Compare commits

...
Author SHA1 Message Date
Brandon Thomas 37e53c20df notify of download start 2025-03-10 21:48:12 -04:00
Brandon Thomas 1e024d449b send more notification details 2025-03-10 21:43:14 -04:00
Brandon Thomas 3176741dc3 cleanup 2025-03-10 21:29:42 -04:00
Brandon Thomas 2ca91097af notify thread 2025-03-10 21:15:52 -04:00
Brandon Thomas 12b2d36934 tempdir download and rename 2025-03-10 21:08:33 -04:00
Brandon Thomas bf380bc416 send downloaded bytes 2025-03-10 20:25:41 -04:00
Brandon Thomas de1d5cbc39 parallel downloads 2025-03-10 19:30:39 -04:00
Brandon Thomas 1482ce1ae3 gross, but should work 2025-03-10 18:29:33 -04:00
Brandon Thomas 1bd44108ea spawn task to download 2025-03-10 17:56:40 -04:00
Brandon Thomas f08c802c66 send progress notification 2025-03-10 16:38:06 -04:00
Brandon Thomas 8385b92178 download code works 2025-03-07 16:39:39 -05:00
Brandon Thomas ee9b7eb675 downloader code fixed 2025-03-07 14:16:36 -05:00
Brandon Thomas 65632d4b81 downloader code from hf 2025-03-07 14:08:10 -05:00
14 changed files with 830 additions and 33 deletions
Generated
+71 -14
View File
@@ -666,12 +666,14 @@ dependencies = [
"cudarc",
"expanduser",
"filesys",
"futures 0.3.29",
"hf-hub 0.3.2",
"image",
"libloading 0.8.6",
"log",
"memory_caching",
"once_cell",
"openssl",
"rand 0.8.5",
"reqwest 0.12.12",
"serde",
@@ -680,7 +682,11 @@ dependencies = [
"tauri",
"tauri-build",
"tauri-plugin-log",
"tempdir",
"tempfile",
"tokenizers 0.15.2",
"tokio",
"tokio-util",
]
[[package]]
@@ -3792,12 +3798,12 @@ dependencies = [
[[package]]
name = "errno"
version = "0.3.9"
version = "0.3.10"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "534c5cf6194dfab3db3242765c03bbe257cf92f22b38f6bc0c58d59108a820ba"
checksum = "33d852cb9b869c2a9b3df2f71a3074817f01e1844f839a144f5fcef059a4eb5d"
dependencies = [
"libc",
"windows-sys 0.52.0",
"windows-sys 0.59.0",
]
[[package]]
@@ -3927,9 +3933,9 @@ dependencies = [
[[package]]
name = "fastrand"
version = "2.1.0"
version = "2.3.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9fc0510504f03c51ada170672ac806f1f105a88aa97a5281117e1ddc3368e51a"
checksum = "37909eebbb50d72f9059c3b6d82c0463f2ff062c9e95845c43a6c9c0355411be"
[[package]]
name = "fdeflate"
@@ -4657,6 +4663,18 @@ dependencies = [
"wasm-bindgen",
]
[[package]]
name = "getrandom"
version = "0.3.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "43a49c392881ce6d5c3b8cb70f98717b7c07aabbdff06687b9030dbfbe2725f8"
dependencies = [
"cfg-if 1.0.0",
"libc",
"wasi 0.13.3+wasi-0.2.2",
"windows-targets 0.52.6",
]
[[package]]
name = "getset"
version = "0.1.4"
@@ -6011,7 +6029,7 @@ dependencies = [
"chumsky",
"email-encoding",
"email_address",
"fastrand 2.1.0",
"fastrand 2.3.0",
"futures-util",
"hostname 0.3.1",
"httpdate",
@@ -6229,6 +6247,12 @@ version = "0.4.14"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "78b3ae25bc7c8c38cec158d1f2757ee79e9b3740fbc7ccf0e59e4b08d793fa89"
[[package]]
name = "linux-raw-sys"
version = "0.9.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "6db9c683daf087dc577b7506e9695b3d556a9f3849903fa28186283afd6809e9"
[[package]]
name = "local-channel"
version = "0.1.5"
@@ -9664,10 +9688,23 @@ dependencies = [
"bitflags 2.5.0",
"errno",
"libc",
"linux-raw-sys",
"linux-raw-sys 0.4.14",
"windows-sys 0.52.0",
]
[[package]]
name = "rustix"
version = "1.0.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f7178faa4b75a30e269c71e61c353ce2748cf3d76f0c44c393f4e60abf49b825"
dependencies = [
"bitflags 2.5.0",
"errno",
"libc",
"linux-raw-sys 0.9.2",
"windows-sys 0.59.0",
]
[[package]]
name = "rustls"
version = "0.21.12"
@@ -11899,14 +11936,16 @@ dependencies = [
[[package]]
name = "tempfile"
version = "3.10.1"
version = "3.18.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "85b77fafb263dd9d05cbeac119526425676db3784113aa9295c88498cbf8bff1"
checksum = "2c317e0a526ee6120d8dabad239c8dadca62b24b6f168914bbbc8e2fb1f0e567"
dependencies = [
"cfg-if 1.0.0",
"fastrand 2.1.0",
"rustix",
"windows-sys 0.52.0",
"fastrand 2.3.0",
"getrandom 0.3.1",
"once_cell",
"rustix 1.0.2",
"windows-sys 0.59.0",
]
[[package]]
@@ -13279,6 +13318,15 @@ version = "0.11.0+wasi-snapshot-preview1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9c8d87e72b64a3b4db28d11ce29237c246188f4f51057d65a7eab63b7987e423"
[[package]]
name = "wasi"
version = "0.13.3+wasi-0.2.2"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "26816d2e1a4a36a2940b96c5296ce403917633dff8f3440e9b236ed6f6bacad2"
dependencies = [
"wit-bindgen-rt",
]
[[package]]
name = "wasite"
version = "0.1.0"
@@ -14048,6 +14096,15 @@ dependencies = [
"windows-sys 0.48.0",
]
[[package]]
name = "wit-bindgen-rt"
version = "0.33.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "3268f3d866458b787f390cf61f4bbb563b922d091359f9608842999eaee3943c"
dependencies = [
"bitflags 2.5.0",
]
[[package]]
name = "wry"
version = "0.48.1"
@@ -14138,8 +14195,8 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "e105d177a3871454f754b33bb0ee637ecaaac997446375fd3e5d43a2ed00c909"
dependencies = [
"libc",
"linux-raw-sys",
"rustix",
"linux-raw-sys 0.4.14",
"rustix 0.38.34",
]
[[package]]
+11 -5
View File
@@ -28,24 +28,30 @@ cudarc = "0.13.7"
# External
anyhow = "1.0"
bytes = "1.10"
base64 = "0.22.1"
bytes = "1.10"
candle-core = { path = "../../vendor/candle/candle-core", features = ["cuda"] }
candle-nn = { path = "../../vendor/candle/candle-nn", features = ["cuda"] }
candle-transformers = { path = "../../vendor/candle/candle-transformers", features = ["cuda"] }
expanduser = "1.2.2"
futures = "0.3"
hf-hub = "0.3.2"
image = "0.25.5"
serde_yaml = "0.9"
log = "0.4"
once_cell = "1.20.3"
expanduser = "1.2.2"
rand = "0.8"
reqwest = "0.12.12"
openssl = { version = "0.10", features = ["vendored"] }
rand = "0.8.5"
reqwest = { version = "0.12.12", features = ["stream"] }
serde = { version = "1.0", features = ["derive"] }
serde_json = "1.0"
serde_yaml = "0.9"
tauri = { version = "2.2.4", features = [] }
tauri-plugin-log = "2"
tempdir = "0.3.7"
tempfile = "3.18.0"
tokenizers = "0.15.2"
tokio = { version = "1.42", features = ["rt", "rt-multi-thread", "fs"] }
tokio-util = { version = "0.7", features = ["codec"] }
[features]
default = ["cuda"]
@@ -6,6 +6,8 @@ use serde::Serialize;
pub enum NotificationModelType {
Unet,
Vae,
