blind_technical_mode']['enabled'] is True
assert 'conversation_feedback' in result['runtime_evidence_log']['blind_technical_mode']['disallowed_sources']
assert result['chart']['special_lagnas']['precision'] == 'sunrise_correct'
@@ -2345,6 +2421,183 @@ def test_consultation_workflow_rectification_entry_sends_empty_objects_before_ch
assert seen['rectification_body']['ascendant'] == {}
def test_consultation_workflow_prashna_entry_uses_prashna_without_compute_chart(monkeypatch) -> None:
handler = _handler()
calls = {'chart': 0, 'prashna': 0}
seen = {}
def fake_chart_compute(body):
calls['chart'] += 1
return {'success': True, 'modules': {}}
def fake_prashna(body):
calls['prashna'] += 1
seen['prashna_body'] = dict(body)
return {
'success': True,
'endpoint': 'prashna',
'question': body.get('question'),
'timing': {'recommendation': '可以进行Prashna分析'},
'judgement': {'summary': '可问'},
}
def fake_thematic_report(body):
seen['theme_body'] = dict(body)
return {
'success': True,
'endpoint': 'thematic_report',
'mode': 'upstream_contract_reuse',
'report': {'sections': []},
}
monkeypatch.setattr(handler, '_compute_chart', fake_chart_compute)
monkeypatch.setattr(handler, '_compute_prashna', fake_prashna)
monkeypatch.setattr(handler, '_compute_thematic_report', fake_thematic_report)
monkeypatch.setattr(handler, '_compute_rectification_gate', lambda body: {'success': True, 'summary': {'recommended_events': []}})
result = handler._compute_consultation_workflow({
'entry_mode': 'prashna',
'question': '这个合作能成吗',
'question_text': '这个合作能成吗',
'theme': ['wealth'],
'year': REDACTED_YEAR,
'month': 4,
'day': 17,
'hour': 14,
'minute': 49,
'lat': 36.42,
'lon': 114.2,
'tz': 8,
})
assert result['success'] is True
assert result['entry_mode'] == 'prashna'
assert result['runtime_planner']['entry_mode'] == 'prashna'
assert result['runtime_planner']['executed_steps'] == ['run_prashna', 'run_thematic_report']
assert calls['chart'] == 0
assert calls['prashna'] == 1
assert seen['prashna_body']['question'] == '这个合作能成吗'
def test_consultation_workflow_builds_audited_remedies_from_guided_topic_gate(monkeypatch) -> None:
handler = _handler()
fake_chart = {
'success': True,
'birth_info': {'date': 'REDACTED_DATE', 'time': 'REDACTED_TIME', 'tz': 8},
'ascendant': {'lon': 92.0, 'sign': 'Cancer'},
'planets': _sample_planets(),
'chart': {
'ascendant': {'lon': 92.0, 'sign': 'Cancer'},
'planets': _sample_planets(),
},
'modules': {
'guided_topics': [
{
'id': 'career',
'title': '事业',
'strict_audit_gate': {
'topic': 'career',
'primary_planets': ['Saturn'],
'active_dasha_lord': 'Saturn',
'strength_context': {
'Saturn': {'total_rupas': 0.42, 'strength_level': 'weak'},
},
'dosha_context': ['delay_signature'],
},
},
],
},
'special_lagnas': {'precision': 'sunrise_correct'},
}
monkeypatch.setattr(handler, '_compute_chart', lambda body: fake_chart)
monkeypatch.setattr(handler, '_compute_rectification_gate', lambda body: {'success': True, 'summary': {'recommended_events': []}})
monkeypatch.setattr(handler, '_compute_thematic_report', lambda body: {'success': True, 'endpoint': 'thematic_report', 'report': {'sections': []}})
result = handler._compute_consultation_workflow({
'entry_mode': 'direct_chart',
'question': '请直接排盘并看事业',
'theme': ['career'],
'year': REDACTED_YEAR,
'month': 4,
'day': 17,
'hour': 14,
'minute': 49,
'lat': 36.42,
'lon': 114.2,
'tz': 8,
})
remedies = result.get('audited_remedies') or {}
assert remedies['status'] == 'ok'
assert remedies['source'] == 'strict_audit_gate'
assert remedies['topic'] == 'career'
assert remedies['active_dasha_lord'] == 'Saturn'
def test_consultation_workflow_timing_route_builds_muhurta_panchanga_sidecar(monkeypatch) -> None:
handler = _handler()
fake_chart = {
'success': True,
'birth_info': {'date': 'REDACTED_DATE', 'time': 'REDACTED_TIME', 'tz': 8},
'ascendant': {'lon': 92.0, 'sign': 'Cancer'},
'planets': _sample_planets(),
'chart': {
'ascendant': {'lon': 92.0, 'sign': 'Cancer'},
'planets': _sample_planets(),
},
'modules': {},
'special_lagnas': {'precision': 'sunrise_correct'},
}
seen = {}
monkeypatch.setattr(handler, '_compute_chart', lambda body: fake_chart)
monkeypatch.setattr(handler, '_compute_rectification_gate', lambda body: {
'success': True,
'endpoint': 'rectification_gate',
'summary': {'recommended_events': []},
})
monkeypatch.setattr(handler, '_compute_thematic_report', lambda body: {
'success': True,
'endpoint': 'thematic_report',
'mode': 'derived_chart_evidence',
'theme_count': len(body.get('theme') or []),
})
def fake_muhurta(body):
seen['muhurta_body'] = dict(body)
return {
'status': 'ok',
'source': 'local_muhurta.py',
'activity': 'business',
'report_mode': 'muhurta_date_range_solver',
'panchanga': {'query_date': '2026-07-08'},
'best_windows': [{'date': '2026-07-08'}],
}
monkeypatch.setattr(handler, '_compute_muhurta_panchanga', fake_muhurta)
result = handler._compute_consultation_workflow({
'entry_mode': 'direct_chart',
'question': '2026年何时适合谈合作和推进项目的应期',
'year': REDACTED_YEAR,
'month': 4,
'day': 17,
'hour': 14,
'minute': 49,
'lat': 36.42,
'lon': 114.2,
'tz': 8,
'reference_date': '2026-07-08',
'theme': ['career'],
})
assert result['success'] is True
assert 'run_muhurta_panchanga' in result['runtime_planner']['executed_steps']
assert result['muhurta_panchanga']['status'] == 'ok'
assert result['muhurta_panchanga']['activity'] == 'business'
assert seen['muhurta_body']['reference_date'] == '2026-07-08'
def test_thematic_report_handles_missing_dasa_convergence_without_crash(monkeypatch) -> None:
handler = _handler()