Crate tangram

The Tangram crate makes it easy to make predictions with your Tangram machine learning model from Rust.

Usage

[dependencies]
tangram = "*"

let model: tangram::Model = tangram::Model::from_path("heart_disease.tangram", None).unwrap();

let input = tangram::predict_input! {
  "age": 63.0,
  "gender": "male",
  // ...
};

let output = model.predict_one(input, None);

For more information, read the docs.

Macros

predict_input

Structs

BagOfWordsCosineSimilarityFeatureContribution

This describes the contribution of a feature from a bag of words cosine similarity feature group.

BagOfWordsFeatureContribution

This describes the contribution of a feature from a bag of words feature group.

BinaryClassificationPredictOutput

This is the output of calling Model::predict on a Model whose task is binary classification.

FeatureContributions

This is a description of the feature contributions for the prediction if the task is regression or binary classification, or for a single class if the task is multiclass classification.

IdentityFeatureContribution

This describes the contribution of a feature from an identity feature group.

LoadModelOptions

These are the options passed when loading a Model.

LogPredictionArgs

This is the type of the argument to Model::log_prediction and Model::enqueue_log_prediction which specifies the details of the prediction to log.

LogTrueValueArgs

This is the type of the argument to Model::log_true_value and Model::enqueue_log_true_value which specifies the details of the true value to log.

Model

Use this struct to load a model, make predictions, and log events to the app.

MulticlassClassificationPredictOutput

This is the output of calling Model::predict on a Model whose task is multiclass classification.

NormalizedFeatureContribution

This describes the contribution of a feature from a normalized feature group.

OneHotEncodedFeatureContribution

PredictInput

This is the input type of Model::predict. A predict input is a map whose keys are the same as the column names in the CSV the model was trained with, and whose values match the type for each column.

PredictOptions

These are the options passed to Model::predict.

RegressionPredictOutput

This is the output of calling Model::predict on a Model whose task is regression.

WordEmbeddingFeatureContribution

This describes the contribution of a feature from a word vector feature group.

Enums

FeatureContributionEntry

This identifies the type of a feature contribution.

NGram

This is a sequence of n tokens. Tangram currently supports unigrams and bigrams.

NumberOrString

PredictInputValue

PredictOutput

This is the output of Model::predict.

Traits

ClassificationOutputValue

Derive Macros

ClassificationOutputValue

PredictInput

PredictInputValue