Json,
ClipEncoder,
}
#[derive(Clone, Serialize)]
@@ -15,6 +17,12 @@ pub enum NotificationEvent<'a> {
model_name: &'a str,
model_type: NotificationModelType,
},
ModelDownloadProgress {
model_name: &'a str,
model_type: NotificationModelType,
currently_downloaded_bytes: usize,
total_file_bytes: usize,
},
ModelDownloadComplete {
model_name: &'a str,
model_type: NotificationModelType,
+1
View File
@@ -3,6 +3,7 @@ pub mod events;
pub mod ml;
pub mod state;
pub mod threads;
pub mod transfer;
use crate::endpoints::download_models::download_models;
use crate::endpoints::image_endpoint::infer_image;
@@ -5,11 +5,11 @@ use crate::state::app_dir::AppDataRoot;
pub async fn download_all_models(app_data_root: &AppDataRoot) -> anyhow::Result<()> {
let weights_dir = app_data_root.weights_dir();
download_model_registry_file(ModelType::ClipJson, weights_dir).await?;
download_model_registry_file(ModelType::SdxlTurboUnet, weights_dir).await?;
download_model_registry_file(ModelType::SdxlTurboVae, weights_dir).await?;
download_model_registry_file(ModelType::SdxlTurboClipEncoder, weights_dir).await?;
download_model_registry_file(ModelType::SdxlTurboClipEncoder2, weights_dir).await?;
//download_model_registry_file(ModelType::ClipJson, weights_dir).await?;
//download_model_registry_file(ModelType::SdxlTurboUnet, weights_dir).await?;
//download_model_registry_file(ModelType::SdxlTurboVae, weights_dir).await?;
//download_model_registry_file(ModelType::SdxlTurboClipEncoder, weights_dir).await?;
//download_model_registry_file(ModelType::SdxlTurboClipEncoder2, weights_dir).await?;
Ok(())
}
+22 -1
View File
@@ -1,5 +1,6 @@
use std::path::PathBuf;
use crate::events::notification_event::NotificationModelType;
use crate::state::app_dir::AppWeightsDir;
use std::path::PathBuf;
#[derive(Copy, Clone, Debug)]
pub enum ModelType {
@@ -11,6 +12,26 @@ pub enum ModelType {
}
impl ModelType {
pub fn get_name(&self) -> &'static str {
match self {
Self::ClipJson => "CLIP (JSON)",
Self::SdxlTurboUnet => "SDXL Turbo UNET",
Self::SdxlTurboVae => "SDXL Turbo VAE",
Self::SdxlTurboClipEncoder => "Clip Encoder #1",
Self::SdxlTurboClipEncoder2 => "Clip Encoder #2",
}
}
pub fn get_notification_type(&self) -> NotificationModelType {
match self {
Self::ClipJson => NotificationModelType::Json,
Self::SdxlTurboUnet => NotificationModelType::Unet,
Self::SdxlTurboVae => NotificationModelType::Vae,
Self::SdxlTurboClipEncoder => NotificationModelType::ClipEncoder,
Self::SdxlTurboClipEncoder2 => NotificationModelType::ClipEncoder,
}
}
pub fn get_hf_id(&self) -> Option<&'static str> {
match self {
Self::ClipJson => Some("laion/CLIP-ViT-bigG-14-laion2B-39B-b160k"),
+50 -2
View File
@@ -1,17 +1,22 @@
use crate::state::os_platform::OsPlatform;
use expanduser::expanduser;
use std::path::{Path, PathBuf};
use tempdir::TempDir;
use tempfile::{Builder, NamedTempFile};
use crate::ml::model_type::ModelType;
const ASSETS_SUBDIRECTORY : &str = "assets";
const WEIGHTS_SUBDIRECTORY : &str = "weights";
const TEMPORARY_SUBDIRECTORY : &str = "temp";
/// The path to the application data directory, which includes "asset" and "weights" data.
