### Install ExplainerDashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/custom_examples.ipynb

Install the library via pip.

```python
#!pip install explainerdashboard
```

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

### Install Explainer Dashboard

Source: https://context7.com/oegedijk/explainerdashboard/llms.txt

Install the explainerdashboard library using pip.

```bash
pip install explainerdashboard
```

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

### Example Usage

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/dashboards.md

Examples of how to run and terminate the ExplainerDashboard.

```APIDOC
### Example

ExplainerDashboard(explainer, mode=’external’).run(port=8050)

ExplainerDashboard.terminate(8050)
```

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

### run() Method

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/dashboards.md

Starts the ExplainerDashboard on a specified port.

```APIDOC
## run(port: int = None, host: str = '0.0.0.0', use_waitress: bool = False, mode: str = None, sagemaker: bool | None = None, **kwargs)

### Description

Start ExplainerDashboard on port.

### Parameters

#### Parameters

* **port** (*int*, optional) – port to run on. If None, then use self.port.
* **host** (*str*, optional) – host to run on. Defaults to ‘0.0.0.0’.
* **use_waitress** (*bool*, optional) – use the waitress python web server instead of the flask development server. Only works with mode=’dash’. Defaults to False.
* **mode** (*str*, {'dash', 'inline', 'jupyterlab', 'external'}, optional) – Type of dash server to start. ‘inline’ runs in a jupyter notebook output cell. ‘jupyterlab’ runs in a jupyterlab pane. ‘external’ runs in an external tab while keeping the notebook interactive. ‘dash’ is the default server. Overrides self.mode, in which case the dashboard will get rebuilt before running it with the right type of dash server. (dash.Dash or JupyterDash). Defaults to None (i.e. self.mode)
* **sagemaker** (*bool*, optional) – if True, apply SageMaker Studio proxy defaults for routes_pathname_prefix and requests_pathname_prefix. If None, use the value from the dashboard initialization.
* **8050.** (*Defaults to None.self.port defaults to*)

### Raises

**ValueError** – if mode is unknown
```

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

### Install explainerdashboard via pip or conda

Source: https://github.com/oegedijk/explainerdashboard/blob/master/README.md

Use these commands to install the package in your environment.

```bash
pip install explainerdashboard
```

```bash
conda install -c conda-forge explainerdashboard
```

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

### Install Project Dependencies

Source: https://github.com/oegedijk/explainerdashboard/blob/master/CONTRIBUTING.md

Upgrade core packaging tools and install the project in editable mode.

```bash
$ python -m pip install -U pip setuptools wheel
$ pip install -e .
```

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

### Run ExplainerDashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/dashboard_examples.ipynb

Instantiate the dashboard with a fitted classifier and start the server.

```python
ExplainerDashboard(
    ClassifierExplainer(RandomForestClassifier().fit(X_train, y_train), X_test, y_test)
).run()
```

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

### Initialize and run ExplainerHub

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/hub.md

Creates a hub instance with multiple dashboards and authentication, then starts the server.

```python
hub = ExplainerHub([db1, db2], logins=[['user', 'password']], secret_key="SECRET")
hub.run()
```

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

### Run ExplainerDashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/explainers.md

Launch an ExplainerDashboard with a pre-initialized explainer instance. This is the most basic way to start the dashboard.

```python
from explainerdashboard import ExplainerDashboard
ExplainerDashboard(explainer).run()
```

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

### Custom Dashboard Example

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/components.md

An example demonstrating how to build a custom dashboard using ExplainerComponent and ShapDependenceComponent.

```APIDOC
## Custom Dashboard Example

### Description
This example shows how to create a custom dashboard layout using `ExplainerComponent` and `ShapDependenceComponent`. It demonstrates initializing a component with specific configurations like hiding titles and categories, and setting an initial feature.

### Code
```python
from explainerdashboard.custom import *

class CustomDashboard(ExplainerComponent):
   def __init__(self, explainer, title="Custom Dashboard", name="None"):
      super().__init__(explainer, title, name=name)
      self.shap_dependence = ShapDependenceComponent(explainer, name=self.name+"dep",
                         hide_title=True, hide_cats=True, hide_highlight=True,
                         cats=True, col='Fare')

   def layout(self):
      return html.Div([
         self.shap_dependence.layout()
     ])

ExplainerDashboard(explainer, CustomDashboard).run()
```
```

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

### Run ExplainerDashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/dashboard_examples.ipynb

Starts the dashboard server for the provided explainer instance.

```python
ExplainerDashboard(explainer).run()
```

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

### Run ExplainerHub via CLI

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/hub.md

Starts the hub server using the command line interface.

```bash
$ explainerhub run hub.yaml
```

