support spec char as delimiter

Signed-off-by: junxiang Mu <1948535941@qq.com>
This commit is contained in:
junxiang Mu
2025-05-06 11:00:18 +08:00
parent fc47ca9dd2
commit 0ac1095c70
11 changed files with 139 additions and 39 deletions
+110 -17
View File
@@ -1,6 +1,7 @@
use async_trait::async_trait;
use bytes::Bytes;
use chrono::Utc;
use common::DEFAULT_DELIMITER;
use ecstore::io::READ_BUFFER_SIZE;
use ecstore::new_object_layer_fn;
use ecstore::store::ECStore;
@@ -24,29 +25,54 @@ use object_store::PutOptions;
use object_store::PutPayload;
use object_store::PutResult;
use object_store::{Error as o_Error, Result};
use pin_project_lite::pin_project;
use s3s::dto::SelectObjectContentInput;
use s3s::s3_error;
use s3s::S3Result;
use std::ops::Range;
use std::pin::Pin;
use std::sync::Arc;
use std::task::ready;
use std::task::Poll;
use tokio::io::AsyncRead;
use tokio_util::io::ReaderStream;
use tracing::info;
use transform_stream::AsyncTryStream;
#[derive(Debug)]
pub struct EcObjectStore {
input: SelectObjectContentInput,
input: Arc<SelectObjectContentInput>,
need_convert: bool,
delimiter: String,
store: Arc<ECStore>,
}
impl EcObjectStore {
pub fn new(input: SelectObjectContentInput) -> S3Result<Self> {
pub fn new(input: Arc<SelectObjectContentInput>) -> S3Result<Self> {
let Some(store) = new_object_layer_fn() else {
return Err(s3_error!(InternalError, "ec store not inited"));
};
Ok(Self { input, store })
let (need_convert, delimiter) = if let Some(csv) = input.request.input_serialization.csv.as_ref() {
if let Some(delimiter) = csv.field_delimiter.as_ref() {
if delimiter.len() > 1 {
(true, delimiter.to_owned())
} else {
(false, String::new())
}
} else {
(false, String::new())
}
} else {
(false, String::new())
};
Ok(Self {
input,
need_convert,
delimiter,
store,
})
}
}
@@ -79,16 +105,6 @@ impl ObjectStore for EcObjectStore {
source: "can not get object info".into(),
})?;
// let stream = stream::unfold(reader.stream, |mut blob| async move {
// match blob.next().await {
// Some(Ok(chunk)) => {
// let bytes = chunk;
// Some((Ok(bytes), blob))
// }
// _ => None,
// }
// })
// .boxed();
let meta = ObjectMeta {
location: location.clone(),
last_modified: Utc::now(),
@@ -98,10 +114,21 @@ impl ObjectStore for EcObjectStore {
};
let attributes = Attributes::default();
Ok(GetResult {
payload: object_store::GetResultPayload::Stream(
let payload = if self.need_convert {
object_store::GetResultPayload::Stream(
bytes_stream(
ReaderStream::with_capacity(ConvertStream::new(reader.stream, self.delimiter.clone()), READ_BUFFER_SIZE),
reader.object_info.size,
)
.boxed(),
)
} else {
object_store::GetResultPayload::Stream(
bytes_stream(ReaderStream::with_capacity(reader.stream, READ_BUFFER_SIZE), reader.object_info.size).boxed(),
),
)
};
Ok(GetResult {
payload,
meta,
range: 0..reader.object_info.size,
attributes,
@@ -154,6 +181,54 @@ impl ObjectStore for EcObjectStore {
}
}
pin_project! {
struct ConvertStream<R> {
inner: R,
delimiter: Vec<u8>,
}
}
impl<R> ConvertStream<R> {
fn new(inner: R, delimiter: String) -> Self {
ConvertStream {
inner,
delimiter: delimiter.as_bytes().to_vec(),
}
}
}
impl<R: AsyncRead + Unpin> AsyncRead for ConvertStream<R> {
#[tracing::instrument(level = "debug", skip_all)]
fn poll_read(
self: Pin<&mut Self>,
cx: &mut std::task::Context<'_>,
buf: &mut tokio::io::ReadBuf<'_>,
) -> std::task::Poll<std::io::Result<()>> {
let me = self.project();
ready!(Pin::new(&mut *me.inner).poll_read(cx, buf))?;
let bytes = buf.filled();
let replaced = replace_symbol(me.delimiter, bytes);
buf.clear();
buf.put_slice(&replaced);
Poll::Ready(Ok(()))
}
}
fn replace_symbol(delimiter: &[u8], slice: &[u8]) -> Vec<u8> {
let mut result = Vec::with_capacity(slice.len());
let mut i = 0;
while i < slice.len() {
if slice[i..].starts_with(delimiter) {
result.push(DEFAULT_DELIMITER);
i += delimiter.len();
} else {
result.push(slice[i]);
i += 1;
}
}
result
}
pub fn bytes_stream<S>(stream: S, content_length: usize) -> impl Stream<Item = Result<Bytes>> + Send + 'static
where
S: Stream<Item = Result<Bytes, std::io::Error>> + Send + 'static,
@@ -175,3 +250,21 @@ where
Ok(())
})
}
#[cfg(test)]
mod test {
use super::replace_symbol;
#[test]
fn test_replace() {
let ss = String::from("dandan&&is&&best");
let slice = ss.as_bytes();
let delimiter = b"&&";
println!("len: {}", "".len());
let result = replace_symbol(delimiter, slice);
match String::from_utf8(result) {
Ok(s) => println!("slice: {}", s),
Err(e) => eprintln!("Error converting to string: {}", e),
}
}
}
+3 -1
View File
@@ -1,3 +1,5 @@
use std::sync::Arc;
use s3s::dto::SelectObjectContentInput;
pub mod analyzer;
@@ -16,7 +18,7 @@ pub mod session;
#[derive(Clone)]
pub struct Context {
// maybe we need transfer some info?
pub input: SelectObjectContentInput,
pub input: Arc<SelectObjectContentInput>,
}
#[derive(Clone)]
+1 -1
View File
@@ -66,7 +66,7 @@ impl SessionCtxFactory {
9,Ivy,24,Marketing,4800
10,Jack,38,Finance,7500";
let data_bytes = data.as_bytes();
// let data = r#""year"╦"gender"╦"ethnicity"╦"firstname"╦"count"╦"rank"
// let data = r#""year"╦"gender"╦"ethnicity"╦"firstname"╦"count"╦"rank"
// "2011"╦"FEMALE"╦"ASIAN AND PACIFIC ISLANDER"╦"SOPHIA"╦"119"╦"1"
// "2011"╦"FEMALE"╦"ASIAN AND PACIFIC ISLANDER"╦"CHLOE"╦"106"╦"2"
// "2011"╦"FEMALE"╦"ASIAN AND PACIFIC ISLANDER"╦"EMILY"╦"93"╦"3"