#[derive(Clone)]
pub struct AppDataRoot {
path: PathBuf,
assets_dir: AppAssetsDir,
weights_dir: AppWeightsDir,
temp_dir: TemporaryDir,
}
#[derive(Clone)]
@@ -24,6 +29,11 @@ pub struct AppWeightsDir {
path: PathBuf,
}
#[derive(Clone)]
pub struct TemporaryDir {
path: PathBuf,
}
impl AppDataRoot {
pub fn create_existing<P: AsRef<Path>>(dir: P) -> anyhow::Result<Self> {
let mut dir = dir.as_ref().to_path_buf();
@@ -51,11 +61,13 @@ impl AppDataRoot {
let assets_dir = AppAssetsDir::create_existing(dir.join(ASSETS_SUBDIRECTORY))?;
let weights_dir = AppWeightsDir::create_existing(dir.join(WEIGHTS_SUBDIRECTORY))?;
let temp_dir = TemporaryDir::create_existing(dir.join(TEMPORARY_SUBDIRECTORY))?;
Ok(Self {
path: dir,
assets_dir,
weights_dir,
temp_dir,
})
}
@@ -66,7 +78,11 @@ impl AppDataRoot {
pub fn weights_dir(&self) -> &AppWeightsDir {
&self.weights_dir
}
pub fn temp_dir(&self) -> &TemporaryDir {
&self.temp_dir
}
pub fn path(&self) -> &Path {
&self.path
}
@@ -121,3 +137,35 @@ impl AppWeightsDir {
self.path.join(model_registry.get_filename())
}
}
impl TemporaryDir {
pub fn create_existing<P: AsRef<Path>>(dir: P) -> anyhow::Result<Self> {
let mut dir = dir.as_ref().to_path_buf();
match dir.canonicalize() {
Ok(d) => dir = d,
Err(err) => {
println!("Error canonicalizing {:?}: {}", dir, err);
}
}
if !dir.exists() {
println!("Creating directory {:?}", dir);
std::fs::create_dir(&dir)?;
}
Ok(Self {
path: dir,
})
}
pub fn with_prefix(&self, prefix: &str) -> anyhow::Result<TempDir> {
let tempdir = TempDir::new_in(&self.path, prefix)?;
Ok(tempdir)
}
pub fn new_named_temp_file(&self) -> anyhow::Result<NamedTempFile> {
let tempfile = Builder::new()
.prefix("temp_")
.suffix(".bin")
.tempfile_in(&self.path)?;
Ok(tempfile)
}
}
@@ -1,11 +1,169 @@
use tauri::{AppHandle, Emitter};
use crate::events::notification_event::{NotificationEvent, NotificationModelType};
use crate::ml::model_type::ModelType;
use crate::state::app_dir::AppDataRoot;
use crate::transfer::download::{download_async, ProgressUpdate};
use futures::stream::FuturesUnordered;
use futures::StreamExt;
use log::{debug, error, info};
use std::collections::VecDeque;
use std::path::PathBuf;
use std::sync::mpsc::Receiver;
use std::sync::{Arc, Mutex};
use tauri::{AppHandle, Emitter};
use tempfile::NamedTempFile;
use tokio::task::JoinHandle;
const NOTIFICATION_CHANNEL_NAME : &str = "notification";
const MAX_FILES : usize = 8;
const CHUNK_SIZE : usize = 1024 * 1024;
const PARALLEL_FAILURES : usize = 8;
const MAX_RETRIES : usize = 30;
struct ModelDownloadStatusChannel {
pub model_name: String,
pub model_type: NotificationModelType,
pub final_file_path: PathBuf,
pub receiver: Receiver<ProgressUpdate>,
}
pub async fn downloader_thread(app_data_root: AppDataRoot, app: AppHandle) -> ! {
loop {
app.emit("testing", 1);
let mut status_check_queue : Arc<Mutex<Vec<ModelDownloadStatusChannel>>> = Arc::new(Mutex::new(Vec::new()));
{
let status_check_queue2 = status_check_queue.clone();
let app2 = app.clone();
tokio::spawn(download_notify_thread(status_check_queue2, app2));
}
let mut download_queue = VecDeque::new();
// TODO: Automatic enqueue of known important models + enqueue new models on-demand
download_queue.push_back(ModelType::ClipJson);
download_queue.push_back(ModelType::SdxlTurboUnet);
download_queue.push_back(ModelType::SdxlTurboVae);
download_queue.push_back(ModelType::SdxlTurboClipEncoder);
download_queue.push_back(ModelType::SdxlTurboClipEncoder2);
let mut handles = FuturesUnordered::new();
while let Some(model) = download_queue.pop_front() {
let url = model.get_download_url().to_string();
let filename = app_data_root.weights_dir().model_path(&model);
if filename.exists() {
info!("File already exists: {:?}", &filename);
continue;
}
let headers = None;
println!(">>> Testing: {:?}", app_data_root.path());
std::thread::sleep(std::time::Duration::from_millis(500));
let (tx, rx) = std::sync::mpsc::channel();
info!("Download: {:?}", filename);
let temp_file = match app_data_root.temp_dir().new_named_temp_file() {
Ok(file) => file,
Err(err) => {
error!("Couldn't create a temporary file: {:?}", err);
continue;
}
};
let result = app.emit(NOTIFICATION_CHANNEL_NAME, NotificationEvent::ModelDownloadStarted {
model_name: model.get_name(),
model_type: model.get_notification_type(),
});
if let Err(err) = result {
error!("Could not emit event: {}", err);
}
let task = tokio::spawn(download_async(
url,
temp_file,
filename.clone(),
MAX_FILES,
CHUNK_SIZE,
PARALLEL_FAILURES,
MAX_RETRIES,
headers,
Some(tx),
));
match status_check_queue.lock() {
Err(err) => error!("Could not unlock: {:?}", err),
Ok(mut lock) => {
lock.push(ModelDownloadStatusChannel {
receiver: rx,
final_file_path: filename.clone(),
model_name: model.get_name().to_string(),
model_type: model.get_notification_type(),
})
}
}
handles.push(task);
}
while let Some(result) = handles.next().await {
match result {
Err(err) => error!("Error downloading: {:?}", err),
Ok(_) => {},
}
}
loop {
// TODO(bt,2025-03-10): Queue management for new files entering the queue.
tokio::time::sleep(std::time::Duration::from_millis(60_000)).await;
}
}
pub async fn download_notify_thread(queue: Arc<Mutex<Vec<ModelDownloadStatusChannel>>>, app: AppHandle) -> ! {
loop {
match queue.lock() {
Err(err) => error!("Could not unlock: {:?}", err),
Ok(mut lock) => {
debug!(">>>>> Notifier has {} items", lock.len());
let mut reenqueue = Vec::with_capacity(lock.len());
for item in (*lock).drain(..) {
if let Ok(update) = item.receiver.try_recv() {
let result = app.emit(NOTIFICATION_CHANNEL_NAME, NotificationEvent::ModelDownloadProgress {
model_name: &item.model_name,
model_type: item.model_type,
currently_downloaded_bytes: update.complete,
total_file_bytes: update.total_length,
});
if let Err(err) = result {
error!("Could not emit event: {}", err);
}
}
if item.final_file_path.exists() {
debug!("No longer sending notifications for {:?}", &item.final_file_path);
let result = app.emit(NOTIFICATION_CHANNEL_NAME, NotificationEvent::ModelDownloadComplete {
model_name: &item.model_name,
model_type: item.model_type,
});
if let Err(err) = result {
error!("Could not emit event: {}", err);
}
} else {
reenqueue.push(item);
}
}
lock.extend(reenqueue);
}
}
tokio::time::sleep(std::time::Duration::from_millis(500)).await;
}
}
@@ -0,0 +1,241 @@
//! Implemented from https://github.com/huggingface/hf_transfer/blob/main/src/lib.rs
use crate::transfer::error::Error;
use crate::transfer::util::{exponential_backoff, BASE_WAIT_TIME, MAX_WAIT_TIME};
use futures::stream::FuturesUnordered;
use futures::StreamExt;
use reqwest::header::{
HeaderMap, HeaderName, HeaderValue, AUTHORIZATION, CONTENT_RANGE,
RANGE,
};
use reqwest::Url;
use std::collections::HashMap;
use std::fmt::Display;
use std::fs::remove_file;
use std::io::SeekFrom;
use std::path::Path;
use std::sync::Arc;
use std::sync::atomic::{AtomicUsize, Ordering};
use std::sync::mpsc::Sender;
use std::time::Duration;
use log::info;
use tempfile::NamedTempFile;
use tokio::fs::OpenOptions;
use tokio::io::AsyncWriteExt;
use tokio::io::AsyncSeekExt;
use tokio::sync::Semaphore;
use tokio::time::sleep;
pub struct ProgressUpdate {
pub complete: usize,
pub total_length: usize,
}
/// max_files: Number of open file handles, which determines the maximum number of parallel downloads
/// parallel_failures: Number of maximum failures of different chunks in parallel (cannot exceed max_files)
/// max_retries: Number of maximum attempts per chunk. (Retries are exponentially backed off + jitter)
///
/// The number of threads can be tuned by the environment variable `TOKIO_WORKER_THREADS` as documented in
/// https://docs.rs/tokio/latest/tokio/runtime/struct.Builder.html#method.worker_threads
#[allow(clippy::too_many_arguments)]
pub async fn download_async<P: AsRef<Path>>(
url: String,
filename: NamedTempFile,
final_filename: P,
max_files: usize,
chunk_size: usize,
parallel_failures: usize,
max_retries: usize,
input_headers: Option<HashMap<String, String>>,
maybe_progress_sender: Option<Sender<ProgressUpdate>>,
) -> Result<(), Error> {
info!(">>> Downloading: {}", filename.as_ref().display());
let client = reqwest::Client::builder()
// https://github.com/hyperium/hyper/issues/2136#issuecomment-589488526
.http2_keep_alive_timeout(Duration::from_secs(15))
.build()?;
let mut headers = HeaderMap::new();
let mut auth_token = None;
if let Some(input_headers) = input_headers {
headers.reserve(input_headers.len());
for (k, v) in input_headers {
let name: HeaderName = k
.try_into()
.map_err(|err| Error::new_err(format!("Invalid header: {err}")))?;
let value: HeaderValue = AsRef::<str>::as_ref(&v)
.try_into()
.map_err(|err| Error::new_err(format!("Invalid header value: {err}")))?;
if name == AUTHORIZATION {
auth_token = Some(value);
} else {
headers.insert(name, value);
}
}
};
let response = if let Some(token) = auth_token.as_ref() {
client.get(&url).header(AUTHORIZATION, token)
} else {
client.get(&url)
}
.headers(headers.clone())
.header(RANGE, "bytes=0-0")
.send()
.await
.map_err(|err| Error::new_err(format!("Error while downloading: {err}")))?
.error_for_status()
.map_err(|err| Error::new_err(err.to_string()))?;
// Only call the final redirect URL to avoid overloading the Hub with requests and also
// altering the download count
let redirected_url = response.url();
if Url::parse(&url)
.map_err(|err| Error::new_err(format!("failed to parse url: {err}")))?
.host()
== redirected_url.host()
{
if let Some(token) = auth_token {
headers.insert(AUTHORIZATION, token);
}
}
let content_range = response
.headers()
.get(CONTENT_RANGE)
.ok_or(Error::new_err("No content length".to_string()))?
.to_str()
.map_err(|err| Error::new_err(format!("Error while downloading: {err}")))?;
let size: Vec<&str> = content_range.split('/').collect();
// Content-Range: bytes 0-0/702517648
// https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Content-Range
let length: usize = size
.last()
.ok_or(Error::new_err(
"Error while downloading: No size was detected".to_string(),
))?
.parse()
.map_err(|err| Error::new_err(format!("Error while downloading: {err}")))?;
info!("File {:?} is {:?} length", filename.as_ref(), length);
let mut handles = FuturesUnordered::new();
let semaphore = Arc::new(Semaphore::new(max_files));
let parallel_failures_semaphore = Arc::new(Semaphore::new(parallel_failures));
let max_chunk_complete = Arc::new(AtomicUsize::new(0));
for start in (0..length).step_by(chunk_size) {
let url = redirected_url.to_string();
let filename = filename.as_ref().to_owned();
let client = client.clone();
let headers = headers.clone();
let stop = std::cmp::min(start + chunk_size - 1, length);
let semaphore = semaphore.clone();
let parallel_failures_semaphore = parallel_failures_semaphore.clone();
let maybe_progress = maybe_progress_sender.clone();
let max_chunk_complete = max_chunk_complete.clone();
handles.push(tokio::spawn(async move {
let permit = semaphore
.acquire_owned()
.await
.map_err(|err| Error::new_err(format!("Error while downloading: {err}")))?;
let mut chunk = download_chunk(&client, &url, &filename, start, stop, headers.clone()).await;
let mut i = 0;
if parallel_failures > 0 {
while let Err(dlerr) = chunk {
if i >= max_retries {
return Err(Error::new_err(format!(
"Failed after too many retries ({max_retries}): {dlerr}"
)));
}
let parallel_failure_permit = parallel_failures_semaphore.clone().try_acquire_owned().map_err(|err| {
Error::new_err(format!(
"Failed too many failures in parallel ({parallel_failures}): {dlerr} ({err})"
))
})?;
let wait_time = exponential_backoff(BASE_WAIT_TIME, i, MAX_WAIT_TIME);
sleep(Duration::from_millis(wait_time as u64)).await;
chunk = download_chunk(&client, &url, &filename, start, stop, headers.clone()).await;
i += 1;
drop(parallel_failure_permit);
if let Some(sender) = maybe_progress.as_ref() {
// NB: Calling fetch_max twice. First call will return the lower value.
