TradingAgents/CHANGELOG.md

Changelog

All notable changes to TradingAgents will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Added

• Trade model for execution history with CGT tracking (Issue #6: DB-5)

  • Trade model with BUY/SELL sides and execution status tracking (PENDING, FILLED, PARTIAL, CANCELLED, REJECTED) file:tradingagents/api/models/trade.py
  • TradeSide, TradeStatus, TradeOrderType enums for type-safe trade operations file:tradingagents/api/models/trade.py:86-137
  • Order type support (MARKET, LIMIT, STOP, STOP_LIMIT) with signal source and confidence tracking
  • Capital Gains Tax (CGT) support for Australian tax compliance file:tradingagents/api/models/trade.py:201-305
  • 50% CGT discount eligibility for holdings >12 months (cgt_discount_eligible field)
  • Australian financial year (FY) calculation (July-June) via tax_year property file:tradingagents/api/models/trade.py:418-441
  • CGT gain/loss tracking with gross_gain, gross_loss, and net_gain after discount
  • Multi-currency support with FX rate to AUD conversion file:tradingagents/api/models/trade.py:306-325
  • High-precision decimal arithmetic (19,4) for monetary values and (19,8) for FX rates
  • Automatic acquisition_date and timestamp management via default values and validators
  • Check constraints for positive values (quantity, price, total_value, fx_rate_to_aud)
  • Signal confidence validation (0-100 range) and holding period non-negative constraint
  • Many-to-one relationship with Portfolio model with cascade delete behavior file:tradingagents/api/models/portfolio.py:202-205
  • Properties for convenient trade type checking (is_buy, is_sell, is_filled) file:tradingagents/api/models/trade.py:443-475
  • Comprehensive validators for enum normalization (side, status, order_type) and symbol/currency normalization file:tradingagents/api/models/trade.py:477-585
  • Event listener validation (before_flush) for cross-field business rule enforcement file:tradingagents/api/models/trade.py:596-665
  • Composite indexes for efficient queries: (portfolio_id, symbol) and (portfolio_id, side) and (status, executed_at)
  • Database migration 005_add_trade_model.py with comprehensive schema definition file:migrations/versions/005_add_trade_model.py
  • Migration with proper upgrade and downgrade functions for reversible schema changes
  • Comprehensive unit test suite covering field validation, defaults, constraints, enums, and CGT calculations file:tests/unit/api/test_trade_model.py (2054 lines, 65 tests)
  • Integration test suite for relationships, cascade delete, and concurrent operations file:tests/integration/api/test_trade_integration.py (1235 lines, 22 tests)
  • Comprehensive docstrings with examples for all Trade model methods and fields
  • Total: 87 tests added for Trade model feature

• FastAPI backend with JWT authentication and strategies CRUD (Issue #48)

  • FastAPI application with async/await support and health check endpoints file:tradingagents/api/main.py
  • JWT authentication with asymmetric RS256 signing algorithm file:tradingagents/api/services/auth_service.py
  • Argon2 password hashing with automatic salt generation for secure credential storage
  • POST /api/v1/auth/login endpoint with username/password authentication returning JWT tokens
  • GET /api/v1/strategies endpoint with pagination, user isolation, and permission-based access control
  • POST /api/v1/strategies endpoint for creating new strategies with JSON parameters support
  • GET /api/v1/strategies/{id} endpoint for retrieving individual strategies with authorization checks
  • PUT /api/v1/strategies/{id} endpoint for updating strategy metadata and parameters
  • DELETE /api/v1/strategies/{id} endpoint for removing strategies with proper cascade behavior
  • SQLAlchemy ORM with async PostgreSQL/SQLite support file:tradingagents/api/models/
  • User model with hashed passwords, email uniqueness, and active status tracking file:tradingagents/api/models/user.py
  • Strategy model with JSON parameters, description, user association, and active/inactive toggling file:tradingagents/api/models/strategy.py
  • Alembic migration system with version control for database schema changes file:migrations/
  • Initial migration creating users and strategies tables with proper constraints file:migrations/versions/
  • Database configuration with environment variable support (DATABASE_URL, SQLALCHEMY_ECHO) file:tradingagents/api/config.py
  • Pydantic schemas for request validation and response serialization file:tradingagents/api/schemas/
  • CORS middleware configuration with environment-based allowed origins
  • Error handling middleware with consistent JSON error responses and proper HTTP status codes
  • Request logging middleware with sanitized credential exclusion and request ID tracking
  • Comprehensive test suite with 208 tests covering authentication, strategies CRUD, models, migrations, middleware, and configuration file:tests/api/
  • API-focused fixtures with async SQLAlchemy session, FastAPI test client, and test user/strategy data file:tests/api/conftest.py
  • Security tests covering SQL injection prevention, XSS payload handling, JWT tampering detection, and rate limiting
  • Integration tests for endpoint authorization, user isolation, pagination, and cascade operations
  • Migration tests validating schema constraints, rollback behavior, and Alembic configuration
  • New dependencies in pyproject.toml: fastapi, uvicorn, sqlalchemy, alembic, pydantic-settings, passlib, argon2-cffi, python-multipart, python-jose, cryptography
  • API documentation generated from FastAPI OpenAPI schema (available at /docs and /redoc)