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

### Extended example with model training and dashboard configuration

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/index.md

A complete workflow including data loading, model training, explainer construction, and dashboard customization.

```python
from sklearn.ensemble import RandomForestClassifier

from explainerdashboard import ClassifierExplainer, ExplainerDashboard
from explainerdashboard.datasets import titanic_survive

X_train, y_train, X_test, y_test = titanic_survive()

model = RandomForestClassifier(n_estimators=50, max_depth=5)
model.fit(X_train, y_train)

explainer = ClassifierExplainer(
                model, X_test, y_test,
                # optional:
                cats=['Sex', 'Deck', 'Embarked'],
                labels=['Not survived', 'Survived'])

db = ExplainerDashboard(explainer, title="Titanic Explainer",
                    whatif=False, # you can switch off tabs with bools
                    shap_interaction=False,
                    decision_trees=False)
db.run(port=8051)
```

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

### Initialize and Run the Dash App

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/custom_examples.ipynb

Register component callbacks to the Dash app instance and start the server.

```python
app = JupyterDash(
    __name__, external_stylesheets=[dbc.themes.FLATLY], assets_url_path=""
)
app.title = "Titanic Explainer"
app.layout = layout

precision.register_callbacks(app)
shap_summary.register_callbacks(app)
shap_dependence.register_callbacks(app)
connector.register_callbacks(app)

app.run_server(port=8050)
```

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

### Install Testing Dependencies

Source: https://github.com/oegedijk/explainerdashboard/blob/master/CONTRIBUTING.md

Install additional libraries required for running the full test suite.

```bash
$ pip install -r requirements_testing.txt
```

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

### ClassifierExplainer Usage Example

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/explainers.md

Example demonstrating how to initialize a ClassifierExplainer and invoke various plotting methods.

```default
explainer = ClassifierExplainer(model, X, y, labels=['Not Survived', 'Survived'])
explainer.plot_confusion_matrix(cutoff=0.6)
explainer.plot_precision(quantiles=10, cutoff=0.6, multiclass=True)
explainer.plot_lift_curve(percentage=True)
explainer.plot_roc_auc(cutoff=0.7)
explainer.plot_pr_auc(cutoff=0.3)
```

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

### Example YAML Configuration Structure

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/hub.md

The structure of the generated hub.yaml file containing dashboard references and settings.

```default
explainerhub:
    title: ExplainerHub
    description: Showing dashboards for both model one and two
    masonry: false
    n_dashboard_cols: 3
    users_file: users.yaml
    db_users: null
    port: 8050
    kwargs: {}
    dashboards:
    - db1_dashboard.yaml
    - db2_dashboard.yaml
```

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

### Heroku Procfile

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/deployment.md

Define the command to start your web server in the `Procfile`. This tells Heroku how to launch your application.

```bash
web: gunicorn dashboard:app
```

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

### Run Dashboard with Production Servers

Source: https://github.com/oegedijk/explainerdashboard/blob/master/README.md

Commands to start the dashboard application using Gunicorn or Waitress.

```sh
    $ gunicorn dashboard:app
```

```sh
    $ waitress-serve dashboard:app
```

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

### Run dashboard with Gunicorn

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/dashboards.md

Command to start the dashboard server using Gunicorn.

```bash
gunicorn dashboard:server
```

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

### Run dashboard with Waitress

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/dashboards.md

Command to start the dashboard server using Waitress on Windows.

```bash
waitress-serve dashboard:server
```

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

### Deploying to Kubernetes (Ingress)

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/index.md

Configuration details for deploying an ExplainerDashboard behind a Kubernetes Ingress controller. This example shows a basic Ingress resource definition.

```yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: explainer-ingress
spec:
  rules:
  - host: explainer.yourdomain.com
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: explainer-service
            port:
              number: 8080

```

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

### Run Dashboard with Gunicorn (Basic)

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/deployment.md

Start the gunicorn server to host your dashboard. Ensure your Flask app is exposed as 'app' in 'dashboard.py'.

```bash
$ gunicorn dashboard:app
```

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

### Alternative Regression Models

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/dashboard_examples.ipynb

Examples of importing and initializing various regression models for use with the dashboard.

```python
from lightgbm.sklearn import LGBMRegressor

model = LGBMRegressor()

# from sklearn.linear_model import LinearRegression
# model = LinearRegression()

# from catboost import CatBoostRegressor
# model = CatBoostRegressor(iterations=100, learning_rate=0.1, verbose=0)

# from xgboost import XGBRegressor
```

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

### Run the ExplainerDashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/simplified_classifier_dashboard_example.ipynb

