graph LR
Qlib_Core_Configuration["Qlib Core & Configuration"]
Data_Management["Data Management"]
Model_Strategy_Development["Model & Strategy Development"]
Simulation_Evaluation["Simulation & Evaluation"]
Deployment_Workflow["Deployment & Workflow"]
Qlib_Core_Configuration -- "Configures data sources and storage." --> Data_Management
Qlib_Core_Configuration -- "Sets up model training environments and parameters." --> Model_Strategy_Development
Data_Management -- "Provides processed features and labels for model training and inference." --> Model_Strategy_Development
Model_Strategy_Development -- "Supplies trained models and generated strategies for backtesting." --> Simulation_Evaluation
Model_Strategy_Development -- "Delivers deployable models and strategies for online execution." --> Deployment_Workflow
Simulation_Evaluation -- "Offers performance feedback for iterative model and strategy refinement." --> Model_Strategy_Development
Simulation_Evaluation -- "Reports backtesting results for workflow tracking." --> Deployment_Workflow
Deployment_Workflow -- "Initiates data collection for online serving." --> Data_Management
Deployment_Workflow -- "Orchestrates experiment runs and manages training tasks." --> Model_Strategy_Development
Deployment_Workflow -- "Manages and triggers backtesting runs." --> Simulation_Evaluation
click Data_Management href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/qlib/Data_Management.md" "Details"
click Model_Strategy_Development href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/qlib/Model_Strategy_Development.md" "Details"
click Simulation_Evaluation href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/qlib/Simulation_Evaluation.md" "Details"
click Deployment_Workflow href "https://github.com/CodeBoarding/GeneratedOnBoardings/blob/main/qlib/Deployment_Workflow.md" "Details"
The Qlib project is structured around a robust financial quantitative research platform, enabling end-to-end workflow from data management to model deployment. The core components facilitate data ingestion and preprocessing, sophisticated model and strategy development (including reinforcement learning), rigorous simulation and evaluation through backtesting, and seamless workflow orchestration for experiment management and online deployment. This modular design ensures clear separation of concerns, promoting maintainability and extensibility, and provides a clear data flow for financial research and trading.
The foundational layer for environment initialization and global settings.
Related Classes/Methods:
Data Management [Expand]
Handles data ingestion, storage, preprocessing, feature engineering, and dataset preparation.
Related Classes/Methods:
scripts/data_collector/base.py:1-9999scripts/dump_bin.py:1-9999qlib/data/dataset/processor.py:1-9999qlib/data/dataset/handler.py:1-9999qlib/data/data.py:1-9999
Model & Strategy Development [Expand]
Encompasses model training (including RL), prediction generation, and trading strategy optimization.
Related Classes/Methods:
qlib/model/trainer.py:1-9999qlib/contrib/model/pytorch_alstm.py:1-9999qlib/contrib/model/gbdt.py:1-9999qlib/rl/trainer/trainer.py:1-9999qlib/model/base.py:1-9999qlib/contrib/strategy/signal_strategy.py:1-9999qlib/contrib/strategy/optimizer/optimizer.py:1-9999
Simulation & Evaluation [Expand]
Performs historical backtesting, simulates strategy performance, and generates comprehensive evaluation reports.
Related Classes/Methods:
qlib/workflow/online/manager.py:1-9999qlib/backtest/backtest.py:1-9999qlib/contrib/evaluate.py:1-9999qlib/contrib/report/analysis_position/report.py:1-9999
Deployment & Workflow [Expand]
Manages experiment orchestration, tracks research workflows, and handles online serving and execution of strategies.
Related Classes/Methods: