QLib

https://qlib.readthedocs.io/en/latest
Admin
QLib is a quantitative finance library that provides tools and frameworks for algorithmic trading...
Tokens:231,865
Snippets:2,538
Trust Score:8.3
Update:4 months ago
Show doc for...
Tokens:
Raw
### Qlib RL Quick Start

Source: https://qlib.readthedocs.io/en/latest/component/model.html

A quick start guide for implementing Reinforcement Learning strategies within Qlib. This section typically provides a minimal working example.

```python
# Placeholder for RL quick start code.
print('Qlib RL Quick Start guide.')
```

--------------------------------

### Qlib Initialization Example

Source: https://qlib.readthedocs.io/en/latest/component/online.html

Demonstrates how to initialize Qlib with custom parameters. Ensure Qlib is installed and configured before running.

```python
from qlib.config import C

# Initialize Qlib with default parameters
# C.set('qlib_data_path', os.path.join(os.getcwd(), 'data'))
# qlib.init(config_file='path/to/your/config.yaml', market='cs')

# Initialize Qlib with default parameters
import qlib
qlib.init()

```

--------------------------------

### Qlib Recorder Start Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Starts a new recording session with QlibRecorder. This is typically the first step before logging any experiment data.

```python
recorder.start()
print('QlibRecorder started.')
```

--------------------------------

### Clone Qlib Repository and Install

Source: https://qlib.readthedocs.io/en/latest/introduction/quick.html

Clone the Qlib repository from GitHub and install it using the setup script.

```bash
git clone https://github.com/microsoft/qlib.git && cd qlib
python setup.py install
```

--------------------------------

### Qlib Server Initialization Example

Source: https://qlib.readthedocs.io/en/latest/component/online.html

Illustrates the basic setup for running Qlib in server mode. This allows for online serving of strategies and data.

```python
import qlib
from qlib.server.manager import OnlineManager

# Initialize Qlib
qlib.init()

# Initialize the OnlineManager
# online_manager = OnlineManager()
# online_manager.start()

# This is a placeholder as starting the server requires specific configuration and environment setup.
print("Placeholder for Qlib server initialization.")

```

--------------------------------

### Qlib MetaGuideModel Initialization Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Demonstrates initializing a MetaGuideModel, used to guide the meta-learning process in Qlib. It often incorporates task-specific guidance.

```python
from qlib.meta.model import MetaGuideModel

meta_guide_model = MetaGuideModel(input_dim=128, output_dim=1)
print(f'MetaGuideModel initialized with input_dim={meta_guide_model.input_dim}')
```

--------------------------------

### Install Qlib from Source

Source: https://qlib.readthedocs.io/en/latest/start/installation.html

Install Qlib from its source code. This involves cloning the repository, installing dependencies like numpy and cython, and then running the setup script.

```bash
pip install numpy
pip install --upgrade cython
git clone https://github.com/microsoft/qlib.git && cd qlib
python setup.py install

```

--------------------------------

### Qlib Recorder Start Experiment Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Starts a specific experiment within the QlibRecorder session. This allows for distinct logging of parameters and metrics for each experiment.

```python
recorder.start_exp(experiment_name='my_first_experiment')
print('Experiment started.')
```

--------------------------------

### Qlib Installation

Source: https://qlib.readthedocs.io/en/latest/changelog/changelog.html

Instructions for installing the Qlib library.

```bash
pip install qlib
```

--------------------------------

### Qlib Recorder Initialization Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Provides a basic example of initializing the QlibRecorder for experiment tracking and management.

```python
from qlib.workflow.recorder import QlibRecorder

recorder = QlibRecorder()
print('QlibRecorder initialized.')
```

--------------------------------

### Qlib Initialization Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Provides a basic example of initializing Qlib. This typically involves setting the data path and calling the init function.

```python
import qlib
import os

# Set the path to your Qlib data directory
qlib_data_path = os.path.expanduser('~/.qlib/qlib_data')
qlib.init(provider_uri='{}/qlib_data'.format(qlib_data_path), region='cn')
print('Qlib initialized successfully.')
```

--------------------------------

### Qlib MetaModel Initialization Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Provides an example of initializing a general MetaModel in Qlib, which is a core component for meta-learning.

```python
from qlib.meta.model import MetaModel

# Assuming input_dim and output_dim are defined based on the task
meta_model = MetaModel(input_dim=128, output_dim=1)
print(f'MetaModel initialized with input_dim={meta_model.input_dim}')
```

--------------------------------

### Install Qlib Dependencies and Source

Source: https://qlib.readthedocs.io/en/latest/_sources/introduction/quick.rst.txt

Installs necessary Python dependencies and clones the Qlib repository from GitHub to build from source.

```bash
pip install numpy
pip install --upgrade cython
git clone https://github.com/microsoft/qlib.git && cd qlib
python setup.py install
```

--------------------------------

### Qlib Recorder Start Experiment

Source: https://qlib.readthedocs.io/en/latest/changelog/changelog.html

Starts a new experiment recording session.

```python
from qlib.workflow.recorder import QlibRecorder

recorder = QlibRecorder()

# Start a new experiment with a specific name
exp_name = 'my_first_experiment'
recorder.start_exp(experiment_name=exp_name)

print(f"Started experiment: {exp_name}")
```

--------------------------------

### POST /QlibRecorder/start_exp

Source: https://qlib.readthedocs.io/en/latest/component/rl/overall.html

Starts a new experiment session within the Qlib recorder framework.

```APIDOC
## POST /QlibRecorder/start_exp

### Description
Initializes a new experiment session to track research progress and results.

### Method
POST

### Endpoint
/QlibRecorder/start_exp

### Parameters
#### Request Body
- **experiment_name** (string) - Required - The name of the experiment to start.
```

--------------------------------

### Finetune Model with qlib.workflow.R

Source: https://qlib.readthedocs.io/en/latest/reference/api.html

Demonstrates a typical workflow for finetuning a model using qlib.workflow.R. It involves starting an experiment to train an initial model, saving it, and then starting a new experiment to finetune the loaded model.

```python
# start exp to train init model
with R.start(experiment_name="init models"):
    model.fit(dataset)
    R.save_objects(init_model=model)
    rid = R.get_recorder().id

# Finetune model based on previous trained model
with R.start(experiment_name="finetune model"):
    recorder = R.get_recorder(recorder_id=rid, experiment_name="init models")
    model = recorder.load_object("init_model")
    model.finetune(dataset, num_boost_round=10)

```

--------------------------------

### Install Qlib Dependencies

Source: https://qlib.readthedocs.io/en/latest/introduction/quick.html

Install necessary Python dependencies for Qlib before building from source.

```bash
pip install numpy
pip install --upgrade  cython
```

--------------------------------

### Start Qlib Recorder Experiment

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Starts a new experiment recording session using QlibRecorder. This should be called before logging any experiment-specific data.

```python
recorder.start_exp(experiment_name='my_experiment')
print('Experiment started')
```

--------------------------------

### Qlib DataLoader Example

Source: https://qlib.readthedocs.io/en/latest/changelog/changelog.html

Demonstrates the usage of QlibDataLoader for data retrieval.

```python
from qlib.data.data_loader import QlibDataLoader

# Instantiate QlibDataLoader
data_loader = QlibDataLoader()

# Example: Retrieve features for a specific instrument and time range
# Note: This is a conceptual example. Actual usage depends on configured data sources.
instrument = 'IF159901'
start_time = '2020-01-01'
end_time = '2020-01-31'
features = ['close', 'open', 'high', 'low']

# data = data_loader.fetch(instruments=instrument, start_time=start_time, end_time=end_time, fields=features)
# print(data)
```

--------------------------------

### Verify Qlib Installation

Source: https://qlib.readthedocs.io/en/latest/_sources/start/installation.rst.txt

A simple Python snippet to verify that Qlib is correctly installed by checking the library version.

```python
import qlib
print(qlib.__version__)
```

--------------------------------

### Start RL Training

Source: https://qlib.readthedocs.io/en/latest/component/rl/quickstart.html

Use this command to start the RL training process. Ensure you have a `train_config.yml` file specifying your training parameters.

```bash
python -m qlib.rl.contrib.train_onpolicy.py --config_path train_config.yml
```

--------------------------------

### Qlib Serialization Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Provides an example of using Qlib's serialization utilities to save and load Python objects. This is crucial for persisting models and configurations.

```python
from qlib.utils.serialization import Serializable

class MyData(Serializable):
    def __init__(self, value):
        self.value = value

    def get_serialization_content(self):
        return self.value

    @classmethod
    def from_serialization_content(cls, content):
        return cls(content)

data_instance = MyData(42)
# To serialize: data_instance.serialize('my_data.pkl')
# To deserialize: loaded_data = MyData.deserialize('my_data.pkl')
print('Qlib serialization example demonstrated.')
```

--------------------------------

### Command Line Interface Examples

Source: https://qlib.readthedocs.io/en/latest/advanced/task_management.html

Provides examples of how to use the TaskManager module from the command line for common operations like checking task status.

```APIDOC
## Command Line Interface

This class can be used as a tool from commandline. Here are several examples. You can view the help of manage module with the following commands: 

```bash
python -m qlib.workflow.task.manage -h # show manual of manage module CLI
python -m qlib.workflow.task.manage wait -h # show manual of the wait command of manage
```

```bash
python -m qlib.workflow.task.manage -t <pool_name> wait
python -m qlib.workflow.task.manage -t <pool_name> task_stat
```
```

--------------------------------

### Qlib Workflow Configuration Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

An example of a Qlib workflow configuration file in YAML format. This defines the dataset, model, and task parameters for a research workflow.

```yaml
qlib_init: |
  from qlib_scripts.workflow_init import init_qlib
  init_qlib()

task: |
  from qlib.workflow import Task
  from qlib.model. Prophet import Prophet
  from qlib.data.dataset import DatasetD

  task = Task(
      dataset=DatasetD(market='cs', freq='day'),
      model=Prophet(),
      name='prophet_task'
  )
  task.train()
  task.test()