Start the ExplainerDashboard on a specified port. The dashboard will be accessible via the provided URL. Ensure the port is not in use.

```python
db.run(port=9000)
```

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

### ExplainerDashboard Startup Output

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/dashboard_examples.ipynb

This output shows the typical messages displayed when ExplainerDashboard starts, including environment detection, layout generation, and server startup information.

```text
Building ExplainerDashboard..\nDetected notebook environment, consider setting mode='external', mode='inline' or mode='jupyterlab' to keep the notebook interactive while the dashboard is running...\nGenerating layout...\nCalculating dependencies...\nReminder: you can store the explainer (including calculated dependencies) with explainer.dump('explainer.joblib') and reload with e.g. ClassifierExplainer.from_file('explainer.joblib')\nRegistering callbacks...\nStarting ExplainerDashboard on http://localhost:8050\nDash is running on http://0.0.0.0:8050/\n\nDash is running on http://0.0.0.0:8050/\n\nDash is running on http://0.0.0.0:8050/\n\n * Serving Flask app "explainerdashboard.dashboards" (lazy loading)\n * Environment: production\n\u001B[31m   WARNING: This is a development server. Do not use it in a production deployment.\n\u001B[2m   Use a production WSGI server instead.\n * Debug mode: off\n\n
```

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

### Run a basic ClassifierExplainer dashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/README.md

A minimal example to fit a model and launch the dashboard in one line.

```python
ExplainerDashboard(ClassifierExplainer(RandomForestClassifier().fit(X_train, y_train), X_test, y_test)).run()
```

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

### Create a Custom Dashboard with ShapDependenceComponent

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/components.md

Example of subclassing ExplainerComponent to build a custom dashboard layout with specific component configurations.

```python
from explainerdashboard.custom import *

class CustomDashboard(ExplainerComponent):
   def __init__(self, explainer, title="Custom Dashboard", name="None"):
      super().__init__(explainer, title, name=name)
      self.shap_dependence = ShapDependenceComponent(explainer, name=self.name+"dep",
                         hide_title=True, hide_cats=True, hide_highlight=True,
                         cats=True, col='Fare')

   def layout(self):
      return html.Div([
         self.shap_dependence.layout()
     ])

 ExplainerDashboard(explainer, CustomDashboard).run()
```

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

### Standalone Dockerfile for ExplainerDashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/deployment.md

Dockerfile for a standalone ExplainerDashboard deployment. Installs the library, copies scripts, generates the dashboard, and sets the run command.

```docker
FROM python:3.8

RUN pip install explainerdashboard

COPY generate_dashboard.py ./
COPY run_dashboard.py ./

RUN python generate_dashboard.py

EXPOSE 9050
CMD ["python", "./run_dashboard.py"]
```

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

### Hugging Face Space Dockerfile

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/deployment.md

Dockerfile for deploying an ExplainerDashboard to Hugging Face Spaces. Installs dependencies and sets up the application server.

```docker
FROM python:3.11-slim

WORKDIR /app
COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt
COPY . .

EXPOSE 7860
CMD ["gunicorn", "dashboard:app", "--bind", "0.0.0.0:7860"]
```

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

### ExplainerDashboard Startup Output (Instance)

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/dashboard_examples.ipynb

This output is similar to the previous startup output, indicating the dashboard is building, detecting the environment, and preparing to run.

```text
Building ExplainerDashboard..\nDetected notebook environment, consider setting mode='external', mode='inline' or mode='jupyterlab' to keep the notebook interactive while the dashboard is running...\nGenerating layout...\nCalculating dependencies...\nReminder: you can store the explainer (including calculated dependencies) with explainer.dump('explainer.joblib') and reload with e.g. ClassifierExplainer.from_file('explainer.joblib')\nRegistering callbacks...\nStarting ExplainerDashboard on http://localhost:8050\nDash is running on http://0.0.0.0:8050/\n\nDash is running on http://0.0.0.0:8050/\n\nDash is running on http://0.0.0.0:8050/\n\nDash is running on http://0.0.0.0:8050/\n\n * Serving Flask app "explainerdashboard.dashboards" (lazy loading)\n * Environment: production\n\u001B[31m   WARNING: This is a development server. Do not use it in a production deployment.\n\u001B[2m   Use a production WSGI server instead.\n * Debug mode: off\n\n
```

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

### Load dashboard from configuration

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/deployment.md

Initializes an ExplainerDashboard instance directly from a previously saved configuration file.

```python
from explainerdashboard import ExplainerDashboard

db = ExplainerDashboard.from_config("dashboard.yaml")
app = db.app.server
```

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

### Get Importances DataFrame

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/explainers.md

A wrapper function to get feature importances, either as SHAP values or permutation importances.

```APIDOC
## get_importances_df(kind='shap', topx=None, cutoff=None, pos_label=None)

### Description
A wrapper function for `get_mean_abs_shap_df()` and `get_permutation_importance_df()`. It allows retrieving feature importances based on either SHAP values or permutation importance.

### Method
Not applicable (this is a method call within an instance)

### Parameters
#### Path Parameters
None

#### Query Parameters
* **kind** (str) - Specifies the type of importance to retrieve. Accepts 'shap' or 'permutations'. Defaults to 'shap'.
* **topx** (int, optional) - Only display the `topx` highest features. Defaults to None.
* **cutoff** (float, optional) - Only display features with importance above `cutoff`. Defaults to None.
* **pos_label** (any, optional) - The positive class label. Defaults to None.

### Request Example
```python
shap_importances = explainer.get_importances_df(kind='shap', topx=5)
permutation_importances = explainer.get_importances_df(kind='permutations', cutoff=0.05)
```

### Response
#### Success Response (200)
* **pd.DataFrame** - A DataFrame containing the requested feature importances.

#### Response Example
```json
{
  "feature1": 0.45,
  "feature2": 0.28,
  "feature3": 0.15
}
```
```

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

### Initialize default ExplainerDashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/dashboards.md

Construct an Explainer instance and launch the dashboard with default compatible tabs.

```python
from explainerdashboard import ClassifierExplainer, ExplainerDashboard

explainer = ClassifierExplainer(model, X_test, y_test)
ExplainerDashboard(explainer).run()
```

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

### Custom Model Tab Example

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/index.md

Defines a custom tab for displaying model-specific information. This example shows how to create a new tab that can be integrated into the ExplainerDashboard.

```python
from explainerdashboard.custom import ExplainerComponent
from dash import html

class CustomModelTab(ExplainerComponent):
    def __init__(self, name, explainer, **kwargs):
        super().__init__(name=name, explainer=explainer, **kwargs)

    def _layout(self):
        return html.Div([
            html.H3("Custom Model Information"),
            html.P(f"Model index: {self.explainer.model_output_names()[0]}")
        ])

```

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

### Load ExplainerDashboard from Configuration

Source: https://github.com/oegedijk/explainerdashboard/blob/master/RELEASE_NOTES.md

Use this class method to initialize a dashboard directly from a YAML configuration file.

```python
ExplainerDashboard.from_config("dashboard.yaml")
```

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

### Custom Predictions Tab Example

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/index.md

Provides an example of a custom tab for displaying prediction-related information. This allows for tailored views of prediction data within the dashboard.

```python
from explainerdashboard.custom import ExplainerComponent
from dash import html

class CustomPredictionsTab(ExplainerComponent):
    def __init__(self, name, explainer, **kwargs):
        super().__init__(name=name, explainer=explainer, **kwargs)

    def _layout(self):
        return html.Div([
            html.H3("Custom Prediction Details"),
            html.P(f"Number of predictions: {len(self.explainer.X)}")
        ])

```

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

### Gunicorn Start Command for Fly.io

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/deployment.md

Specifies the command to start the web application using gunicorn. Ensure the port matches the internal port configured during `fly launch`.

```bash
web: gunicorn dashboard:app --bind 0.0.0.0:${PORT:-8080}
```

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

### Deploying to Fly.io

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/index.md

Instructions and considerations for deploying an ExplainerDashboard application to Fly.io. This typically involves creating a `fly.toml` configuration file and building a Docker image.

```toml
[render]
  # ... other configurations
  [http_service]
    internal_port = 8080
    force_https = true
    [http_service.concurrency]
      max_per_server = 50
      soft_limit = 50
      hard_limit = 100

[[services.ports]]
  port = "80"
  handlers = ["http"]
[[services.ports]]
  port = "443"
  handlers = ["http"]

```

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

### GET get_permutation_importances_df

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/explainers.md

Retrieves a dataframe containing features ordered by their permutation importance.

```APIDOC
## GET get_permutation_importances_df

### Description
Returns a dataframe with features ordered by permutation importance.

### Parameters
#### Query Parameters
- **topx** (int) - Optional - Only return topx most important features
- **cutoff** (float) - Optional - Only return features with importance of at least cutoff
- **pos_label** (any) - Optional - Positive label for classification

### Response
#### Success Response (200)
- **importance_df** (pd.DataFrame) - Dataframe containing permutation importances
```

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

### Run a RegressionExplainer dashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/README.md

Example for regression models, including setting target units.

```python
X_train, y_train, X_test, y_test = titanic_fare()
model = RandomForestRegressor().fit(X_train, y_train)

explainer = RegressionExplainer(model, X_test, y_test,
                                cats=['Deck', 'Embarked', 'Sex'],
                                descriptions=feature_descriptions,
                                units = "$", # defaults to ""
                                )

ExplainerDashboard(explainer).run()
```