let _old = max_chunk_complete.fetch_max(stop, Ordering::SeqCst);
let max_chunk = max_chunk_complete.fetch_max(stop, Ordering::SeqCst);
let progress_update = ProgressUpdate {
complete: max_chunk,
total_length: length,
};
if let Err(_err) = sender.send(progress_update) {
// Fail silently, not essential to issue progress updates
}
}
}
}
drop(permit);
chunk.map_err(|e| Error::new_err(format!("Downloading error {e}"))).and(Ok(stop - start))
}));
}
// Output the chained result
while let Some(result) = handles.next().await {
match result {
Ok(Ok(size)) => {
//if let Some(ref callback) = callback {
// callback.call((size,), None)?;
//}
}
Ok(Err(py_err)) => {
return Err(py_err);
}
Err(err) => {
return Err(Error::new_err(format!(
"Error while downloading: {err}"
)));
}
}
}
let file = filename.persist(final_filename)
.map_err(|e| Error::new_err(format!("Could not rename to final filename: {:?}", e)))?;
file.sync_all()?;
Ok(())
}
async fn download_chunk<P: AsRef<Path>>(
client: &reqwest::Client,
url: &str,
filename: P,
start: usize,
stop: usize,
headers: HeaderMap,
) -> Result<(), Error> {
// Process each socket concurrently.
let range = format!("bytes={start}-{stop}");
let mut file = OpenOptions::new()
.write(true)
.truncate(false)
.create(true)
.open(filename)
.await?;
file.seek(SeekFrom::Start(start as u64)).await?;
let response = client
.get(url)
.headers(headers)
.header(RANGE, range)
.send()
.await?
.error_for_status()?;
let content = response.bytes().await?;
file.write_all(&content).await?;
Ok(())
}
@@ -0,0 +1,49 @@
//! Implemented from https://github.com/huggingface/hf_transfer/blob/main/src/lib.rs
use std::fmt::Display;
use reqwest::header::ToStrError;
#[derive(Debug)]
pub enum Error {
Io(std::io::Error),
Request(reqwest::Error),
ToStrError(ToStrError),
Misc(String),
}
impl From<std::io::Error> for Error {
fn from(value: std::io::Error) -> Self {
Self::Io(value)
}
}
impl From<reqwest::Error> for Error {
fn from(value: reqwest::Error) -> Self {
Self::Request(value)
}
}
impl From<ToStrError> for Error {
fn from(value: ToStrError) -> Self {
Self::ToStrError(value)
}
}
impl Display for Error {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
match self {
Self::Io(io) => write!(f, "Io: {io}"),
Self::Request(req) => write!(f, "Request: {req}"),
Self::ToStrError(req) => write!(f, "Response non ascii: {req}"),
Self::Misc(err) => write!(f, "Misc error: {err}"),
}
}
}
impl std::error::Error for Error {}
impl Error {
pub fn new_err(reason: String) -> Self {
Self::Misc(reason)
}
}
@@ -0,0 +1,4 @@
pub mod download;
pub mod upload;
pub mod util;
pub mod error;
@@ -0,0 +1,189 @@
//! Implemented from https://github.com/huggingface/hf_transfer/blob/main/src/lib.rs
use crate::transfer::error::Error;
use crate::transfer::util::{exponential_backoff, BASE_WAIT_TIME, MAX_WAIT_TIME};
use futures::stream::FuturesUnordered;
use futures::StreamExt;
use reqwest::header::CONTENT_LENGTH;
use std::collections::HashMap;
use std::fmt::Display;
use std::io::SeekFrom;
use std::sync::Arc;
use std::time::Duration;
use tokio::fs::OpenOptions;
use tokio::io::{AsyncReadExt, AsyncSeekExt};
use tokio::sync::Semaphore;
use tokio::time::sleep;
use tokio_util::codec::{BytesCodec, FramedRead};
/// parts_urls: Dictionary consisting of part numbers as keys and the associated url as values
/// completion_url: The url that should be called when the upload is finished
/// max_files: Number of open file handles, which determines the maximum number of parallel uploads