```

--------------------------------

### Building Formulaic Alphas in Qlib Example

Source: https://qlib.readthedocs.io/en/latest/component/online.html

Provides an example of how to build custom formulaic alpha factors within the Qlib framework. This involves defining mathematical expressions based on financial data.

```python
from qlib.data import D
from qlib.strategy.portfolio import PortfolioStrategy
from qlib.backtest.executor import BacktestExecutor
from qlib.model.expression import Expression, ExpressionParam

# Define a formulaic alpha
# Example: Momentum factor (close price change over N days)
alpha_expr = Expression("($close - $close.shift(10)) / $close.shift(10)")

# You would then use this alpha within a Qlib strategy and backtest it.
# This is a simplified example; actual implementation involves more setup.
print("Placeholder for building formulaic alphas example.")

```

--------------------------------

### Building Formulaic Alphas in Qlib

Source: https://qlib.readthedocs.io/en/latest/changelog/changelog.html

Example demonstrating how to build custom formulaic alphas within Qlib.

```python
from qlib.expression import Expression, H, F

# Define a simple formulaic alpha: (close - open) / open
# alpha_expr = (H.close - H.open) / H.open

# More complex example: Momentum factor
# alpha_expr = F.ts_delay(H.close, 1) / H.close - 1

# To use this alpha, you would typically integrate it into a Qlib workflow
# For example, by defining it in a configuration file or programmatically.
print("Formulaic alpha definition example (commented out).")
```

--------------------------------

### Install and Run Flake8 for Code Style Checking

Source: https://qlib.readthedocs.io/en/latest/_sources/developer/code_standard_and_dev_guide.rst.txt

Installs flake8 and runs it on the Qlib directory, ignoring specific error codes. Flake8 checks for style guide violations and potential errors in Python code.

```bash
flake8 --ignore E501,F541,E402,F401,W503,E741,E266,E203,E302,E731,E262,F523,F821,F811,F841,E713,E265,W291,E712,E722,W293 qlib
```

--------------------------------

### Qlib Recorder Get URI Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Retrieves the current storage URI for the QlibRecorder. This indicates where experiment records are being saved.

```python
current_uri = recorder.get_uri()
print(f'Current recorder URI: {current_uri}')
```

--------------------------------

### Execute Daily Backtest with Qlib

Source: https://qlib.readthedocs.io/en/latest/_sources/component/strategy.rst.txt

Demonstrates how to initialize Qlib, configure a TopkDropoutStrategy, and perform a daily backtest followed by risk analysis.

```python
from pprint import pprint
import qlib
import pandas as pd
from qlib.contrib.evaluate import backtest_daily, risk_analysis
from qlib.contrib.strategy import TopkDropoutStrategy

qlib.init(provider_uri='<qlib data dir>')

STRATEGY_CONFIG = {
    "topk": 50,
    "n_drop": 5,
    "signal": pred_score,
}

strategy_obj = TopkDropoutStrategy(**STRATEGY_CONFIG)
report_normal, positions_normal = backtest_daily(
    start_time="2017-01-01", end_time="2020-08-01", strategy=strategy_obj
)

analysis = {
    "excess_return_without_cost": risk_analysis(report_normal["return"] - report_normal["bench"]),
    "excess_return_with_cost": risk_analysis(report_normal["return"] - report_normal["bench"] - report_normal["cost"])
}

analysis_df = pd.concat(analysis)
pprint(analysis_df)
```

--------------------------------

### LocalPITProvider.period_feature

Source: https://qlib.readthedocs.io/en/latest/reference/api.html

Gets historical period data series for a given instrument and field between specified start and end indices, and current time.

```APIDOC
## LocalPITProvider.period_feature

### Description
get the historical periods data series between start_index and end_index 

### Method
period_feature

### Parameters
#### Path Parameters
* **instrument** (str) - 
* **field** (str) - 
* **start_index** (str) - 
* **end_index** (str) - 
* **cur_time** (str) - 
* **period** (str) - 

### Returns

```

--------------------------------

### Initialize Qlib Model Instance from Configuration (Python)

Source: https://qlib.readthedocs.io/en/latest/_sources/component/workflow.rst.txt

Demonstrates how to initialize a Qlib model, specifically LGBModel, using a dictionary of keyword arguments. This mirrors the configuration typically found in YAML files for Qlib.

```python
from qlib.contrib.model.gbdt import LGBModel
kwargs = {
    "loss": "mse" ,
    "colsample_bytree": 0.8879,
    "learning_rate": 0.0421,
    "subsample": 0.8789,
    "lambda_l1": 205.6999,
    "lambda_l2": 580.9768,
    "max_depth": 8,
    "num_leaves": 210,
    "num_threads": 20,
}
LGBModel(kwargs)
```

--------------------------------

### Qlib Recorder Get Recorder Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Retrieves a specific recorder instance by its ID. This allows direct access to a particular recorder's functionalities.

```python
specific_recorder = recorder.get_recorder('recorder_id_xyz')
print(f'Retrieved recorder: {specific_recorder}')
```

