langchain-rust 7.0.0

LangChain for Rust, the easiest way to write LLM-based programs in Rust
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54

use async_trait::async_trait;
use text_splitter::ChunkConfig;
use tiktoken_rs::tokenizer::Tokenizer;

use super::{SplitterOptions, TextSplitter, TextSplitterError};

#[derive(Debug, Clone)]
pub struct TokenSplitter {
    splitter_options: SplitterOptions,
}

impl Default for TokenSplitter {
    fn default() -> Self {
        TokenSplitter::new(SplitterOptions::default())
    }
}

impl TokenSplitter {
    pub fn new(options: SplitterOptions) -> TokenSplitter {
        TokenSplitter {
            splitter_options: options,
        }
    }

    #[deprecated = "Use `SplitterOptions::get_tokenizer_from_str` instead"]
    pub fn get_tokenizer_from_str(&self, s: &str) -> Option<Tokenizer> {
        match s.to_lowercase().as_str() {
            "cl100k_base" => Some(Tokenizer::Cl100kBase),
            "p50k_base" => Some(Tokenizer::P50kBase),
            "r50k_base" => Some(Tokenizer::R50kBase),
            "p50k_edit" => Some(Tokenizer::P50kEdit),
            "gpt2" => Some(Tokenizer::Gpt2),
            _ => None,
        }
    }
}

#[async_trait]
impl TextSplitter for TokenSplitter {
    async fn split_text(&self, text: &str) -> Result<Vec<String>, TextSplitterError> {
        let options = self.splitter_options.clone();
        let text = text.to_string();
        // Tokenizing and splitting a large document is CPU-bound; keep it off the async
        // worker thread.
        crate::blocking::run_blocking(move || {
            let chunk_config = ChunkConfig::try_from(&options)?;
            Ok(text_splitter::TextSplitter::new(chunk_config)
                .chunks(&text)
                .map(|x| x.to_string())
                .collect())
        })
        .await
    }
}