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

### Create Basic ExplainerDashboard

Source: https://context7.com/oegedijk/explainerdashboard/llms.txt

Launch a full interactive dashboard with default tabs for model explanation. Configure which tabs to display and run the dashboard on a specified port.

```python
from explainerdashboard import ClassifierExplainer, ExplainerDashboard
from sklearn.ensemble import RandomForestClassifier
from explainerdashboard.datasets import titanic_survive

# Prepare data and model
X_train, y_train, X_test, y_test = titanic_survive()
model = RandomForestClassifier(n_estimators=50, max_depth=5).fit(X_train, y_train)

# Create explainer
explainer = ClassifierExplainer(model, X_test, y_test)

# Create and run dashboard with all default tabs
dashboard = ExplainerDashboard(
    explainer,
    title="Titanic Survival Classifier",
    description="Explains passenger survival predictions",
    # Control which tabs to show
    importances=True,      # Feature importances tab
    model_summary=True,    # Model performance metrics tab
    contributions=True,    # Individual predictions tab
    whatif=True,           # What-if analysis tab
    shap_dependence=True,  # SHAP dependence plots tab
    shap_interaction=False,  # Disable slow interaction tab
    decision_trees=True    # Decision trees visualization
)

# Run on specified port
dashboard.run(port=8050)
```

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

### GET get_shap_values_df

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/explainers.md

Calculates and returns SHAP values for the model using the shap library.

```APIDOC
## GET get_shap_values_df

### Description
Calculates SHAP values using the shap library.

### Parameters
#### Query Parameters
- **pos_label** (any) - Optional - Positive label for classification

### Response
#### Success Response (200)
- **shap_values** (pd.DataFrame) - SHAP values calculated for the model
```

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

### View Server Output

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/custom_examples.ipynb

Expected console output when the Dash application starts successfully.

```text
Output:
Dash app running on http://127.0.0.1:8050/
```

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

### Initialize RandomForest Explainer

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/explainer_examples.ipynb

Sets up a RandomForestClassifier and initializes the corresponding explainer.

```python
from sklearn.ensemble import RandomForestClassifier
from explainerdashboard.datasets import (
    titanic_survive,
    titanic_names,
    feature_descriptions,
)

X_train, y_train, X_test, y_test = titanic_survive()
train_names, test_names = titanic_names()

model = RandomForestClassifier(n_estimators=50, max_depth=5)
model.fit(X_train, y_train)

explainer = ClassifierExplainer(
    model, X_test, y_test, cats=["Sex", "Deck", "Embarked"], idxs=test_names
)
```

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

### Save and Load Explainer

Source: https://context7.com/oegedijk/explainerdashboard/llms.txt

Demonstrates how to save an explainer object to disk using 'dump' and load it back using 'from_file' for faster dashboard startup. Supports multiple file formats.

```python
from explainerdashboard import ClassifierExplainer, ExplainerDashboard
from sklearn.ensemble import RandomForestClassifier
from explainerdashboard.datasets import titanic_survive

# Create and train model
X_train, y_train, X_test, y_test = titanic_survive()
model = RandomForestClassifier(n_estimators=50).fit(X_train, y_train)

# Create explainer and calculate all properties (can take time)
explainer = ClassifierExplainer(model, X_test, y_test)

# Pre-calculate expensive properties
explainer.calculate_properties()  # Calculate all
# Or calculate specific properties:
# explainer.shap_values()
# explainer.permutation_importances()
# explainer.shap_interaction_values()

# Save to disk (uses joblib by default)
explainer.dump('explainer.joblib')

# Load from disk - much faster than recalculating
loaded_explainer = ClassifierExplainer.from_file('explainer.joblib')

# Alternative formats
explainer.dump('explainer.pkl')  # Pickle format
explainer.dump('explainer.dill')  # Dill format (for complex objects)

# Create dashboard from loaded explainer
dashboard = ExplainerDashboard(loaded_explainer, title="Loaded Dashboard")
dashboard.run()
```

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

### Run ExplainerDashboard on Specific Port

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/dashboard_examples.ipynb

Launch the ExplainerDashboard on a specified port, for example, port 8051.

```python
db.run(8051)
```