--------------------------------

### ExpManager.start_exp

Source: https://qlib.readthedocs.io/en/latest/component/recorder.html

Starts a new experiment or resumes an existing one. It handles getting or creating an experiment and setting it as active. It also manages the active experiment URI.

```APIDOC
## ExpManager.start_exp

### Description
Starts an experiment. This method includes first get_or_create an experiment, and then set it to be active. Maintaining _active_exp_uri is included in start_exp, remaining implementation should be included in _end_exp in subclass.

### Method
start_exp

### Parameters
* **experiment_id** (str) - id of the active experiment.
* **experiment_name** (str) - name of the active experiment.
* **recorder_id** (str) - id of the recorder to be started.
* **recorder_name** (str) - name of the recorder to be started.
* **uri** (str) - the current tracking URI.
* **resume** (boolean) - whether to resume the experiment and recorder.

### Return type
An active experiment.
```

--------------------------------

### Qlib Recorder Get Experiment Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Retrieves a specific experiment record by its unique ID. This allows detailed inspection of a past experiment's metadata.

```python
experiment_record = recorder.get_exp('some_experiment_id')
print(experiment_record)
```

--------------------------------

### Initialize QLib with Data Provider

Source: https://qlib.readthedocs.io/en/latest/start/getdata.html

Initializes QLib and specifies the path to the QLib data. Ensure data is downloaded and the path is correct.

```python
import qlib
qlib.init(provider_uri='~/.qlib/qlib_data/cn_data')
```

--------------------------------

### Qlib Data Retrieval Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Retrieves historical data for a given instrument and field using Qlib's Data API. Specify the instrument, field, start, and end dates.

```python
from qlib.data import D

data = D.features(instruments='IF000', fields='close', start_time='2017-01-01', end_time='2017-01-10')
print(data)
```

--------------------------------

### Get Recorder Instance

Source: https://qlib.readthedocs.io/en/latest/component/recorder.html

Retrieves a recorder instance based on provided IDs or names. If multiple recorders match the query (e.g., by experiment name) and an MLflow backend is used, the recorder with the latest start time will be returned.

```python
recorder = R.get_recorder(recorder_id='2e7a4efd66574fa49039e00ffaefa99d', experiment_name='test')
```

--------------------------------

### Qlib Data Preparation Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Shows how to prepare data for Qlib, converting CSV files into Qlib's native format. This is a crucial step before using Qlib for analysis.

```python
from qlib_scripts.data_handler import CSV2Bin

# Define input CSV file and output bin file paths
csv_path = 'path/to/your/data.csv'
bin_path = 'path/to/your/data.bin'

# Create a CSV2Bin object and run the conversion
c2b = CSV2Bin(csv_path, bin_path)
c2b.run()
print('Data conversion complete.')
```

--------------------------------

### Example Qlib Workflow Configuration with Filter

Source: https://qlib.readthedocs.io/en/latest/component/data.html

This snippet shows how to define and use an ExpressionDFilter within a Qlib workflow configuration file. It specifies filter parameters like rule expression, start and end times, and whether to keep filtered data.

```yaml
filter: &filter
    filter_type: ExpressionDFilter
    rule_expression: "Ref($close, -2) / Ref($close, -1) > 1"
    filter_start_time: 2010-01-01
    filter_end_time: 2010-01-07
    keep: False

data_handler_config: &data_handler_config
    start_time: 2010-01-01
    end_time: 2021-01-22
    fit_start_time: 2010-01-01
    fit_end_time: 2015-12-31
    instruments: *market
    filter_pipe: [*filter]

```

--------------------------------

### Qlib Online Manager Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Illustrates the initialization and basic usage of the Online Manager in Qlib, which handles online serving functionalities.

```python
from qlib.online import OnlineManager

manager = OnlineManager()
# manager.start() # Example of starting the manager
print('OnlineManager initialized.')
```

--------------------------------

### Install Qlib using pip

Source: https://qlib.readthedocs.io/en/latest/start/installation.html

Use this command to install Qlib directly from pip. Ensure you have Python 3.8 or earlier installed.

```bash
pip install pyqlib

```

--------------------------------

### Qlib MetaGuideModel Initialization

Source: https://qlib.readthedocs.io/en/latest/component/online.html

Shows the initialization of `MetaGuideModel`, another specialized meta-model, likely used for guiding the meta-learning process.

```python
from qlib_scripts.workflow_tech.meta_model import MetaGuideModel

# Initialize MetaGuideModel
# meta_guide_model = MetaGuideModel(config_path='path/to/your/meta_guide_model_config.yaml')

# This is a placeholder as the actual initialization requires a configuration file.
print("Placeholder for MetaGuideModel initialization.")

