This change set introduces a versioned result contract, shared config schema/loading, provider/data adapter seams, and a no-strategy application-service skeleton so the current research graph, orchestrator layer, and dashboard backend stop drifting further apart. It also keeps the earlier MiniMax compatibility and compact-prompt work aligned with the new contract shape and extends regression coverage so degradation, fallback, and service migration remain testable during the next phases.
Constraint: Must preserve existing FastAPI entrypoints and fallback behavior while introducing an application-service seam
Constraint: Must not turn application service into a new strategy or learning layer
Rejected: Full backend rewrite to service-only execution now | too risky before contract and fallback paths stabilize
Rejected: Leave provider/data/config logic distributed across scripts and endpoints | continues boundary drift and weakens verification
Confidence: high
Scope-risk: broad
Directive: Keep future application-service changes orchestration-only; move any scoring, signal fusion, or learning logic to orchestrator or tradingagents instead
Tested: python -m compileall orchestrator tradingagents web_dashboard/backend
Tested: python -m pytest orchestrator/tests/test_signals.py orchestrator/tests/test_llm_runner.py orchestrator/tests/test_quant_runner.py orchestrator/tests/test_contract_v1alpha1.py orchestrator/tests/test_application_service.py orchestrator/tests/test_provider_adapter.py web_dashboard/backend/tests/test_main_api.py web_dashboard/backend/tests/test_portfolio_api.py web_dashboard/backend/tests/test_api_smoke.py web_dashboard/backend/tests/test_services_migration.py -q
Not-tested: live MiniMax/provider execution against external services
Not-tested: full dashboard/manual websocket flow against a running frontend
Not-tested: omx team runtime end-to-end in the primary workspace
* feat(dashboard): apply Apple design system to all 4 pages
- Font: replace SF Pro with DM Sans (web-available) throughout
- Typography: consistent DM Sans stack, monospace data display
- ScreeningPanel: add horizontal scroll for mobile, fix stat card hover
- AnalysisMonitor: Apple progress bar, stage pills, decision badge
- BatchManager: add copy-to-clipboard for task IDs, fix error tooltip truncation, add CTA to empty state
- ReportsViewer: Apple-styled modal, search bar consistency
- Keyboard: add Escape to close modals
- CSS: progress bar ease-out, sidebar collapse button icon-only mode
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(dashboard): secure API key handling and add stage progress streaming
- Pass ANTHROPIC_API_KEY via env dict instead of CLI args (P1 security fix)
- Add monitor_subprocess() coroutine with fcntl non-blocking reads
- Inject STAGE markers (analysts/research/trading/risk/portfolio) into script stdout
- Update task stage state and broadcast WebSocket progress at each stage boundary
- Add asyncio.Event for monitor cancellation on task completion/cancel
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat(dashboard): persist task state to disk for restart recovery
- Add TASK_STATUS_DIR for task state JSON files
- Lifespan startup: restore task states from disk
- Task completion/failure: write state to disk
- Task cancellation: delete persisted state
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(dashboard): correct stage key mismatch, add created_at, persist cancelled tasks
- Fix ANALYSIS_STAGES key 'trader' → 'trading' to match backend STAGE markers
- Add created_at field to task state at creation, sort list_tasks by it
- Persist task state before broadcast in cancel path (closes restart race)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat(dashboard): add portfolio panel - watchlist, positions, and recommendations
New backend:
- api/portfolio.py: watchlist CRUD, positions with live P&L, recommendations
- POST /api/portfolio/analyze: batch analysis of watchlist tickers
- GET /api/portfolio/positions: live price from yfinance + unrealized P&L
New frontend:
- PortfolioPanel.jsx with 3 tabs: 自选股 / 持仓 / 今日建议
- portfolioApi.js service
- Route /portfolio (keyboard shortcut: 5)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat(dashboard): add CSV and PDF report export
- GET /api/reports/export: CSV with ticker,date,decision,summary
- GET /api/reports/{ticker}/{date}/pdf: PDF via fpdf2 with DejaVu fonts
- ReportsViewer: CSV export button + PDF export in modal footer
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(dashboard): address 4 critical issues found in pre-landing review
1. main.py: move API key validation before task state creation —
prevents phantom "running" tasks when ANTHROPIC_API_KEY is missing
2. portfolio.py: make get_positions() async and fetch yfinance prices
concurrently via run_in_executor — no longer blocks event loop
3. portfolio.py: add fcntl.LOCK_EX around all JSON read-modify-write
operations on watchlist.json and positions.json — eliminates TOCTOU
lost-write races under concurrent requests
4. main.py: use tempfile.mkstemp with mode 0o600 instead of world-
readable /tmp/analysis_{task_id}.py — script content no longer
exposed to other users on shared hosts
Also: remove unused UploadFile/File imports, undefined _save_to_cache
function, dead code in _delete_task_status, and unused
get_or_create_default_account helper.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(dashboard): use secure temp file for batch analysis scripts
Batch portfolio analysis was writing scripts to /tmp with default
permissions (0o644), exposing the API key to other local users.