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

### ExplainerDashboard Initialization and Usage

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/dashboard_examples.ipynb

Demonstrates how to initialize and run the ExplainerDashboard for classification tasks, including customizing components.

```APIDOC
## ExplainerDashboard Initialization and Usage

### Description
This section details the initialization of the `ExplainerDashboard` class for classification models. It shows how to enable a simple mode that displays a predefined set of components and how to customize specific components.

### Method
`ExplainerDashboard(explainer, simple=False, classifier_custom_component=None, ...)`

### Endpoint
N/A (Class initialization)

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
None

### Request Example
```python
# Basic initialization for simple mode
db = ExplainerDashboard(explainer, simple=True)
db.run()

# With a customized classification component
db = ExplainerDashboard(explainer, simple=True, classifier_custom_component="precision_graph")
db.run()
```

### Response
#### Success Response (200)
N/A (This is a class initialization and method call, not an HTTP endpoint)

#### Response Example
N/A
```

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

### Get Decision Path Dataframe

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/explainer_examples.ipynb

Retrieves the decision path details for a specific tree and instance.

```python
explainer.get_decisionpath_df(tree_idx=20, index=name)
```

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

### Run Custom Dashboard from YAML

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/cli.md

Launch a custom explainer dashboard using a previously created YAML configuration file.

```bash
$ explainerdashboard run dashboard.yaml
```

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

### Initialize ExplainerDashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/dashboard_examples.ipynb

Create an ExplainerDashboard instance from an explainer object. This pre-calculates necessary metrics for the dashboard.

```python
db = ExplainerDashboard(explainer)
```

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

### Run Dashboard with Waitress

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/deployment.md

Start the waitress server to host your dashboard, binding to port 8050.

```bash
$ waitress-serve --port=8050 dashboard:app
```

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

### Configure ClassifierExplainer with LogisticRegression

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/explainers.md

Example of initializing a ClassifierExplainer with specific background data and model output settings.

```python
from sklearn.linear_model import LogisticRegression

model = LogisticRegression()
model.fit(X_train, y_train)

explainer = ClassifierExplainer(model, X_test, y_test,
                                    shap='linear',
                                    X_background=X_train,
                                    model_output='logodds')
ExplainerDashboard(explainer).run()
```

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

### ExplainerHub.from_config

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/hub.md

Instantiate an ExplainerHub based on a configuration file or dictionary.

```APIDOC
## classmethod from_config

### Description
Instantiate an ExplainerHub based on a config file or dictionary.

### Parameters
- **config** (Union[dict, str, Path]) - Required - Either a dict or a .yaml config file to load.
- **update_params** (dict) - Optional - Additional kwargs to override stored settings.

### Response
- **Returns** (ExplainerHub) - New instance of ExplainerHub according to the config.
```

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

### ExplainerHub Initialization

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/hub.md

Parameters for configuring and initializing an ExplainerHub instance.

```APIDOC
## ExplainerHub Initialization

### Description
Initializes an ExplainerHub to manage multiple ExplainerDashboard instances.

### Parameters
- **dashboards** (List[ExplainerDashboard]) - Required - List of ExplainerDashboard to include in ExplainerHub.
- **title** (str) - Optional - Title to display. Defaults to “ExplainerHub”.
- **description** (str) - Optional - Short description of ExplainerHub.
- **masonry** (bool) - Optional - Lay out dashboard cards in fluid bootstrap masonry responsive style. Defaults to False.
- **n_dashboard_cols** (int) - Optional - Organize cards in rows and columns. Defaults to 3 columns.
- **users_file** (Path) - Optional - A .yaml or .json file used to store user and (hashed) password data. Defaults to ‘users.yaml’.
- **user_json** (Path) - Optional - Deprecated. A .json file used to store user and (hashed) password data.
- **logins** (List[List[str, str]]) - Optional - List of [‘login’, ‘password’] pairs.
- **db_users** (dict) - Optional - Dictionary limiting access to certain dashboards to a subset of users.
- **dbs_open_by_default** (bool) - Optional - Only force logins for dashboard with defined db_users. Defaults to False.
- **port** (int) - Optional - Port to run hub on. Defaults to 8050.
- **min_height** (int) - Optional - Defaults to 3000 pixels.
- **secret_key** (str) - Required - Flask secret key to pass to dashboard in order to persist logins.
- **no_index** (bool) - Optional - Do not add the “/” route, but only mount the dashboards.
- **bootstrap** (str) - Optional - URL with custom bootstrap css.
- **fluid** (bool) - Optional - Let the bootstrap container fill the entire width of the browser. Defaults to True.
- **base_route** (str) - Optional - Base route for dashboard. Defaults to “dashboards”.
- **index_to_base_route** (bool) - Optional - Dispatches Hub to “/base_route/index” instead of the default “/”.
- **static_to_base_route** (bool) - Optional - Dispatches Hub to “/base_route/static” instead of the default “/static”.
- **max_dashboards** (int) - Optional - Max number of dashboards in the hub.
- **add_dashboard_route** (bool) - Optional - Open a route /add_dashboard and /remove_dashboard.
- **add_dashboard_pattern** (str) - Optional - A str pattern with curly brackets for dashboard file location.
```

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

### Get Decision Path Summary

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/explainer_examples.ipynb

Retrieves a summary of the decision path adjustments for a specific tree and instance.

```python
explainer.get_decisionpath_summary_df(tree_idx=5, index=name)
```

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

### Constructor: ExplainerDashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/dashboards.md

Initializes a new ExplainerDashboard instance to visualize model explanations.

```APIDOC
## Constructor: ExplainerDashboard

### Description
Creates an explainerdashboard out of an Explainer object. The dashboard can be configured as a single-page or multi-tab layout with various display modes.

### Parameters
#### Request Body
- **explainer** (BaseExplainer) - Required - The explainer object to visualize.
- **tabs** (ExplainerComponent | List[ExplainerComponent]) - Optional - Single component or list of components to display.
- **title** (str) - Optional - Title of the dashboard, defaults to 'Model Explainer'.
- **mode** (str) - Optional - Type of dash server to start: 'dash', 'inline', 'jupyterlab', or 'external'.
- **fluid** (bool) - Optional - Whether to stretch the layout to available space, defaults to True.
- **width** (int) - Optional - Width of notebook output cell in pixels, defaults to 1000.
- **height** (int) - Optional - Height of notebook output cell in pixels, defaults to 800.
- **bootstrap** (str) - Optional - Link to bootstrap URL.
- **external_stylesheets** (List[str]) - Optional - Additional external stylesheets.
```

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

### Construct and run a basic ClassifierExplainer

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/index.md

Initialize an explainer object with a model and test data, then launch the dashboard.

```python
from explainerdashboard import ClassifierExplainer, ExplainerDashboard
explainer = ClassifierExplainer(model, X_test, y_test)
```

```python
ExplainerDashboard(explainer).run()
```

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

### Build Explainer from Configuration (CLI)

Source: https://context7.com/oegedijk/explainerdashboard/llms.txt

Build explainers from YAML configuration files, useful for CI/CD pipelines. Can specify both explainer and dashboard configurations.

```bash
# Build explainer from configuration (for CI/CD)
explainerdashboard build explainer.yaml
```

```bash
# Build with specific dashboard configuration
explainerdashboard build explainer.yaml dashboard.yaml
```

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

### Persist and restore hub configuration

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/hub.md

Saves the current hub setup to a YAML file and restores it later.

```python
hub.to_yaml('hub.yaml')
```

```python
hub2 = ExplainerHub.from_config('hub.yaml')
```

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

### Initialize and Run ExplainerDashboard

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/custom_examples.ipynb

Use this snippet to initialize an ExplainerDashboard with a given explainer and custom dashboard class. The bootstrap theme and header visibility can be configured. The dashboard will run on http://localhost:8050.

```python
from explainerdashboard import ExplainerDashboard

ExplainerDashboard(
    explainer,
    CustomDashboard,
    boostrap=dbc.themes.FLATLY,
    hide_header=True
).run()
```

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

### Get Base Explainer Metrics

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/explainers.md

Retrieves a dictionary of performance metrics for the explainer. This method is implemented by both ClassifierExplainer and RegressionExplainer.

```python
metrics()
```

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

### Create and Activate Virtual Environment

Source: https://github.com/oegedijk/explainerdashboard/blob/master/CONTRIBUTING.md

Initialize a new Python virtual environment to isolate project dependencies.

```bash
$ python -m venv venv
$ source venv/bin/activate
```

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

### Server Request Logs

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/simplified_classifier_dashboard_example.ipynb