```

--------------------------------

### Qlib Recorder Methods Example

Source: https://qlib.readthedocs.io/en/latest/component/online.html

Demonstrates various methods of the QlibRecorder for managing experiments, including searching, listing, and deleting records.

```python
from qlib.workflow.recorder import QlibRecorder

recorder = QlibRecorder()

# Search for records
records = recorder.search_records(experiment_name='my_experiment')
print(f"Found {len(records)} records.")

# List all experiments
experiments = recorder.list_experiments()
print(f"Experiments: {experiments}")

# List all recorders
recorders = recorder.list_recorders()
print(f"Recorders: {recorders}")

# Get a specific experiment
exp = recorder.get_exp(experiment_name='my_experiment')

# Delete an experiment
# recorder.delete_exp(experiment_name='my_experiment')

```

--------------------------------

### Install Development Dependencies and Pre-commit Hooks

Source: https://qlib.readthedocs.io/en/latest/developer/code_standard_and_dev_guide.html

Installs Qlib and its development dependencies in editable mode, and then installs git pre-commit hooks. This ensures code is automatically formatted before committing.

```bash
pip install -e .[dev]
pre-commit install
```

--------------------------------

### Command-line Interface Examples

Source: https://qlib.readthedocs.io/en/latest/advanced/task_management.html

Demonstrates how to use the TaskManager module from the command line to manage tasks. Use '-h' for detailed help.

```bash
python -m qlib.workflow.task.manage -h # show manual of manage module CLI
python -m qlib.workflow.task.manage wait -h # show manual of the wait command of manage

```

```bash
python -m qlib.workflow.task.manage -t <pool_name> wait
python -m qlib.workflow.task.manage -t <pool_name> task_stat

```

--------------------------------

### Install Qlib with Development Dependencies

Source: https://qlib.readthedocs.io/en/latest/_sources/developer/code_standard_and_dev_guide.rst.txt

Installs Qlib in editable mode, allowing changes to reflect immediately without reinstallation. It also installs development-related packages like pytest and sphinx.

```bash
pip install -e .[dev]
```

--------------------------------

### Qlib MetaGuideModel Initialization

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Initializes a MetaGuideModel, which is used to guide the meta-learning process. This model often incorporates task-specific information.

```python
from qlib.meta.model import MetaGuideModel

meta_guide_model = MetaGuideModel(input_dim=10, output_dim=1)
print(meta_guide_model)
```

--------------------------------

### Initialize Qlib with Experiment Manager

Source: https://qlib.readthedocs.io/en/latest/_sources/start/initialization.rst.txt

This Python example demonstrates initializing Qlib with an experiment manager, specifically MLflow. It requires specifying the class, module path, and keyword arguments for the experiment manager, including the tracking URI.

```python
import qlib
from qlib.constant import REG_CN
provider_uri = "~/.qlib/qlib_data/cn_data"  # target_dir
qlib.init(provider_uri=provider_uri, region=REG_CN, exp_manager= {
    "class": "MLflowExpManager",
    "module_path": "qlib.workflow.expm",
    "kwargs": {
        "uri": "python_execution_path/mlruns",
        "default_exp_name": "Experiment",
    }
})
```

--------------------------------

### Qlib Format Data Example

Source: https://qlib.readthedocs.io/en/latest/changelog/changelog.html

Example of converting CSV data to Qlib format.

```python
from qlib.data.data_handler import DataHandlerLP
from qlib.data.data import D
import pandas as pd

# Assuming you have a CSV file named 'your_data.csv'
data_path = 'your_data.csv'
qlib_data_path = 'path/to/save/qlib/data'

# Convert CSV to Qlib format
# This is a simplified example; refer to Qlib documentation for detailed options
# Example: Convert daily stock data
# You would typically use a more robust script for actual data conversion
# For demonstration, let's assume a DataFrame structure
# df = pd.read_csv(data_path)
# ... process df and save to qlib format ...

# Example of loading data after conversion (conceptual)
# D.instruments('csi300') # Load instruments
# D.features(['close', 'open'], start_time='2017-01-01', end_time='2017-01-31') # Load features
```

--------------------------------

### Qlib Online Strategy Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Shows how to define and use an Online Strategy within Qlib's online serving framework. Strategies dictate trading decisions in real-time.

```python
from qlib.online.strategy import OnlineStrategy

class MyOnlineStrategy(OnlineStrategy):
    def __init__(self, **kwargs):
        super().__init__(**kwargs)

    def predict(self, data):
        # Implement prediction logic here
        return super().predict(data)

strategy = MyOnlineStrategy()
print('OnlineStrategy defined.')
```

--------------------------------

### Initialize Qlib

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Initializes Qlib with specified parameters. Ensure all required parameters are correctly set before use.

```python
from qlib.config import C
C.set('qlib_data_path', os.path.join(os.getcwd(), 'data'))
init_qlib()
```

--------------------------------

### Install Qlib in Editable Mode for Development

Source: https://qlib.readthedocs.io/en/latest/developer/code_standard_and_dev_guide.html

Installs Qlib in editable mode, allowing changes to reflect directly in the environment without reinstallation. The '[dev]' option installs additional packages needed for development.

```bash
pip install -e ".[dev]"
```

--------------------------------

### Install Pre-commit Hooks for Qlib Development

Source: https://qlib.readthedocs.io/en/latest/_sources/developer/code_standard_and_dev_guide.rst.txt