• FRED API integration for economic data (Issue #8: DATA-7)

  • Federal Reserve Economic Data (FRED) API wrapper module with core utilities file:tradingagents/dataflows/fred_common.py (346 lines)
  • Data retrieval functions module with high-level economic data access file:tradingagents/dataflows/fred.py (396 lines)
  • Custom exceptions for API rate limiting and invalid series errors file:tradingagents/dataflows/fred_common.py:52-67
  • FredRateLimitError exception with retry_after metadata for client-side rate limit handling
  • FredInvalidSeriesError exception with series_id tracking for debugging invalid requests
  • Retry logic with exponential backoff for transient API failures file:tradingagents/dataflows/fred_common.py:146-250
  • get_api_key() function for secure FRED_API_KEY environment variable retrieval file:tradingagents/dataflows/fred_common.py:74-83
  • Date formatting utility for converting various date formats to FRED-compatible YYYY-MM-DD file:tradingagents/dataflows/fred_common.py:90-144
  • Request wrapper _make_fred_request() with retry logic and exponential backoff (up to 3 attempts with 1-2-4s delays)
  • Local file caching with 24-hour TTL to reduce API quota consumption file:tradingagents/dataflows/fred_common.py:42-48
  • Interest rates function get_interest_rates() for Federal Funds Rate data file:tradingagents/dataflows/fred.py:104-142
  • Treasury rates function get_treasury_rates() for 2Y/5Y/10Y/30Y yield curves file:tradingagents/dataflows/fred.py:143-185
  • Money supply functions get_money_supply() for M1/M2 monetary aggregates file:tradingagents/dataflows/fred.py:186-228
  • Economic growth function get_gdp() for nominal and real GDP data file:tradingagents/dataflows/fred.py:229-271
  • Inflation measurement functions get_inflation() for CPI and PCE indices file:tradingagents/dataflows/fred.py:272-314
  • Labor market function get_unemployment() for unemployment rate data file:tradingagents/dataflows/fred.py:315-352
  • Generic series function get_fred_series() for accessing any FRED series by ID file:tradingagents/dataflows/fred.py:353-396
  • All functions return pandas DataFrames with 'date' and 'value' columns on success, error strings on failure
  • Optional date range filtering (start_date, end_date) in YYYY-MM-DD format
  • Automatic caching control via use_cache parameter
  • Comprehensive docstrings with examples for all functions and utilities
  • Unit test suite for core utilities covering caching, retry logic, and date formatting file:tests/unit/dataflows/test_fred_common.py (594 lines, 40 tests)
  • Unit test suite for data retrieval functions covering all economic data functions file:tests/unit/dataflows/test_fred.py (634 lines, 42 tests)
  • Integration test suite for FRED API with live endpoint testing file:tests/integration/dataflows/test_fred_integration.py (560 lines, 26 tests)
  • Test coverage including rate limit handling, caching behavior, and date range filtering
  • Total: 108 tests added for FRED API feature

• Benchmark data retrieval and analysis (Issue #10)