Switch to tempfile.mkstemp + chmod 0o600, matching the single-analysis
pattern. Also fix cancel_task cleanup to use glob patterns for
tempfile-generated paths.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(dashboard): remove fake fallback data from ReportsViewer
ReportsViewer showed fabricated Chinese text when a report failed to load,
making fake data appear indistinguishable from real analysis. Now shows
an error message instead.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(dashboard): reliability fixes - cross-platform PDF fonts, API timeouts, yfinance concurrency, retry logic
- PDF: try multiple DejaVu font paths (macOS + Linux) instead of hardcoded macOS
- Frontend: add 15s AbortController timeout to all API calls + proper error handling
- yfinance: cap concurrent price fetches at 5 via asyncio.Semaphore
- Batch analysis: retry failed stock analyses up to 2x with exponential backoff
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: resolve 4 critical security/correctness bugs in web dashboard
1. Mass position deletion (portfolio.py): remove_position now rejects
empty position_id — previously position_id="" matched all positions
and deleted every holding for a ticker across ALL accounts.
2. Path traversal in get_recommendation (portfolio.py): added ticker/date
validation (no ".." or path separators) + resolved-path check against
RECOMMENDATIONS_DIR to prevent ../../etc/passwd attacks.
3. Path traversal in get_report_content (main.py): same ticker/date
validation + resolved-path check against get_results_dir().
4. china_data import stub (interface.py + new china_data.py): the actual
akshare implementation lives in web_dashboard/backend/china_data.py
(different package); tradingagents/dataflows/china_data.py was missing
entirely, so _china_data_available was always False. Added stub file
and AttributeError to the import exception handler so the module
gracefully degrades instead of silently hiding the missing vendor.
Magic numbers also extracted to named constants:
- MAX_RETRY_COUNT, RETRY_BASE_DELAY_SECS (main.py)
- MAX_CONCURRENT_YFINANCE_REQUESTS (portfolio.py)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Apply review suggestions: use concise `or` pattern for API key
resolution, consolidate tests into parameterized subTest, move
import to module level per PEP 8.
GoogleClient now accepts the unified `api_key` parameter used by
OpenAI and Anthropic clients, mapping it to the provider-specific
`google_api_key` that ChatGoogleGenerativeAI expects. Legacy
`google_api_key` still works for backward compatibility.
Resolves TODO.md item #2 (inconsistent parameter handling).
Add effort parameter (high/medium/low) for Claude 4.5+ and 4.6 models,
consistent with OpenAI reasoning_effort and Google thinking_level.
Also add content normalization for Anthropic responses.
- Point requirements.txt to pyproject.toml as single source of truth
- Resolve welcome.txt path relative to module for CLI portability
- Include cli/static files in package build
- Extract shared normalize_content for OpenAI Responses API and
Gemini 3 list-format responses into base_client.py
- Update README install and CLI usage instructions
Enable use_responses_api for native OpenAI provider, which supports
reasoning_effort with function tools across all model families.
Removes the UnifiedChatOpenAI subclass workaround.
Closes#403
- Add http_client and http_async_client parameters to all LLM clients
- OpenAIClient, GoogleClient, AnthropicClient now support custom httpx clients
- Fixes SSL certificate verification errors on Windows Conda environments
- Users can now pass custom httpx.Client with verify=False or custom certs
Fixes#369
- OpenAI: add GPT-5.4, GPT-5.4 Pro; remove o-series and legacy GPT-4o
- Anthropic: add Claude Opus 4.6, Sonnet 4.6; remove legacy 4.1/4.0/3.x
- Google: add Gemini 3.1 Pro, 3.1 Flash Lite; remove deprecated
gemini-3-pro-preview and Gemini 2.0 series
- xAI: clean up model list to match current API
- Simplify UnifiedChatOpenAI GPT-5 temperature handling
- Add missing tradingagents/__init__.py (fixes pip install building)
Add _clean_dataframe() to normalize stock DataFrames before stockstats:
coerce invalid dates/prices, drop rows missing Close, fill price gaps.
Also add on_bad_lines="skip" to all cached CSV reads.
LLMs (especially smaller models) sometimes pass multiple indicator
names as a single comma-separated string instead of making separate
tool calls. Split and process each individually at the tool boundary.
InvestDebateState was missing bull_history, bear_history, judge_decision.
RiskDebateState was missing aggressive_history, conservative_history,
neutral_history, latest_speaker, judge_decision. This caused KeyError
in _log_state() and reflection, especially with edge-case config values.
Prevents UnicodeEncodeError on Windows where the default encoding
(cp1252/gbk) cannot handle Unicode characters in LLM output.
Closes#77, closes#114, closes#126, closes#215, closes#332
- Add StatsCallbackHandler for tracking LLM calls, tool calls, and tokens
- Integrate callbacks into TradingAgentsGraph and all LLM clients
- Dynamic agent/report counts based on selected analysts
- Fix report completion counting (tied to agent completion)
陈少杰
Shaojie
Yijia-Xiao
Yijia Xiao
CadeYu
javierdejesusda
阳虎
makk9
Ljx-007
Edward Sun
luohy15