Installs Qlib in editable mode with development dependencies and then installs git pre-commit hooks. These hooks automatically format code using black and flake8 before each commit.

```bash
pip install -e .[dev]
pre-commit install
```

--------------------------------

### Qlib MetaTaskDataset Initialization Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Demonstrates initializing a MetaTaskDataset, which is used to prepare datasets for meta-learning tasks in Qlib.

```python
from qlib.meta.task import MetaTaskDataset

meta_dataset = MetaTaskDataset(dataset_name='example_dataset')
print(f'MetaTaskDataset initialized for dataset: {meta_dataset.dataset_name}')
```

--------------------------------

### Qlib Server Initialization

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Initializes the Qlib server for online serving. This setup is required for real-time data and strategy execution.

```python
from qlib_server.server import QlibServer

server = QlibServer()
server.start()
```

--------------------------------

### Data Handler Initialization with Processors

Source: https://qlib.readthedocs.io/en/latest/reference/api.html

Illustrates how to initialize DataHandlerLP with different processor lists for inference and learning. Shows examples of processor configurations using class names, kwargs, or instance objects.

```python
from qlib.data.dataset.handler import DataLoader, DataLoaderLP
from qlib.data.dataset.processor import MinMaxNorm, DropnaFeature

# Example 1: Using class names and kwargs
infer_processors_1 = [
    {
        "class": "MinMaxNorm",
        "kwargs": {
            "fit_start_time": "20080101",
            "fit_end_time": "20121231"
        }
    },
    "DropnaFeature"
]

# Example 2: Using object instances
minmax_norm_processor = MinMaxNorm(fit_start_time="20080101", fit_end_time="20121231")
learn_processors_2 = [minmax_norm_processor, DropnaFeature()]

# Example initialization (assuming other parameters are set)
# handler_lp = DataLoaderLP(
#     data_loader=...,
#     infer_processors=infer_processors_1,
#     learn_processors=learn_processors_2,
#     process_type='append',
#     drop_raw=False
# )
```

--------------------------------

### Get Default Experiment Instance

Source: https://qlib.readthedocs.io/en/latest/component/recorder.html

Retrieves the default experiment instance. This is a shorthand for getting the experiment when no specific name is provided.

```python
exp = R.get_exp() -> a default experiment.
```

--------------------------------

### Qlib Workflow Configuration Example

Source: https://qlib.readthedocs.io/en/latest/component/online.html

Illustrates a basic Qlib workflow configuration file. This defines the model, dataset, and recorder settings for a research task.

```yaml
import qlib
from qlib.config import C

# Initialize Qlib
qlib.init()

# Example of a configuration file for Qlib workflow
# This is a conceptual representation and would typically be in a .yaml file
config = {
    'task': {
        'model': {
            'class': 'GRU',
            'module_path': 'qlib.contrib.model.grudnn',
            'kwargs': {
                'hidden_size': 256,
                'num_layers': 2
            }
        },
        'dataset': {
            'class': 'DatasetD', 
            'module_path': 'qlib.data.dataset.dataset', 
            'kwargs': {
                'handler': {
                    'class': 'Alpha158', 
                    'module_path': 'qlib.data.dataset.handler.alpha158',
                    'kwargs': {
                        'fields_group': 'close', 
                        'feature_group_name': 'features'
                    }
                },
                'infer_processor': {
                    'class': 'PosiSharpміну', 
                    'module_path': 'qlib.data.dataset.processor'
                }
            }
        },
        'record': {
            'experiment_name': 'example_experiment',
            'recorder': {
                'class': 'MLflowRecorder',
                'module_path': 'qlib.workflow.recorder'
            }
        }
    }
}

# To use this config, you would typically save it to a YAML file and pass its path to qlib.workflow.init()
# For example:
# with open('config.yaml', 'w') as f:
#     yaml.dump(config, f)
# qlib.workflow.init(config_file='config.yaml')

```

--------------------------------

### Qlib MetaTask Initialization Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Shows how to initialize a MetaTask object, which is used in Qlib's meta-learning framework. It requires a task name and dataset name.

```python
from qlib.meta.task import MetaTask

meta_task = MetaTask(task_name='example_task', dataset_name='example_dataset')
print(f'MetaTask initialized: {meta_task.task_name}')
```

--------------------------------

### DataHandlerLP.setup_data

Source: https://qlib.readthedocs.io/en/latest/component/data.html

Sets up the data for initialization, particularly useful when running initialization for multiple times. It supports different initialization types and optional caching.

```APIDOC
## DataHandlerLP.setup_data

### Description
Sets up the data for initialization, particularly useful when running initialization for multiple times. It supports different initialization types and optional caching.

### Method
setup_data

### Parameters
- **init_type** (str) - Optional - The type of initialization. Defaults to 'fit_seq'.
- **enable_cache** (bool) - Optional - If true, processed data will be saved on disk for faster loading in subsequent calls.
- **kwargs** - Additional keyword arguments.

### Request Example
```json
{
  "init_type": "fit_seq",
  "enable_cache": false
}
```

### Response
This method does not explicitly define a return value in the provided documentation, but it sets up the data handler.
```