  • Benchmark data module for SPY, sector ETF, and analysis functions file:tradingagents/dataflows/benchmark.py (441 lines)
  • Sector ETF mappings for all 11 SPDR sector funds (communication, consumer discretionary/staples, energy, financials, healthcare, industrials, materials, real estate, technology, utilities) file:tradingagents/dataflows/benchmark.py:48-59
  • get_benchmark_data() function for fetching OHLCV data via yfinance with date validation file:tradingagents/dataflows/benchmark.py:67-115
  • get_spy_data() convenience wrapper for S&P 500 benchmark data file:tradingagents/dataflows/benchmark.py:117-136
  • get_sector_etf_data() function for retrieving sector-specific benchmark data with sector validation file:tradingagents/dataflows/benchmark.py:138-186
  • calculate_relative_strength() function with IBD-style weighted ROC formula file:tradingagents/dataflows/benchmark.py:188-285
  • Relative strength calculation using weighted periods (40% 63-day, 20% 126-day, 20% 189-day, 20% 252-day ROC)
  • Customizable ROC periods with default IBD-style weighting
  • Data alignment via inner join with validation for overlapping dates
  • calculate_rolling_correlation() function for time-series correlation analysis file:tradingagents/dataflows/benchmark.py:287-349
  • Configurable rolling window sizes with default 60-day window
  • Comprehensive validation for data alignment and minimum data requirements
  • calculate_beta() function for volatility and systematic risk measurement file:tradingagents/dataflows/benchmark.py:351-441
  • Beta calculation using covariance-variance approach with optional smoothing
  • Optional rolling beta calculation with customizable window (default 252 days)
  • Markdown rolling window implementation for efficient computation
  • All functions return DataFrames/Series/floats on success, error strings on failure
  • Comprehensive error handling with descriptive messages and validation logic
  • Comprehensive docstrings with examples for all public functions
  • Unit test suite for benchmark functions file:tests/unit/dataflows/test_benchmark.py (753 lines, 28 tests)
  • Integration test suite for benchmark workflows file:tests/integration/dataflows/test_benchmark_integration.py (593 lines, 7 tests)
  • Test coverage includes data fetching, sector validation, relative strength calculation, correlation analysis, and beta calculation
  • Total: 35 tests added for benchmark data feature

• Multi-timeframe OHLCV aggregation functions (Issue #9)

  • Multi-timeframe aggregation module for daily to weekly/monthly resampling file:tradingagents/dataflows/multi_timeframe.py (320 lines)
  • Core OHLCV aggregation validation function _validate_ohlcv_dataframe() file:tradingagents/dataflows/multi_timeframe.py:38-75
  • Core resampling function _resample_ohlcv() with proper aggregation rules file:tradingagents/dataflows/multi_timeframe.py:78-139
  • OHLCV aggregation rules: Open=first, High=max, Low=min, Close=last, Volume=sum file:tradingagents/dataflows/multi_timeframe.py:105-111
  • Weekly aggregation function aggregate_to_weekly() with configurable week anchor file:tradingagents/dataflows/multi_timeframe.py:142-220
  • Support for Sunday/Monday week anchors with automatic day-of-week mapping file:tradingagents/dataflows/multi_timeframe.py:180-201
  • Month boundary handling for weeks spanning multiple months
  • Monthly aggregation function aggregate_to_monthly() with period end/start labeling file:tradingagents/dataflows/multi_timeframe.py:223-320
  • Support for month-end (ME) and month-start (MS) labels via period_end parameter
  • Timezone preservation for both UTC and localized datetime indices
  • Input validation with descriptive error messages for missing/invalid data
  • OHLCV value rounding to 2 decimal places for data consistency
  • Return type: DataFrame on success, error string on validation failure
  • Comprehensive docstrings with examples for all public functions
  • Unit test suite for aggregation functions file:tests/unit/dataflows/test_multi_timeframe.py (29 tests)
  • Integration test suite for multi-timeframe workflows file:tests/integration/dataflows/test_multi_timeframe_integration.py (13 tests)
  • Test coverage includes validation, weekly aggregation (multiple anchors), monthly aggregation, timezone handling, and partial period aggregation
  • Total: 42 tests added for multi-timeframe aggregation feature

• User model enhancement with profile and API key management (Issue #3)

  • Extended User model with tax_jurisdiction and timezone fields file:tradingagents/api/models/user.py:47-54
  • Tax jurisdiction field supporting country (e.g., "US", "AU") and state/province level codes (e.g., "US-CA", "AU-NSW")
  • IANA timezone identifier field (e.g., "America/New_York", "Australia/Sydney") with automatic validation
  • Email verification status tracking via is_verified boolean field file:tradingagents/api/models/user.py:60-64
  • Secure API key management with bcrypt hashing and unique constraints file:tradingagents/api/models/user.py:55-59
  • API key service module with generate_api_key(), hash_api_key(), and verify_api_key() functions file:tradingagents/api/services/api_key_service.py
  • API key generation using secrets.token_urlsafe() with 256-bit entropy and 'ta_' prefix
  • Bcrypt-based API key hashing using pwdlib.PasswordHash for secure storage
  • Constant-time verification to prevent timing attacks on API keys
  • Timezone validator using IANA zoneinfo database file:tradingagents/api/services/validators.py:134-191
  • Tax jurisdiction validator supporting 50+ country codes and state/province subdivisions file:tradingagents/api/services/validators.py:193-283
  • Utility functions get_available_timezones() and get_available_tax_jurisdictions() for UI dropdowns file:tradingagents/api/services/validators.py:285-333
  • Database migration 002_add_user_profile_fields.py with proper defaults and constraints file:migrations/versions/002_add_user_profile_fields.py
  • Migration rollback support for reversible schema changes
  • Comprehensive docstrings and security considerations for all functions

