Crate replicate_rust

An Unofficial Rust client for Replicate. Provides a type-safe interface by deserializing API responses into Rust structs.

Getting Started

Add replicate_rust to Cargo.toml:

[dependencies]
replicate-rust = "0.0.5"

Grab your token from replicate.com/account and set it as an environment variable:

export REPLICATE_API_TOKEN=<your token>

Here’s an example using replicate_rust to run a model:

use replicate_rust::{config::Config, Replicate, errors::ReplicateError};

fn main() -> Result<(), ReplicateError> {
   let config = Config::default();
   // Instead of using the default config ( which reads API token from env variable), you can also set the token directly:
   // let config = Config {
   //     auth: String::from("REPLICATE_API_TOKEN"),
   //     ..Default::default()
   // };

   let replicate = Replicate::new(config);

   // Construct the inputs.
   let mut inputs = std::collections::HashMap::new();
   inputs.insert("prompt", "a  19th century portrait of a wombat gentleman");

   let version = "stability-ai/stable-diffusion:27b93a2413e7f36cd83da926f3656280b2931564ff050bf9575f1fdf9bcd7478";

   // Run the model.
   let result = replicate.run(version, inputs)?;

   // Print the result.
   println!("{:?}", result.output);
   // Some(Array [String("https://pbxt.replicate.delivery/QLDGe2rXuIQ9ByMViQEXrYCkKfDi9I3YWAzPwWsDZWMXeN7iA/out-0.png")])```

   Ok(())
}

Usage

See the reference docs for detailed API documentation.

Examples

• Run a model in the background:

  // Construct the inputs.
  let mut inputs = std::collections::HashMap::new();
  inputs.insert("prompt", "a 19th century portrait of a wombat gentleman");
  
  let version = "stability-ai/stable-diffusion:27b93a2413e7f36cd83da926f3656280b2931564ff050bf9575f1fdf9bcd7478";
  
  // Run the model.
  let mut prediction = replicate.predictions.create(version, inputs)?;
  
  println!("{:?}", prediction.status);
  // 'starting'
  
  prediction.reload()?;
  println!("{:?}", prediction.status);
  // 'processing'
  
  println!("{:?}", prediction.logs);
  // Some("Using seed: 3599
  // 0%|          | 0/50 [00:00<?, ?it/s]
  // 4%|▍         | 2/50 [00:00<00:04, 10.00it/s]
  // 8%|▊         | 4/50 [00:00<00:03, 11.56it/s]
  
  
  let prediction = prediction.wait()?;
  
  println!("{:?}", prediction.status);
  // 'succeeded'
  
  println!("{:?}", prediction.output);

• Cancel a prediction:

  // Construct the inputs.
  let mut inputs = std::collections::HashMap::new();
  inputs.insert("prompt", "a 19th century portrait of a wombat gentleman");
  
  let version = "stability-ai/stable-diffusion:27b93a2413e7f36cd83da926f3656280b2931564ff050bf9575f1fdf9bcd7478";
  
  // Run the model.
  let mut prediction = replicate.predictions.create(version, inputs)?;
  
  println!("{:?}", prediction.status);
  // 'starting'
  
  prediction.cancel()?;
  
  prediction.reload()?;
  
  println!("{:?}", prediction.status);
  // 'cancelled'

• List predictions:

  let predictions = replicate.predictions.list()?;
  println!("{:?}", predictions);
  // ListPredictions { ... }

• Get model Information:

  let model = replicate.models.get("replicate", "hello-world")?;
  println!("{:?}", model);
  // GetModel { ... }

• Get Versions List:

  let versions = replicate.models.versions.list("replicate", "hello-world")?;
  println!("{:?}", versions);
  // ListModelVersions { ... }

• Get Model Version Information:

  let model = replicate.models.versions.get("kvfrans",
  "clipdraw",
  "5797a99edc939ea0e9242d5e8c9cb3bc7d125b1eac21bda852e5cb79ede2cd9b",)?;
  println!("{:?}", model);
  // GetModelVersion { ... }

• Get Collection Information:

  let collection = replicate.collections.get("audio-generation")?;
  println!("{:?}", collection);
  // GetCollectionModels { ... }//!   ```

• Get Collection Lists:

  let collections = replicate.collections.list()?;
  println!("{:?}", collections);
  // ListCollectionModels { ... }

Modules

api_definitions

This module contains the definition of the API responses by the Replicate API. The responses the documented in the HTTP API reference.

collection

Used to interact with the Collection Endpoints.

config

The config module contains the Config struct, which is used to initialize configuration for the API. Currently contains the API token, the user agent and the base url.

errors

Custom errors for the crate.

model

Used to interact with the Model Endpoints.

prediction

Used to interact with the Prediction Endpoints.

prediction_client

Helper struct for the prediction struct

retry

Helper struct for the prediction struct. Used to retry pooling the api for latest prediction status until it is completed.

training

Used to interact with the Training Endpoints.

version

Used to interact with the Model Versions Endpoints.

Structs

Replicate

Rust Client for interacting with the Replicate API. Currently supports the following endpoints: