Crate seedframe

Help

Source

Expand description

§Seedframe - Core API Documentation

Seedframe is a clean, macro-driven Rust library for building LLM applications.

§Features

Declarative API through straight forward proc-macros
Modular Architecture with clearly defined components:
- Loaders: Data ingestion from various sources (files, APIs, etc.)
- Vector Stores: Embedding storage and retrieval (In-memory, Redis, etc.)
- Embedders: Text embedding providers
- LLM Clients: Unified interface for different LLM providers
- Tools: Function calling abstractions with state management and automatic documentation
- Extractors: Structured output generation from LLM responses

§Examples

The seedframe repo contains a number of examples that show how to put all the pieces together.

§Building a simple RAG

use seedframe::prelude::*;
use seedframe::providers::{completions::OpenAI, embeddings::OpenAIEmbedding};
use seedframe::vector_store::InMemoryVectorStore;

// Declare file loader that doesnt check for updates, loading files that match the glob pattern
#[loader(kind = "FileOnceLoader", path = "/tmp/data/**/*.txt")]
pub struct MyLoader;

#[vector_store(store = "InMemoryVectorStore")]
pub struct MyVectorStore;

#[embedder(provider = "OpenAIEmbedding")]
struct MyEmbedder {
    #[vector_store]
    my_vector_store: MyVectorStore,
    #[loader]
    my_loader: MyLoader,
}

#[client(provider = "OpenAI")]
struct MyClient {
    #[embedder]
    my_embedder: MyEmbedder,
}

#[tokio::main]
async fn main() {
    let mut client = MyClient::build(
        "You are a helpful assistant".to_string()
    ).await;
     
    tokio::time::sleep(Duration::from_secs(5)).await;
    let response = client.prompt("Explain quantum computing").send().await.unwrap();
}

§Tool calls and Extractors

use seedframe::{providers::completions::OpenAI, prelude::*};

#[client(provider = "OpenAI", tools("analyze"))]
struct ToolClient;

/// Perform sentiment analysis on text
/// # Arguments
/// * `text`: Input text to analyze
/// * `language`: Language of the text (ISO 639-1)
#[tool]
fn analyze(text: String, language: String) -> String {
    todo!("implementation");
}

#[derive(Extractor)]
struct PersonData {
    /// Age in years
    age: u8,
    /// Email address
    email: String
}

#[tokio::main]
async fn main() -> Result<()> {
    let mut client = ToolClient::build("You're a data analyst".to_string())
        .await
        .with_state(AppState::new())?;

    // Tool call
    client.prompt("Analyze this: 'I love Rust!' (en)")
        .append_tool_response(true) // append result of tool execution to the message history
        // more prompt modifiers if you want
        .send()
        .await?;

    // Structured extraction
    let person = client.prompt("John is 30, email john@example.com")
        .extract::<PersonData>()
        .await?;
}

You can pass state to tools by adding arguments of type State<_> to them, the only catch is that there can only be one type of State<T> attached to the client.

use seedframe::{providers::completions::OpenAI, prelude::*};

#[client(provider = "OpenAI", tools("greet"))]
struct ToolClient;

/// Greets a user
/// # Arguments
/// * `name`: name of the user
#[tool]
fn greet(name: String, State(count): State<u32>) -> String {
    for _ in range 0..count { println("Hello {name}!!")};
}

#[tokio::main]
async fn main() -> Result<()> {
    let mut client = ToolClient::build("You're a helpful assistant".to_string())
        .await
        .with_state(3u32)?
        .with_state(7u32)? // this is an error since there's already a State of type u32 attached
        .with_state("some other state".to_string())?;

    // Tool call
    client.prompt("Say hi to jack for me".to_string())
        .send()
        .await?;
}

To share mutable state you can use types with interior mutablity, eg Mutexs

/// Greets a user
/// # Arguments
/// * `name`: name of the user
#[tool]
fn greet(name: String, State(count): State<u32>) -> String {
    let mut count = count.lock().unwrap();
    println!("{name}! This is my {count}th time saying hello");
    *count += 1;
}

struct AppState {
    count: std::sync::Mutex<u32>
}

#[tokio::main]
async fn main() -> Result<()> {
    let mut client = ToolClient::build("You're a helpful assistant".to_string())
        .await
        .with_state(AppState { count: std::sync::Mutex::new(0u32) })?;

    // Tool call
    client.prompt("Say hi to jack for me".to_string())
        .send()
        .await?;

    // Tool call
    client.prompt("Say hi to jack for me".to_string())
        .send()
        .await?;
}

§Feature flags

seedframe uses a set of [feature flags] to reduce the amount of compiled and optional dependencies.

The following optional features are available:

Name	Description	Default?
`pdf`	enables file loaders to parse PDFs	No

Modules§

completion: Language model completion and conversation management
document: Document processing and representation utilities
embeddings: Text embeddings support
error: Error types for all library operations
loader: Resource loading utilities
prelude: Convenience prelude exports
providers: Builtin completion and embedding model providers
tools: Function calling and tool execution support
vector_store: Vector storage and retrieval

Crate seedframeCopy item path