• Portfolio model for managing trading portfolios (Issue #4: DB-3)

  • Portfolio model with support for LIVE, PAPER, and BACKTEST portfolio types file:tradingagents/api/models/portfolio.py
  • PortfolioType enum for type-safe portfolio categorization file:tradingagents/api/models/portfolio.py:95-107
  • PreciseNumeric custom SQLAlchemy type decorator for Decimal(19,4) precision file:tradingagents/api/models/portfolio.py:63-92
  • High-precision monetary value handling (19 total digits, 4 decimal places) for initial_capital and current_value fields
  • Automatic current_value default to initial_capital on portfolio creation
  • Unique constraint on (user_id, name) to prevent duplicate portfolio names per user
  • Check constraints for non-negative capital and value amounts
  • Composite indexes for efficient queries: (user_id, is_active) and (user_id, portfolio_type)
  • Relationship to User model with cascade delete behavior
  • Currency code field with ISO 4217 validation (3-letter codes, e.g., AUD, USD)
  • Portfolio status tracking via is_active boolean field with default True
  • Comprehensive validators for currency normalization, portfolio type validation, and business rule enforcement file:tradingagents/api/models/portfolio.py:249-296
  • Event listener validation (before_flush) for cross-field and database constraint checks file:tradingagents/api/models/portfolio.py:300-347
  • User model updated with portfolios relationship for one-to-many portfolio association file:tradingagents/api/models/user.py:72-76
  • Database migration 003_add_portfolio_model.py with comprehensive schema definition file:migrations/versions/003_add_portfolio_model.py
  • Migration with proper upgrade and downgrade functions for reversible schema changes
  • Comprehensive unit and integration test suites file:tests/unit/api/test_portfolio_model.py file:tests/integration/api/test_portfolio_integration.py
  • Unit tests covering field validation, defaults, constraints, and enum handling
  • Integration tests for relationships, cascade delete, and concurrent operations

• Settings model for user preferences and risk management (Issue #5: DB-4)

  • Settings model for managing user trading preferences, risk tolerance, and alert configurations file:tradingagents/api/models/settings.py
  • RiskProfile enum for user risk tolerance profiles (CONSERVATIVE, MODERATE, AGGRESSIVE) file:tradingagents/api/models/settings.py:70-82
  • One-to-one relationship with User model with cascade delete behavior
  • Risk management fields: risk_score (0-10 scale), risk_profile, max_position_pct (0-100), max_portfolio_risk_pct (0-100)
  • Investment horizon field in years for long-term portfolio planning
  • Alert preferences JSON field for configurable email/SMS/push notifications
  • Unique constraint on user_id enforcing one settings record per user
  • Check constraints for valid numeric ranges: risk_score 0-10, position/portfolio risk 0-100, horizon >= 0
  • Automatic timestamps via TimestampMixin (created_at, updated_at)
  • Validators for risk_profile enum normalization and type safety file:tradingagents/api/models/settings.py:246-279
  • Event listener validation (before_flush) for business rule enforcement file:tradingagents/api/models/settings.py:284-313
  • User model updated with settings relationship for one-to-one association file:tradingagents/api/models/user.py
  • Database migration 004_add_settings_model.py with comprehensive schema definition file:migrations/versions/004_add_settings_model.py
  • Migration with proper upgrade and downgrade functions for reversible schema changes
  • Comprehensive unit test suite covering field validation, defaults, constraints, and enum handling file:tests/unit/api/test_settings_model.py
  • Integration test suite for relationships, cascade delete, and concurrent operations file:tests/integration/api/test_settings_integration.py

• Test fixtures directory with centralized mock data (Issue #51)

