Skip to content

Fixing the tokenization routes token (offsets are in bytes, not in #576

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Apr 9, 2025
Merged

Fixing the tokenization routes token (offsets are in bytes, not in #576

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
7ed87cb
Fixing the tokenization routes token (offsets are in bytes, not in
Narsil Apr 9, 2025
63d9fde
Fixing the tokenization in both routes.
Narsil Apr 9, 2025
5a1149d
Typo.
Narsil Apr 9, 2025
46dd4ec
Casting.
Narsil Apr 9, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
162 changes: 162 additions & 0 deletions core/src/tokenization.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,6 +16,16 @@ pub struct Tokenization {
sender: async_channel::Sender<TokenizerRequest>,
}

#[derive(Debug)]
#[cfg_attr(test, derive(PartialEq))]
pub struct SimpleToken {
pub id: u32,
pub text: String,
pub special: bool,
pub start: Option<usize>,
pub stop: Option<usize>,
}

impl Tokenization {
pub fn new(
workers: usize,
Expand Down Expand Up @@ -485,3 +495,155 @@ enum TokenizerRequest {
Span,
),
}

pub fn into_tokens(encoding: tokenizers::Encoding, input: &str) -> Vec<SimpleToken> {
encoding
.get_ids()
.iter()
.zip(encoding.get_offsets())
.zip(encoding.get_special_tokens_mask())
.zip(encoding.get_tokens())
.map(|(((&id, &(start, stop)), special), token)| {
let special = *special == 1;
match special {
true => SimpleToken {
id,
text: token.clone(),
special,
start: None,
stop: None,
},
false => {
let text: Vec<u8> = input.bytes().skip(start).take(stop - start).collect();
let text: String = String::from_utf8_lossy(&text).to_string();
SimpleToken {
id,
text,
special,
start: Some(start),
stop: Some(stop),
}
}
}
})
.collect()
}

#[cfg(test)]
mod tests {
use super::*;
use hf_hub::api::sync::ApiBuilder;

#[test]
fn tokenizer() {
let api = ApiBuilder::from_env().build().unwrap();
let filename = api
.model("BAAI/bge-m3".to_string())
.get("tokenizer.json")
.unwrap();
let string = "这是一个文本向量化的测试句子";
let tokenizer = Tokenizer::from_file(filename).unwrap();

let encoded = tokenizer.encode(string, true).unwrap();
assert_eq!(
encoded.get_offsets(),
vec![
(0, 0),
(0, 3),
(0, 12),
(12, 18),
(18, 21),
(21, 24),
(24, 30),
(30, 36),
(36, 39),
(39, 42),
(0, 0)
]
);

let tokens = into_tokens(encoded, &string);
assert_eq!(
tokens,
vec![
SimpleToken {
id: 0,
text: "<s>".to_string(),
special: true,
start: None,
stop: None
},
SimpleToken {
id: 6,
text: "这".to_string(),
special: false,
start: Some(0),
stop: Some(3)
},
SimpleToken {
id: 100013,
text: "这是一个".to_string(),
special: false,
start: Some(0),
stop: Some(12)
},
SimpleToken {
id: 189061,
text: "文本".to_string(),
special: false,
start: Some(12),
stop: Some(18)
},
SimpleToken {
id: 2110,
text: "向".to_string(),
special: false,
start: Some(18),
stop: Some(21)
},
SimpleToken {
id: 3272,
text: "量".to_string(),
special: false,
start: Some(21),
stop: Some(24)
},
SimpleToken {
id: 41904,
text: "化的".to_string(),
special: false,
start: Some(24),
stop: Some(30)
},
SimpleToken {
id: 49125,
text: "测试".to_string(),
special: false,
start: Some(30),
stop: Some(36)
},
SimpleToken {
id: 27683,
text: "句".to_string(),
special: false,
start: Some(36),
stop: Some(39)
},
SimpleToken {
id: 1344,
text: "子".to_string(),
special: false,
start: Some(39),
stop: Some(42)
},
SimpleToken {
id: 2,
text: "</s>".to_string(),
special: true,
start: None,
stop: None
}
]
);
}
}
46 changes: 19 additions & 27 deletions router/src/grpc/server.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,9 @@ use std::future::Future;
use std::net::SocketAddr;
use std::time::{Duration, Instant};
use text_embeddings_core::infer::Infer;
use text_embeddings_core::tokenization::EncodingInput;
use text_embeddings_core::tokenization::{
into_tokens, EncodingInput, SimpleToken as CoreSimpleToken,
};
use tokio::sync::{mpsc, oneshot, OwnedSemaphorePermit};
use tokio_stream::wrappers::UnboundedReceiverStream;
use tokio_stream::StreamExt;
Expand Down Expand Up @@ -340,32 +342,22 @@ impl TextEmbeddingsService {
.map_err(ErrorResponse::from)?;
let inputs = encoded_inputs.unwrap_or(inputs);

let tokens: Vec<SimpleToken> = encoding
.get_ids()
.iter()
.zip(encoding.get_offsets())
.zip(encoding.get_special_tokens_mask())
.zip(encoding.get_tokens())
.map(|(((&id, &(start, stop)), special), token)| {
let special = *special == 1;
match special {
true => SimpleToken {
id,
text: token.clone(),
special,
start: None,
stop: None,
},
false => {
let text: String = inputs.chars().skip(start).take(stop - start).collect();
SimpleToken {
id,
text,
special,
start: Some(start as u32),
stop: Some(stop as u32),
}
}
let tokens: Vec<SimpleToken> = into_tokens(encoding, &inputs)
.into_iter()
.map(|t| {
let CoreSimpleToken {
id,
text,
special,
start,
stop,
} = t;
SimpleToken {
id,
text,
special,
start: start.map(|s| s as u32),
stop: stop.map(|s| s as u32),
}
})
.collect();
Expand Down
43 changes: 17 additions & 26 deletions router/src/http/server.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -34,6 +34,7 @@ use text_embeddings_backend::BackendError;
use text_embeddings_core::infer::{
AllEmbeddingsInferResponse, Infer, InferMetadata, PooledEmbeddingsInferResponse,
};
use text_embeddings_core::tokenization::{into_tokens, SimpleToken as CoreSimpleToken};
use text_embeddings_core::TextEmbeddingsError;
use tokio::sync::OwnedSemaphorePermit;
use tower_http::cors::{AllowOrigin, CorsLayer};
Expand Down Expand Up @@ -1295,32 +1296,22 @@ async fn tokenize(
.map_err(ErrorResponse::from)?;
let input = encoded_input.unwrap_or(input);

let tokens: Vec<SimpleToken> = encoding
.get_ids()
.iter()
.zip(encoding.get_offsets())
.zip(encoding.get_special_tokens_mask())
.zip(encoding.get_tokens())
.map(|(((&id, &(start, stop)), special), token)| {
let special = *special == 1;
match special {
true => SimpleToken {
id,
text: token.clone(),
special,
start: None,
stop: None,
},
false => {
let text: String = input.chars().skip(start).take(stop - start).collect();
SimpleToken {
id,
text,
special,
start: Some(start),
stop: Some(stop),
}
}
let tokens: Vec<SimpleToken> = into_tokens(encoding, &input)
.into_iter()
.map(|t| {
let CoreSimpleToken {
id,
text,
special,
start,
stop,
} = t;
SimpleToken {
id,
text,
special,
start,
stop,
}
})
.collect();
Expand Down
Loading