SDMetrics

https://docs.sdv.dev/sdmetrics
Admin
SDMetrics is an open source Python library for evaluating tabular synthetic data through statistical...
Tokens:47,952
Snippets:593
Trust Score:9
Update:6 months ago
Show doc for...
Tokens:
Raw
### Example Properties Dictionary

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/multi-table-api

Example output from `get_properties()`, showing scores for 'Data Validity', 'Data Structure', and 'Relationship Validity'.

```json
{
  "Data Validity": 1.0,
  "Data Structure": 1.0,
  "Relationship Validity": 1.0
}
```

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

### Install Development Dependencies

Source: https://docs.sdv.dev/sdmetrics/resources/contributions/development

Install the SDMetrics library in development mode using the provided Makefile. It is recommended to use a virtual environment.

```bash
cd SDMetrics/
make install-develop
```

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

### Install SDMetrics via Package Managers

Source: https://docs.sdv.dev/sdmetrics/getting-started/installation

Use these commands to install the library using either pip or conda.

```shell
pip install sdmetrics
```

```shell
conda install -c pytorch -c conda-forge sdmetrics
```

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

### Example Detailed Property Results

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/multi-table-api

Example output from `get_details()` for 'Data Validity' on the 'users' table, showing column-specific metrics and scores.

```text
Table        Column        Metric                 Score
users        user_id       KeyUniqueness          1.0   
users        age           BoundaryAdherence      1.0     
users        height        BoundaryAhderence      1.0            
users        card_type     CategoryAdherence      1.0
...
```

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

### Example Overall Score

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/multi-table-api

An example of the output from `get_score()`, indicating a perfect score of 1.0.

```text
1.0
```

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

### Install pomegranate dependency

Source: https://docs.sdv.dev/sdmetrics/data-metrics/metrics-in-beta/data-likelihood/bnlikelihood

The pomegranate library is a required dependency for the BNLikelihood metric.

```bash
pip install pomegranate
```

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

### Quality Report Output Example

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality

This is an example of the output generated by the QualityReport. It shows scores for Column Shapes and Column Pair Trends, leading to an Overall Score.

```text
Generating report ...

(1/2) Evaluating Column Shapes: |██████████| 9/9 [00:00<00:00, 273.13it/s]|
Column Shapes Score: 89.11%

(2/2) Evaluating Column Pair Trends: |██████████| 36/36 [00:00<00:00, 57.42it/s]|
Column Pair Trends Score: 88.3%

Overall Score (Average): 88.7%
```

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

### Verify SDMetrics Version

Source: https://docs.sdv.dev/sdmetrics/getting-started/installation

Check the installed version of the library to ensure it matches the latest release.

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

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

### Diagnostic Report Output Example

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report

This is an example of the output you might see when generating a diagnostic report. It shows the progress and final scores for data validity and structure.

```text
Generating report ...

(1/2) Evaluating Data Validity: |██████████| 9/9 [00:00<00:00, 458.92it/s]|
Data Validity Score: 100.0%

(2/2) Evaluating Data Structure: |██████████| 1/1 [00:00<00:00, 104.60it/s]|
Data Structure Score: 100.0%

Overall Score (Average): 100.0%
```

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

### LinearRegression and MLPRegressor Compute Method

Source: https://docs.sdv.dev/sdmetrics/data-metrics/metrics-in-beta/ml-efficacy-single-table/regression

This example demonstrates how to compute regression metrics using the `compute` method for both LinearRegression and MLPRegressor models.

```APIDOC
## POST /compute

### Description
Computes the regression metric for the given data.

### Method
POST

### Endpoint
/compute

### Parameters
#### Query Parameters
- **test_data** (pandas.DataFrame) - Required - A pandas.DataFrame containing the full data to test on. This should include the column that you are trying to predict.
- **train_data** (pandas.DataFrame) - Required - A pandas.DataFrame containing the full data to train on. This should include the column that you are trying to predict.
- **target** (string) - Required - A string representing the name of the column that you want to predict. This must be a numerical column.
- **metadata** (dict) - Optional - A description of the dataset.

### Request Example
```python
from sdmetrics.single_table import LinearRegression, MLPRegressor

LinearRegression.compute(
    test_data=real_data,
    train_data=synthetic_data,
    target='numerical_column_name',
    metadata=metadata
)
```

### Response
#### Success Response (200)
- **score** (float) - The r2 test score.

#### Response Example
```json
{
  "score": 0.85
}
```
```

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

### Define Sequential Table Metadata

Source: https://docs.sdv.dev/sdmetrics/getting-started/metadata/sequential-metadata

A complete example of a metadata dictionary for a sequential table, including sequence keys, indices, and column definitions.

```python
{
    "sequence_key": "Patient ID",
    "sequence_index": "Time",
    "columns": {
        "Patient ID": {
            "sdtype": "id",
            "regex_format": "ID_[0-9]{3}"
        },
        "Address": {
            "sdtype": "address",
            "pii": True
        },
        "Smoker": {
            "sdtype": "boolean"
        },
        "Time": {
            "sdtype": "datetime",
            "datetime_format": "%m/%d/%Y"
        },
        "Heart Rate": {
            "sdtype": "categorical"
        },
        "Systolic BP": {
            "sdtype": "numerical"
        }
    }
}
```

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

### Define Full Table Metadata

Source: https://docs.sdv.dev/sdmetrics/getting-started/metadata/single-table-metadata

A complete example of a metadata dictionary for a single table, including primary key and various column types.

```python
{
    "primary_key": "user_id",
    "columns": {
        "user_id": {
            "sdtype": "id",
            "regex_format": "U_[0-9]{3}"
        },
        "age": {
            "sdtype": "numerical"
        },
        "address": {
            "sdtype": "address",
            "pii": True
        }, 
        "tier": {
            "sdtype": "categorical"
        },
        "active": {
            "sdtype": "boolean"
        },
        "paid_amt": {
            "sdtype": "numerical"
        },
        "renew_date": {
            "sdtype": "datetime",
            "datetime_format": "%Y-%m-%d"
        }
    }
}
```

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

### Multi-Table Iteration Result

Source: https://docs.sdv.dev/sdmetrics/resources/contributions/defining-your-metric

Example of a nested dictionary structure for metrics applied to multi-table datasets.

```python
{
    'users': {
        ('column_1', 'column_2'): 0.1234,
        ('column_1', 'column_3'): 0.5678
    },
    'sessions': ...
}
```

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

### Example Output Structure

Source: https://docs.sdv.dev/sdmetrics/data-metrics/metrics-in-beta/equalizedoddsimprovement

The compute_breakdown method returns a dictionary containing the overall score and detailed prediction counts for both real and synthetic datasets.

```json
{
  'score': 0.5071,
  'real_training_data': {
    'equalized_odds': 0.8857,
    'prediction_counts_validation': {
      'Asian=True': {
        'true_positive': 10, # TPR = 0.4
        'false_positive': 5, # FPR = 0.0588
        'true_negative': 80,
        'false_negative': 15
      },
      'Asian=False': {
        'true_positive': 8, # TPR = 0.2857
        'false_positive': 4, # FPR = 0.0506
        'true_negative': 75,
        'false_negative': 20
      }
    }
  },
  'synthetic_data': {
    'equalized_odds': 0.9,
    # same format as the real_training_data
  }
}
```

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

### Multi Table Report Generation Progress

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/multi-table-api

Example output showing the progress and scores during the generation of a multi-table diagnostic report. Scores range from 0 to 100%.

```text
Generating report ...

(1/3) Evaluating Data Validity: |██████████| 15/15 [00:00<00:00, 603.69it/s]|
Data Validity Score: 100.0%

(2/3) Evaluating Data Structure: |██████████| 2/2 [00:00<00:00, 151.49it/s]|
Data Structure Score: 100.0%

(3/3) Evaluating Relationship Validity: |██████████| 1/1 [00:00<00:00, 68.51it/s]|
Relationship Validity Score: 100.0%

Overall Score (Average): 100.0%
```

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

### Get Property Details

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/single-table-api

Retrieve a pandas DataFrame containing detailed metrics for a specific property.

```python
report.get_details(property_name='Data Validity')
```

```python
Column        Metric                 Score
user_id       KeyUniqueness          1.0   
age           BoundaryAdherence      1.0     
height        BoundaryAhderence      1.0            
card_type     CategoryAdherence      1.0
...
```

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

### Generate Diagnostic Report

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report

Use this snippet to generate a diagnostic report for your dataset. Ensure you have the 'sdmetrics' library installed and your real and synthetic data, along with metadata, are prepared.

```python
from sdmetrics.reports.single_table import DiagnosticReport

report = DiagnosticReport()
report.generate(real_data, synthetic_data, metadata)
```

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

### Run Full Test Suite

Source: https://docs.sdv.dev/sdmetrics/resources/contributions/development

Execute the complete test suite for the SDMetrics library using the provided Makefile. This verifies that your changes do not break existing functionality.

```bash
make test
```

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

### Initialize Multi Table QualityReport

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/multi-table-api

Import and instantiate the QualityReport class from the multi_table module.

```python
from sdmetrics.reports.multi_table import QualityReport

report = QualityReport()
```

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

### Initialize QualityReport

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/single-table-api

Import and instantiate the QualityReport object from the single table reports module.

```python
from sdmetrics.reports.single_table import QualityReport

report = QualityReport()
```

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

### Retrieve Property Scores

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/single-table-api

Get a dictionary containing the scores for each evaluated property.

```python
report.get_properties()
```

```python
{
  'Data Validity': 1.0,
  'Data Structure': 1.0,
}
```

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

### QualityReport() - Initialization

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/multi-table-api

Initializes a QualityReport object. This is the first step before generating any reports.

```APIDOC
## QualityReport() - Initialization

### Description
Create your report object by importing it from the multi table reports module.

### Method
```python
from sdmetrics.reports.multi_table import QualityReport

report = QualityReport()
```
```

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

### Initialize DiagnosticReport

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/single-table-api

Import and instantiate the DiagnosticReport class to begin the evaluation process.

```python
from sdmetrics.reports.single_table import DiagnosticReport

report = DiagnosticReport()
```

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

### Retrieve Property Scores

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/single-table-api

Get a pandas DataFrame listing each evaluated property and its score.

```python
report.get_properties()
```

```python
Property                Score
Column Shapes           0.8278
Column Pair Trends      0.7872
```

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

### get_properties() - Get Property Scores

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/multi-table-api

Retrieves the scores for each individual property evaluated in the report.

```APIDOC
## get_properties() - Get Property Scores

### Description
Use this method at any point to retrieve each property that the report evaluated.

### Method
```python
report.get_properties()
```

### Parameters
None

### Request Example
```python
report.get_properties()
```

### Response
#### Success Response (200)
- **properties** (pandas.DataFrame) - A pandas.DataFrame that lists each property name and its associated score.

#### Response Example
```
Property                        Score
Column Shapes                   0.792262
Column Pair Trends              0.424967
Cardinality                     0.800000
Intertable Trends               0.482407
```
```

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

### Retrieve Overall Score

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/single-table-api

Get the overall quality score as a float between 0 and 1.

```python
report.get_score()
```

```python
1.0
```

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

### Clone SDMetrics Repository

Source: https://docs.sdv.dev/sdmetrics/resources/contributions/development

Clone the SDMetrics repository from GitHub to your local machine. This is the first step in setting up your development environment.

```bash
git clone git@github.com:your_username/SDMetrics
```

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

### Load Quality Report

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/multi-table-api

Loads a QualityReport object from a specified file. Ensure the filepath points to a valid '.pkl' file saved previously.

```python
from sdmetrics.reports.multi_table import QualityReport

report = QualityReport.load('results/quality_report.pkl')
```

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

### GET get_visualization

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/multi-table-api

Generates a plotly visualization for a specific property and table within the quality report.

```APIDOC
## GET get_visualization

### Description
Use this method to visualize the details about a property.

### Parameters
#### Path Parameters
- **property_name** (string) - Required - The name of the property. One of: 'Column Shapes', 'Column Pair Trends', 'Cardinality' or 'Intertable Trends'
- **table_name** (string) - Required - The name of the table

### Response
- **plotly.Figure** (object) - A plotly.Figure object representing the visualization.
```

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

### Commit and Push Changes

Source: https://docs.sdv.dev/sdmetrics/resources/contributions/development

Stage all changes, commit them with a descriptive message, and push your feature branch to GitHub. This prepares your work for a pull request.

```bash
git add .
git commit -m "<Descriptive details about the new metric>"
git push origin issue-123-your-new-metric-name
```

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

### Get Evaluated Properties

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/multi-table-api

Retrieve a dictionary listing each property evaluated in the report and its corresponding score.

```python
report.get_properties()
```

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

### Load DiagnosticReport from file

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/single-table-api

Loads a previously saved DiagnosticReport object from a .pkl file.

```python
from sdmetrics.reports.single_table import DiagnosticReport

report = DiagnosticReport.load('results/diagnostic_report.pkl')
```

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

### Retrieve Detailed Property Metrics

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/multi-table-api

Get a detailed breakdown of scores for a specific property, optionally filtered by table name.

```python
report.get_details(
    property_name='Column Shapes',
    table_name='users')
```

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

### Import DiagnosticReport

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/multi-table-api

Import the DiagnosticReport class from the sdmetrics.reports.multi_table module to begin creating reports.

```python
from sdmetrics.reports.multi_table import DiagnosticReport

report = DiagnosticReport()
```

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

### Get Overall Report Score

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/multi-table-api

Retrieve the overall quality score of the synthetic data as a floating-point value between 0 and 1.

```python
report.get_score()
```

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

### POST QualityReport.load

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/multi-table-api

Loads a previously saved QualityReport object from a file.

```APIDOC
## POST QualityReport.load

### Description
Load the report from the file.

### Parameters
#### Request Body
- **filepath** (string) - Required - The name of the file where the report is stored

### Response
- **QualityReport** (object) - A QualityReport object.
```

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

### Load QualityReport from File

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/single-table-api

Loads a QualityReport object from a specified file path. This allows you to resume analysis or access previously saved report data.

```python
from sdmetrics.reports.single_table import QualityReport

report = QualityReport.load('results/quality_report.pkl')
```

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

### get_details(property_name, table_name) - Get Property Details

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/multi-table-api

Retrieves detailed scores for a specific property, optionally filtered by table name.

```APIDOC
## get_details(property_name, table_name) - Get Property Details

### Description
Use this method to get more details about a particular property.

### Method
```python
report.get_details(property_name, table_name)
```

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
* **property_name** (str) - Required - A string with the name of the property. One of: `'Column Shapes'`, `'Column Pair Trends'`, `'Cardinality'` or `'Intertable Trends'`
* **table_name** (str) - Optional - A string with the name of the table. If provided, you'll receive filtered results for the table.

### Request Example
```python
report.get_details(
    property_name='Column Shapes',
    table_name='users')
```

### Response
#### Success Response (200)
- **details** (pandas.DataFrame) - A pandas.DataFrame that returns more details about the property for the given table.

#### Response Example
```
Table        Column         Metric         Score
users        purchase_amt   KSComplement   0.880
users        card_type      TVComplement   0.690
users        start_date     KSComplement   0.790
...
```
```

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

### Get Visualization

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/multi-table-api

Retrieves a Plotly Figure object for visualizing property details. Supports 'Data Validity' and 'Relationship Validity' properties.

```APIDOC
## GET /get_visualization

### Description
Use this method to visualize the details about a property.

### Method
GET

### Endpoint
/get_visualization

### Parameters
#### Query Parameters
- **property_name** (string) - Required - The name of the property to visualize. Supported values: 'Data Validity', 'Relationship Validity'.
- **table_name** (string) - Required - The name of the table for which to generate the visualization.

### Request Example
```python
# Assuming 'report' is an instance of DiagnosticReport
fig = report.get_visualization(property_name='Data Validity', table_name='users')
fig.show()
```

### Response
#### Success Response (200)
- **plotly.Figure** (object) - A Plotly Figure object representing the visualization.
```

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

### DiagnosticReport()

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/single-table-api

Initializes a DiagnosticReport object. This is the first step before generating any reports.

```APIDOC
## Initialize DiagnosticReport

### Description
Create your report object by importing it from the single table reports module.

### Method
```python
from sdmetrics.reports.single_table import DiagnosticReport

report = DiagnosticReport()
```
```

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

### Generate Coverage Report

Source: https://docs.sdv.dev/sdmetrics/resources/contributions/development

Generate a full code coverage report using the provided Makefile. This helps identify areas of your code that are not adequately tested.

```bash
make coverage
```

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

### Save and Load Metadata

Source: https://docs.sdv.dev/sdmetrics/getting-started/metadata/multi-table-metadata

Use the json module to persist metadata dictionaries to files and retrieve them for use in SDMetrics.

```python
import json

with open('my_metadata_file.json', 'w') as f:
    json.dump(my_metadata_dict, f)
```

```python
import json 

with open('my_metadata_file.json') as f:
    my_metadata_dict = json.load(f)

# use my_metadata_dict in the SDMetrics library
```

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

### Get Detailed Property Results

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/multi-table-api

Retrieve detailed results for a specific property, optionally filtered by table name. Returns a pandas DataFrame.

```python
report.get_details(
    property_name='Data Validity',
    table_name='users'
)
```

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

### Get Visualization for Property

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/multi-table-api

Use this method to visualize details about a property like 'Column Shapes', 'Column Pair Trends', 'Cardinality', or 'Intertable Trends' for a specified table. Returns a plotly.Figure object.

```python
fig = report.get_visualization(
    property_name='Column Shapes',
    table_name='users')
fig.show()
```

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

### QualityReport.get_visualization

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/single-table-api

Visualizes the details of a specific property.

```APIDOC
## get_visualization(property_name)

### Description
Visualizes the details about a property.

### Parameters
- **property_name** (string) - Required - The name of the property (e.g., 'Column Shapes' or 'Column Pair Trends').

### Response
- **figure** (plotly.graph_objects.Figure) - An interactive plotly figure object.
```

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

### Compute Disclosure Protection Estimate Breakdown

Source: https://docs.sdv.dev/sdmetrics/data-metrics/privacy/disclosureprotectionestimate

Use this method to get a detailed breakdown of the disclosure protection score, including CAP and baseline estimates. Specify continuous columns if they require discretization.

```python
from sdmetrics.single_table import DisclosureProtectionEstimate

score = DisclosureProtectionEstimate.compute_breakdown(
    real_data=real_table,
    synthetic_data=synthetic_table,
    known_column_names=['age', 'gender'],
    sensitive_column_names=['political_affiliation'],
    continuous_column_names=['age'],
    num_rows_subsample=2500,
    num_iterations=100,
    verbose=True
)
```

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

### Sequential Metadata Structure

Source: https://docs.sdv.dev/sdmetrics/getting-started/metadata/sequential-metadata

The minimal required structure for defining sequential metadata in SDMetrics.

```python
{
    "sequence_key": "Patient ID",
    "sequence_index": "Time",
    "columns": { <column information> }
}
```

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

### Define Datetime Column

Source: https://docs.sdv.dev/sdmetrics/getting-started/metadata/single-table-metadata

Metadata configuration for a datetime column, requiring a format string.

```python
"renew_date": {
    "sdtype": "datetime",
    "format": "%Y-%m-%d"
}
```

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

### Compute DCRBaselineProtection Breakdown

Source: https://docs.sdv.dev/sdmetrics/data-metrics/privacy/dcrbaselineprotection

Use this method to compute the DCRBaselineProtection metric. Provide real data, synthetic data, and metadata. Optional parameters include subsampling rows and the number of iterations for computation.

```python
from sdmetrics.single_table import DCRBaselineProtection

score = DCRBaselineProtection.compute_breakdown(
    real_data=real_table,
    synthetic_data=synthetic_table,
    metadata=metadata_dictionary
)
```

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

### QualityReport.get_details

Source: https://docs.sdv.dev/sdmetrics/data-metrics/quality/quality-report/single-table-api

Retrieves detailed metrics for a specific property.

```APIDOC
## get_details(property_name)

### Description
Gets more details about a particular property.

### Parameters
- **property_name** (string) - Required - The name of the property (e.g., 'Column Shapes' or 'Column Pair Trends').

### Response
- **details** (pandas.DataFrame) - A DataFrame containing detailed metrics for the specified property.
```

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

### Multi Table Metadata Dictionary

Source: https://docs.sdv.dev/sdmetrics/getting-started/metadata/multi-table-metadata

This is the metadata dictionary for the illustrated table, defining two tables 'users' and 'sessions' with their respective columns, primary keys, and relationships.

```python
{
    "tables": {
        "users": {
            "primary_key": "user_id",
            "columns": {
                "user_id": {
                    "sdtype": "id",
                    "regex_format": "U_[0-9]{3}"
                },
                "age": {
                    "sdtype": "numerical"
                },
                "address": {
                    "sdtype": "address",
                    "pii": True
                }
            }
        },
        "sessions": {
            "primary_key": "session_id",
            "columns": {
                "session_id": {
                    "sdtype": "id"
                },
                "user": {
                    "sdtype": "id",
                    "regex_format": "U_[0-9]{3}"
                },
                "date": {
                    "sdtype": "datetime",
                    "datetime_format": "%Y-%m-%d"
                },
                "browser": {
                    "sdtype": "categorical"
                },
                "bounced": {
                    "sdtype": "boolean"
                }
            }
        }
    },
    "relationships": [{
        "parent_table_name": "users",
        "parent_primary_key": "user_id",
        "child_table_name": "sessions",
        "child_foreign_key": "user_id"
    ]}
}
```

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

### Create a New Feature Branch

Source: https://docs.sdv.dev/sdmetrics/resources/contributions/development

Create a new Git branch for developing your metric. The branch name should follow the convention 'issue-<x>' where <x> is the GitHub issue number.

```bash
git checkout -b issue-123-adding-your-new-metric
```

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

### DiagnosticReport.get_details

Source: https://docs.sdv.dev/sdmetrics/data-metrics/diagnostic/diagnostic-report/multi-table-api

Retrieves detailed information about a specific property.

```APIDOC
## DiagnosticReport.get_details

### Description
Returns detailed results for a specific property, optionally filtered by table name.

### Parameters
#### Query Parameters
- **property_name** (string) - Required - The name of the property (e.g., 'Data Validity', 'Data Structure', 'Relationship Validity').
- **table_name** (string) - Optional - The name of the table to filter results by.

### Response
- **details** (pandas.DataFrame) - A DataFrame containing detailed metrics for the specified property.
```

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

### Compute BNLikelihood metric

Source: https://docs.sdv.dev/sdmetrics/data-metrics/metrics-in-beta/data-likelihood/bnlikelihood

Use the compute method from the BNLikelihood class to evaluate synthetic data against real data.

```python
from sdmetrics.single_table import BNLikelihood

BNLikelihood.compute(
    real_data=real_table,
    synthetic_data=synthetic_table
)
```

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

### Define Other Data Types

Source: https://docs.sdv.dev/sdmetrics/getting-started/metadata/single-table-metadata

Metadata configuration for custom data types, such as addresses, including PII sensitivity settings.

```python
"address": {
    "sdtype": "address",
    "pii": True
}
```

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

### Load Metadata Dictionary from JSON

Source: https://docs.sdv.dev/sdmetrics/getting-started/metadata/single-table-metadata

Use this snippet to load a Python dictionary from a JSON file. The 'json' library must be imported. The loaded dictionary can then be used with the SDMetrics library.

```python
import json 

with open('my_metadata_file.json') as f:
    my_metadata_dict = json.load(f)

# use my_metadata_dict in the SDMetrics library
```

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

### Metadata Structure: Individual Table Definition

Source: https://docs.sdv.dev/sdmetrics/getting-started/metadata/multi-table-metadata

Each table within the 'tables' dictionary is defined by its name as a key, and its value is a dictionary specifying the primary key and a detailed 'columns' description.

```python
{
    "tables": {
        "users": {
            "primary_key": "user_id",
            "columns": { <column information> }
         },
        "sessions": {
            "primary_key": "session_id",
            "columns": { <column information> }
        }
    },
    ...
}
```

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

### Lint Code Style

Source: https://docs.sdv.dev/sdmetrics/resources/contributions/development

Check the style of your code using the linting tool provided by the Makefile. This ensures code consistency and adherence to project standards.

```bash
make lint
```
Tokens:
Raw