  • FixtureLoader class for loading JSON fixtures with automatic datetime parsing file:tests/fixtures/init.py
  • Stock data fixtures: US market OHLCV, Chinese market OHLCV, standardized data file:tests/fixtures/stock_data/
  • Metadata fixtures: Complete analysis, partial analysis, batch analysis, error scenarios file:tests/fixtures/metadata/analysis_metadata.json
  • Report section fixtures: Market, sentiment, news, fundamentals, investment plans file:tests/fixtures/report_sections/complete_reports.json
  • API response fixtures: OpenAI embeddings (single, batch, error responses) file:tests/fixtures/api_responses/openai_embeddings.json
  • Configuration fixtures: Vendor-specific and LLM provider configurations file:tests/fixtures/configurations/default_config.json
  • Comprehensive FixtureLoader class with convenience functions and edge case support
  • DataFrame conversion utilities for stock data with automatic datetime indexing
  • UTF-8 encoding support for Chinese market data with column standardization
  • Updated docs/testing/writing-tests.md with fixtures section and usage examples file:docs/testing/writing-tests.md
  • Fixture-based testing best practices documentation

• DeepSeek API support for LLM provider integration (Issue #41)

  • DeepSeek provider integration using ChatOpenAI with base_url pointing to DeepSeek API file:tradingagents/graph/trading_graph.py:105-145
  • DEEPSEEK_API_KEY environment variable handling with validation and helpful error messages
  • Support for DeepSeek models: deepseek-chat and deepseek-reasoner with custom attribution headers
  • Embedding fallback chain for providers without native embeddings (OpenAI -> HuggingFace -> disable memory) file:tradingagents/agents/utils/memory.py:16-57
  • Optional HuggingFace sentence-transformers integration (all-MiniLM-L6-v2 model) for offline embeddings
  • Graceful degradation with informative warnings when embedding backends unavailable
  • Comprehensive test suite for DeepSeek integration file:tests/integration/test_deepseek.py

• Test directory restructuring into unit/integration/e2e (Issue #50)

  • Organized tests into unit/, integration/, and e2e/ subdirectories by test type
  • Unit tests (5 files) - Fast, isolated tests: conftest_hierarchy, documentation_structure, exceptions, logging_config, report_exporter
  • Integration tests (3 files) - Component interaction tests: akshare, cli_error_handling, openrouter
  • End-to-end tests - Complete workflow tests with dedicated e2e/README.md guidelines
  • Hierarchical conftest.py structure for each test directory with type-specific fixtures
  • Updated pytest.ini with test discovery paths (tests, tests/unit, tests/integration, tests/e2e)
  • Custom markers registered: unit, integration, e2e, llm, chromadb, slow, requires_api_key
  • Updated docs/testing/README.md with new directory structure diagram and fixture organization
  • Improved test isolation with directory-specific fixtures and configurations

• pytest conftest.py hierarchy for organized test fixtures (Issue #49)

  • Root-level conftest.py with shared fixtures (environment variables, LangChain/ChromaDB mocking, configuration)
  • Unit-level conftest.py with data vendor mocking (akshare, yfinance, sample DataFrames)
  • Integration-level conftest.py with live ChromaDB and temporary directory fixtures
  • Fixture scope management (function, session, module) for test isolation and performance
  • Comprehensive docstrings for all fixtures with usage examples and scope documentation
  • pytest.ini configuration with custom markers (unit, integration, e2e, llm, chromadb, slow, requires_api_key)
  • Test suite validating fixture accessibility across test directories file:tests/test_conftest_hierarchy.py
  • Updated testing documentation with conftest.py hierarchy section file:docs/testing/README.md
  • Fixture usage examples in writing-tests.md file:docs/testing/writing-tests.md

• Comprehensive documentation structure (Issue #52)

  • Organized docs/ directory with structured documentation sections
  • Quick start guide at docs/QUICKSTART.md
  • Architecture documentation in docs/architecture/ (multi-agent-system, data-flow, llm-integration)
  • API reference documentation in docs/api/ (trading-graph, agents, dataflows)
  • Developer guides in docs/guides/ (adding-new-analyst, adding-llm-provider, adding-data-vendor, configuration)
  • Testing documentation in docs/testing/ (README, running-tests, writing-tests)
  • Development setup guide in docs/development/
  • Central documentation index at docs/README.md with navigation and key concepts
  • Updated PROJECT.md DOCUMENTATION MAP section to reference new docs/ structure
  • Added Documentation section to README.md with links to key guides

• Export reports to file with metadata (Issue #21)