--------------------------------

### Qlib Initialization API

Source: https://qlib.readthedocs.io/en/latest/component/online.html

Details on initializing Qlib and its parameters.

```APIDOC
## Initialization API

### Description
Provides details on how to initialize the Qlib library and its associated parameters.

### Method
GET

### Endpoint
/api/qlib/initialization

### Parameters
#### Query Parameters
- **config_path** (string) - Optional - Path to the Qlib configuration file.

### Response
#### Success Response (200)
- **status** (string) - Initialization status message.

#### Response Example
{
  "status": "Qlib initialized successfully."
}
```

--------------------------------

### Instrument Provider Filter Example

Source: https://qlib.readthedocs.io/en/latest/reference/api.html

Example configuration for filtering instruments based on market and dynamic rules. This configuration can be used with the InstrumentProvider.

```python
{
'market': 'csi500',
'filter_pipe': [{'filter_type': 'ExpressionDFilter',
'rule_expression': '$open<40',
'filter_start_time': None,
'filter_end_time': None,
'keep': False},
{'filter_type': 'NameDFilter',
'name_rule_re': 'SH[0-9]{4}55',
'filter_start_time': None,
'filter_end_time': None}]}

```

--------------------------------

### Perform Backtest and Risk Analysis

Source: https://qlib.readthedocs.io/en/latest/component/recorder.html

Example demonstrating how to perform a backtest using `TopkDropoutStrategy` and analyze the results, including excess returns with and without costs. Requires importing strategy and backtest modules from `qlib.contrib.strategy.strategy` and `qlib.contrib.evaluate`.

```python
from qlib.contrib.strategy.strategy import TopkDropoutStrategy
from qlib.contrib.evaluate import (
    backtest as normal_backtest,
    risk_analysis,
)

# backtest
STRATEGY_CONFIG = {
    "topk": 50,
    "n_drop": 5,
}
BACKTEST_CONFIG = {
    "limit_threshold": 0.095,
    "account": 100000000,
    "benchmark": BENCHMARK,
    "deal_price": "close",
    "open_cost": 0.0005,
    "close_cost": 0.0015,
    "min_cost": 5,
}

strategy = TopkDropoutStrategy(**STRATEGY_CONFIG)
report_normal, positions_normal = normal_backtest(pred_score, strategy=strategy, **BACKTEST_CONFIG)

# analysis
analysis = dict()
analysis["excess_return_without_cost"] = risk_analysis(report_normal["return"] - report_normal["bench"])
analysis["excess_return_with_cost"] = risk_analysis(report_normal["return"] - report_normal["bench"] - report_normal["cost"])
analysis_df = pd.concat(analysis)  # type: pd.DataFrame
print(analysis_df)
```

--------------------------------

### Install Compatible python-socketio Version

Source: https://qlib.readthedocs.io/en/latest/FAQ/FAQ.html

Ensure `python-socketio` versions match between Qlib and Qlib-server to avoid `BadNamespaceError`. Install a specific version using pip.

```bash
pip install -U python-socketio==<qlib-server python-socketio version>
```

--------------------------------

### Rebase FeatureStorage with start and end indices

Source: https://qlib.readthedocs.io/en/latest/reference/api.html

Adjusts the start and end indices of a FeatureStorage. The interval is inclusive. This can be used to truncate or shift the data range.

```python
feature:
    3   4
    4   5
    5   6

>>> self.rebase(start_index=4)

    feature:
        4   5
        5   6

>>> self.rebase(start_index=3)

    feature:
        3   np.nan
        4   5
        5   6

>>> self.write([3], index=3)

    feature:
        3   3
        4   5
        5   6

>>> self.rebase(end_index=4)

    feature:
        3   3
        4   5

>>> self.write([6, 7, 8], index=4)

    feature:
        3   3
        4   6
        5   7
        6   8

>>> self.rebase(start_index=4, end_index=5)

    feature:
        4   6
        5   7

```

--------------------------------

### Qlib Recorder Initialization

Source: https://qlib.readthedocs.io/en/latest/changelog/changelog.html

Initialize the QlibRecorder to manage experiments.

```python
from qlib_scripts.workflow_init import init_qlib

# Initialize Qlib and set up the recorder
init_qlib(config_path='your_qlib_config.yaml')

from qlib.workflow.recorder import QlibRecorder

# Create a recorder instance
recorder = QlibRecorder()
print(recorder)
```

--------------------------------

### Qlib Online Tool Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Demonstrates the usage of an Online Tool within Qlib's online serving system. Tools provide utility functions for online operations.

```python
from qlib.online.tool import OnlineTool

tool = OnlineTool()
# Example usage: tool.get_latest_data(...)
print('OnlineTool initialized.')
```

--------------------------------

### Query Task by ID Example

Source: https://qlib.readthedocs.io/en/latest/advanced/task_management.html

Example of querying a specific task in the collection using its MongoDB _id. This can raise a CursorNotFound exception if the cursor is too old.

```bash
python -m qlib.workflow.task.manage -t <your task pool> query ‘{“_id”: “615498be837d0053acbc5d58”}’

