# Julia API

## Catch22 Module

The `Catch22` module provides functionality for extracting the *catch22* feature set from time series data.

### Types

#### FeatureSet

```julia
FeatureSet(funcs::Vector{<:Function}, names::Vector{Symbol}, keywords::Vector{Vector{String}}, descriptions::Vector{String})
```

A collection of feature extraction functions, their names, keywords, and descriptions.

#### SuperFeatureSet

```julia
SuperFeatureSet(funcs::Vector{<:Function}, names::Vector{Symbol}, descriptions::Vector{String}, keywords::Vector{Vector{String}}, transform::Function)
```

A hierarchichal`FeatureSet` in which one or more evaluated features are non-independent, relying on some (potentially) common preprocessing transformations; e.g. the partial autocorrelation at successive lags. Feature sets of this type detect any common preprocessing transformations, as to only perform each calculation once. `Catch22.jl`uses this to compute the zscore transformation, required by all *catch22* features.

Further details on the custom types used by the `Catch22.jl` package can be found at [`TimeseriesFeatures.jl`](https://github.com/brendanjohnharris/TimeseriesFeatures.jl).

### Functions

#### catch22

```julia
catch22(𝐱::Vector)
catch22(X::Array)
catch22[featurename::Symbol](X::Array)
```

Evaluate all *catch22* features for a time series vector `𝐱` or the columns of an array `X`. Can be indexed by feature names (as symbols) to return a subset of the available features.

#### catch24

```
catch24(𝐱::Vector)
catch24(X::Array)
```

Evaluate all *catch22* features along with the mean (`DN_Mean`) and standard deviation (`DN_Spread_Std`) for a time series vector `𝐱` or the columns of an array `X`.

#### c22

```julia
c22(𝐱::Vector)
c22(X::Array)
```

Evaluate all *catch22* features with **shortened** names for a time series vector `𝐱` or the columns of an array `X`.

#### c24

```
c24(𝐱::Vector)
c24(X::Array)
```

Evaluate all *catch22* features with **shortened** names, along with the mean (`mean`) and standard deviation (`std`) for a time series vector `𝐱` or the columns of an array `X`.

#### Individual Feature Methods

The `Catch22` module also provides direct access to the individual feature extraction methods. These methods can be called directly using `Catch22.{name}`, where `{name}` is the name of the feature method (as given by the **long** name in the [table of features](https://time-series-features.gitbook.io/catch22/information-about-catch22/feature-descriptions/feature-overview-table)).&#x20;

| Method                | Parameters    | Return Type | Description                                                                |
| --------------------- | ------------- | ----------- | -------------------------------------------------------------------------- |
| `DN_Mean`             | `𝐱` (Vector) | Float64     | Computes the mean of the input time series.                                |
| `DN_Spread_Std`       | `𝐱` (Vector) | Float64     | Computes the standard deviation of the input time series.                  |
| `DN_HistogramMode_5`  | `𝐱` (Vector) | Float64     | Computes the mode of the input time series using a histogram with 5 bins.  |
| `DN_HistogramMode_10` | `𝐱` (Vector) | Float64     | Computes the mode of the input time series using a histogram with 10 bins. |
| ...                   | ...           | ...         | ...                                                                        |

***

## Example Usage

```julia
using Catch22

# Load test data
𝐱 = Catch22.testdata[:test]

# Evaluate all catch22 features
features = catch22(𝐱)

# Evaluate specific feature
mode_5 = DN_HistogramMode_5(𝐱)

# Evaluate catch24 features (with mean and std)
features_with_stats = catch24(𝐱)

# Evaluate features with shortened names
short_features = c22(𝐱)

```

***