Sample output showing HTTP GET and POST requests handled by the ExplainerDashboard server.

```text
192.168.1.21 - - [07/May/2021 15:47:12] "GET /_dash-component-suites/dash_renderer/react-dom@16.v1_9_1m1618962237.14.0.min.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:12] "GET /_dash-component-suites/dash_renderer/polyfill@7.v1_9_1m1618962237.8.7.min.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:12] "GET /_dash-component-suites/dash_renderer/prop-types@15.v1_9_1m1618962237.7.2.min.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:12] "GET /_dash-component-suites/dash_core_components/dash_core_components.v1_16_0m1618962238.min.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:12] "GET /_dash-component-suites/dash_core_components/dash_core_components-shared.v1_16_0m1618962238.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:12] "GET /_dash-component-suites/dash_table/bundle.v4_11_3m1618962241.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:12] "GET /_dash-component-suites/dash_bootstrap_components/_components/dash_bootstrap_components.v0_12_0m1618962292.min.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:12] "GET /_dash-component-suites/dash_html_components/dash_html_components.v1_1_3m1618962240.min.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:12] "GET /_dash-component-suites/dash_renderer/dash_renderer.v1_9_1m1618962237.min.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:13] "GET /_dash-layout HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:13] "GET /_dash-dependencies HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:13] "GET /assets/favicon.ico?m=1620419845.4037383 HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "GET /_dash-component-suites/dash_core_components/async-dropdown.v1_16_0m1617903285.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "GET /_dash-component-suites/dash_core_components/async-graph.v1_16_0m1617903285.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "GET /_dash-component-suites/dash_core_components/async-slider.v1_16_0m1617903285.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "GET /_dash-component-suites/dash_core_components/async-plotlyjs.v1_16_0m1617903285.js HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:14] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:15] "POST /_dash-update-component HTTP/1.1" 204 -
192.168.1.21 - - [07/May/2021 15:47:15] "POST /_dash-update-component HTTP/1.1" 204 -
192.168.1.21 - - [07/May/2021 15:47:15] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:15] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:15] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:15] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:15] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:16] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:16] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:16] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:16] "POST /_dash-update-component HTTP/1.1" 200 -
192.168.1.21 - - [07/May/2021 15:47:16] "POST /_dash-update-component HTTP/1.1" 200 -
```

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

### Run ExplainerDashboard from CLI

Source: https://context7.com/oegedijk/explainerdashboard/llms.txt

Use the command-line interface to run dashboards from explainer files or YAML configurations. Options include specifying ports, disabling browser launch, and running on AWS SageMaker.

```bash
# Run dashboard from explainer file
explainerdashboard run explainer.joblib
```

```bash
# Run dashboard from YAML configuration
explainerdashboard run dashboard.yaml
```

```bash
# Run with specific port and no browser launch
explainerdashboard run dashboard.yaml --port 8051 --no-browser
```

```bash
# Run on AWS SageMaker with proxy settings
explainerdashboard run dashboard.yaml --sagemaker
```

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

### Customize ExplainerDashboard Classification Component

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/dashboard_examples.ipynb

Customize the classification component by passing a string to `classifier_custom_component`. For example, use 'precision_graph' for the PrecisionComponent.

```python
db = ExplainerDashboard(
    explainer, simple=True, classifier_custom_component="precision_graph"
)
db.run()
```

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

### Configure Dashboard Tabs

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/dashboards.md

Example of mixing component classes, string identifiers, and custom tab instances for the tabs parameter.

```python
tabs=[ImportancesTab, “contributions”, custom_tab]
```

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

### ExplainerDashboard.from_config

Source: https://github.com/oegedijk/explainerdashboard/blob/master/docs/source/cli.md

Loads a dashboard from a configuration YAML file or directly from an explainer file.

```APIDOC
## ExplainerDashboard.from_config

### Description
Loading a dashboard from a configuration .yaml file. You can either pass both an explainer and a yaml file generated with ExplainerDashboard.to_yaml("dashboard.yaml"), or just the .yaml if the explainerfile was specified in to_yaml, or load the explainerfile separately.

### Method
Class Method

### Endpoint
N/A

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
None

### Request Example
```python
# Example 1: Passing explainer and yaml file
db = ExplainerDashboard.from_config(explainer, "dashboard.yaml")

# Example 2: Passing only yaml file (if explainerfile was specified in to_yaml)
db = ExplainerDashboard.from_config("dashboard.yaml")

# Example 3: Passing explainer file and yaml file separately
db = ExplainerDashboard.from_config("explainer.joblib", "dashboard.yaml")

# Example 4: Overriding parameters
db = ExplainerDashboard.from_config("dashboard.yaml", update_params={"title": "My Custom Dashboard"})
```

### Response
#### Success Response (200)
An instance of ExplainerDashboard.

#### Response Example
```python
# Example of a returned ExplainerDashboard instance (conceptual)
print(db)
```
```

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

### Train Model and Build Explainer

Source: https://github.com/oegedijk/explainerdashboard/blob/master/notebooks/custom_examples.ipynb

Prepare data, train a Random Forest model, and initialize the ClassifierExplainer.

```python

X_train, y_train, X_test, y_test = titanic_survive()
model = RandomForestClassifier(n_estimators=50, max_depth=15).fit(X_train, y_train)
explainer = ClassifierExplainer(
    model,
    X_test,
    y_test,
    cats=["Sex", "Deck", "Embarked"],
    descriptions=feature_descriptions,
    labels=["Not survived", "Survived"],
)

explainer.plot_contributions(0)
```