```

--------------------------------

### Qlib MetaTaskModel Initialization Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Shows how to initialize a MetaTaskModel, a specialized meta-model for specific tasks within Qlib's meta-learning framework.

```python
from qlib.meta.model import MetaTaskModel

meta_task_model = MetaTaskModel(input_dim=128, output_dim=1)
print(f'MetaTaskModel initialized with input_dim={meta_task_model.input_dim}')
```

--------------------------------

### Qlib Workflow Configuration Example (YAML)

Source: https://qlib.readthedocs.io/en/latest/_sources/component/workflow.rst.txt

A sample YAML configuration file for Qlib's `qrun` command. This file defines parameters for data handling, model training (e.g., LGBModel), dataset segmentation, and recording results like signal and portfolio analysis.

```YAML
qlib_init:
    provider_uri: "~/.qlib/qlib_data/cn_data"
    region: cn
market: &market csi300
benchmark: &benchmark SH000300
data_handler_config: &data_handler_config
    start_time: 2008-01-01
    end_time: 2020-08-01
    fit_start_time: 2008-01-01
    fit_end_time: 2014-12-31
    instruments: *market
port_analysis_config: &port_analysis_config
    strategy:
        class: TopkDropoutStrategy
        module_path: qlib.contrib.strategy.strategy
        kwargs:
            topk: 50
            n_drop: 5
            signal: <PRED>
    backtest:
        start_time: 2017-01-01
        end_time: 2020-08-01
        account: 100000000
        benchmark: *benchmark
        exchange_kwargs:
            limit_threshold: 0.095
            deal_price: close
            open_cost: 0.0005
            close_cost: 0.0015
            min_cost: 5
task:
    model:
        class: LGBModel
        module_path: qlib.contrib.model.gbdt
        kwargs:
            loss: mse
            colsample_bytree: 0.8879
            learning_rate: 0.0421
            subsample: 0.8789
            lambda_l1: 205.6999
            lambda_l2: 580.9768
            max_depth: 8
            num_leaves: 210
            num_threads: 20
    dataset:
        class: DatasetH
        module_path: qlib.data.dataset
        kwargs:
            handler:
                class: Alpha158
                module_path: qlib.contrib.data.handler
                kwargs: *data_handler_config
            segments:
                train: [2008-01-01, 2014-12-31]
                valid: [2015-01-01, 2016-12-31]
                test: [2017-01-01, 2020-08-01]
    record:
        - class: SignalRecord
          module_path: qlib.workflow.record_temp
          kwargs: {}
        - class: PortAnaRecord
          module_path: qlib.workflow.record_temp
          kwargs:
              config: *port_analysis_config
```

--------------------------------

### Begin Task Training

Source: https://qlib.readthedocs.io/en/latest/reference/api.html

Starts task training by initializing a recorder and saving the task configuration. Use this to begin the training process for a specific task.

```python
qlib.model.trainer.begin_task_train(_task_config : dict_, _experiment_name : str_, _recorder_name : str | None = None_)
```

--------------------------------

### get_longest_back_rolling

Source: https://qlib.readthedocs.io/en/latest/reference/api.html

Get the longest length of historical data the feature has accessed. This is designed for getting the needed range of the data to calculate the features in specific range at first.

```APIDOC
## get_longest_back_rolling()

### Description
Get the longest length of historical data the feature has accessed. This is designed for getting the needed range of the data to calculate the features in specific range at first. However, situations like Ref(Ref($close, -1), 1) can not be handled rightly. So this will only used for detecting the length of historical data needed.

### Returns

the longest length of historical data the feature has accessed

### Return type

int
```

--------------------------------

### Install Qlib via pip

Source: https://qlib.readthedocs.io/en/latest/_sources/start/installation.rst.txt

The standard method to install the Qlib library using the Python package manager. This command fetches the latest stable release from PyPI.

```bash
pip install pyqlib
```

--------------------------------

### Qlib Recorder Initialization

Source: https://qlib.readthedocs.io/en/latest/component/online.html

Shows how to initialize the QlibRecorder for experiment management. This recorder helps log parameters, metrics, and artifacts.

```python
from qlib.workflow.recorder import QlibRecorder

# Initialize QlibRecorder
recorder = QlibRecorder()

# Start a new experiment
recorder.start_exp(experiment_name='my_experiment')

# Log parameters
recorder.log_params({'param1': 10, 'param2': 'value'})

# Log metrics
recorder.log_metrics({'metric1': 0.95, 'metric2': 0.05})

# End the experiment
recorder.end_exp()

```

--------------------------------

### Qlib Recorder Log Parameters Example

Source: https://qlib.readthedocs.io/en/latest/component/model.html

Logs parameters for the current experiment using QlibRecorder. This is essential for tracking experiment configurations.

```python
params = {'learning_rate': 0.001, 'epochs': 50}
recorder.log_params(params)
print('Parameters logged.')
```
Tokens:
Raw