  • YAML frontmatter formatting for report metadata file:tradingagents/utils/report_exporter.py:63-111
  • Report creation with combined YAML frontmatter and markdown content file:tradingagents/utils/report_exporter.py:112-136
  • Safe filename generation with date prefixes and sanitization file:tradingagents/utils/report_exporter.py:137-185
  • JSON metadata serialization with datetime handling and directory creation file:tradingagents/utils/report_exporter.py:186-220
  • Comprehensive report generation combining multiple sections with table of contents file:tradingagents/utils/report_exporter.py:221-325
  • Support for organizing report sections by team (Analyst, Research, Trading, Portfolio)
  • Datetime-to-ISO-string conversion for YAML/JSON serialization
  • Helper functions for basic YAML formatting when PyYAML is unavailable
  • Comprehensive test suite for all report export functions file:tests/test_report_exporter.py
  • Public API exports in utils/init.py for easy access

• Rate limit error handling for LLM APIs (Issue #39)

  • Unified exception hierarchy for handling rate limit errors across providers (OpenAI, Anthropic, OpenRouter) file:tradingagents/utils/exceptions.py
  • Dual-output logging configuration supporting both terminal and file outputs file:tradingagents/utils/logging_config.py
  • Automatic rotating log files with 5MB rotation and 3 backups
  • Terminal logging at INFO level and file logging at DEBUG level
  • API key sanitization in log messages to prevent credential leaks
  • Error recovery utilities for saving partial analysis state on errors file:tradingagents/utils/error_recovery.py
  • User-friendly error message formatting for rate limit errors file:tradingagents/utils/error_messages.py
  • Comprehensive test suite for exceptions and logging configuration file:tests/test_exceptions.py file:tests/test_logging_config.py

• AKShare data vendor integration for US and Chinese stock market data (Issue #16)

  • Unified AKShare vendor module with support for both US and Chinese markets file:tradingagents/dataflows/akshare.py
  • Date format conversion utility for YYYYMMDD compatibility file:tradingagents/dataflows/akshare.py:34-67
  • Exponential backoff retry mechanism with configurable attempts and delays file:tradingagents/dataflows/akshare.py:70-108
  • US stock data retrieval via get_akshare_stock_data_us() file:tradingagents/dataflows/akshare.py:114-211
  • Chinese stock data retrieval via get_akshare_stock_data_cn() file:tradingagents/dataflows/akshare.py:213-320
  • Auto-market detection with get_akshare_stock_data() for automatic routing file:tradingagents/dataflows/akshare.py:322-372
  • Rate limit error handling via AKShareRateLimitError exception with vendor fallback file:tradingagents/dataflows/akshare.py:28-30
  • Integration with interface.py vendor routing system file:tradingagents/dataflows/interface.py
  • Comprehensive test suite for all AKShare functions file:tests/test_akshare.py

• OpenRouter API provider support for unified access to multiple LLM models

  • Support for provider/model-name format (e.g., anthropic/claude-sonnet-4.5)
  • Proper API key handling with OPENROUTER_API_KEY environment variable
  • Custom headers for OpenRouter attribution (HTTP-Referer, X-Title)
  • Embedding fallback to OpenAI when using OpenRouter (since OpenRouter lacks embeddings)
  • Comprehensive test suite for OpenRouter provider integration file:tests/test_openrouter.py

• Expanded .env.example with all supported LLM provider API keys

• Detailed LLM Provider Options section in README.md with examples for:

  • OpenAI (default)
  • Anthropic
  • OpenRouter (new)
  • Google Generative AI
  • Ollama (local)

• OpenRouter configuration example in Python usage section

• Documentation updates in PROJECT.md for OpenRouter support

Changed

• Updated trading_graph.py LLM provider initialization to handle OpenRouter separately with proper API key and header management file:tradingagents/graph/trading_graph.py:75-105
• Enhanced memory.py embedding logic to support OpenRouter's embedding fallback behavior file:tradingagents/agents/utils/memory.py:6-27
• Main.py now includes OpenRouter configuration example (commented out) for easy reference

Fixed

• ChromaDB collection persistence issue by using get_or_create_collection() instead of create_collection() to prevent "collection already exists" errors and enable persistent memory across application restarts file:tradingagents/agents/utils/memory.py:29 (Issue #30)
• Improved error messages for missing OPENROUTER_API_KEY when using openrouter provider
• Better embedding client initialization for different LLM providers

---

[1.0.0] - 2025-01-01 (Example - Update with actual release date)

Added

• Initial multi-agent trading framework release
• Support for multiple LLM providers
• Analyst team (fundamental, sentiment, news, technical)
• Researcher debate mechanism
• Risk management workflow
• CLI interface
• Integration with financial data APIs