/// parallel_failures: Number of maximum failures of different chunks in parallel (cannot exceed max_files)
/// max_retries: Number of maximum attempts per chunk. (Retries are exponentially backed off + jitter)
///
/// The number of threads can be tuned by the environment variable `TOKIO_WORKER_THREADS` as documented in
/// https://docs.rs/tokio/latest/tokio/runtime/struct.Builder.html#method.worker_threads
///
/// See https://docs.aws.amazon.com/AmazonS3/latest/userguide/mpuoverview.html for more information
/// on the multipart upload
#[allow(clippy::too_many_arguments)]
fn multipart_upload(
file_path: String,
parts_urls: Vec<String>,
chunk_size: u64,
max_files: usize,
parallel_failures: usize,
max_retries: usize,
) -> Result<Vec<HashMap<String, String>>, Error> {
if parallel_failures > max_files {
return Err(Error::new_err(
"Error parallel_failures cannot be > max_files".to_string(),
));
}
if (parallel_failures == 0) != (max_retries == 0) {
return Err(Error::new_err(
"For retry mechanism you need to set both `parallel_failures` and `max_retries`"
.to_string(),
));
}
tokio::runtime::Builder::new_multi_thread()
.enable_all()
.build()?
.block_on(async {
upload_async(
file_path,
parts_urls,
chunk_size,
max_files,
parallel_failures,
max_retries,
)
.await
})
}
#[allow(clippy::too_many_arguments)]
async fn upload_async(
file_path: String,
parts_urls: Vec<String>,
chunk_size: u64,
max_files: usize,
parallel_failures: usize,
max_retries: usize,
) -> Result<Vec<HashMap<String, String>>, Error> {
let client = reqwest::Client::new();
let mut handles = FuturesUnordered::new();
let semaphore = Arc::new(Semaphore::new(max_files));
let parallel_failures_semaphore = Arc::new(Semaphore::new(parallel_failures));
for (part_number, part_url) in parts_urls.iter().enumerate() {
let url = part_url.to_string();
let path = file_path.to_owned();
let client = client.clone();
let start = (part_number as u64) * chunk_size;
let semaphore = semaphore.clone();
let parallel_failures_semaphore = parallel_failures_semaphore.clone();
handles.push(tokio::spawn(async move {
let permit = semaphore
.clone()
.acquire_owned()
.await
.map_err(|err| Error::new_err(format!("Error acquiring semaphore: {err}")))?;
let mut chunk = upload_chunk(&client, &url, &path, start, chunk_size).await;
let mut i = 0;
if parallel_failures > 0 {
while let Err(ul_err) = chunk {
if i >= max_retries {
return Err(Error::new_err(format!(
"Failed after too many retries ({max_retries}): {ul_err}"
)));
}
let parallel_failure_permit = parallel_failures_semaphore.clone().try_acquire_owned().map_err(|err| {
Error::new_err(format!(
"Failed too many failures in parallel ({parallel_failures}): {ul_err} ({err})"
))
})?;
let wait_time = exponential_backoff(BASE_WAIT_TIME, i, MAX_WAIT_TIME);
sleep(Duration::from_millis(wait_time as u64)).await;
chunk = upload_chunk(&client, &url, &path, start, chunk_size).await;
i += 1;
drop(parallel_failure_permit);
}
}
drop(permit);
chunk.map_err(|e|{
//match e {
// Error::Io(io) => Error::new_err(format!("Io error {io}")),
// Error::Request(req) => Error::new_err(format!("Error while sending chunk {req}")),
// Error::ToStrError(req) => Error::new_err(format!("Response header contains non ASCII chars: {req}")),
//}
e
}).map(|chunk| (part_number, chunk, chunk_size))
}));
}
let mut results: Vec<HashMap<String, String>> = vec![HashMap::default(); parts_urls.len()];
while let Some(result) = handles.next().await {
match result {
Ok(Ok((part_number, headers, size))) => {
//if let Some(ref callback) = callback {
// callback.call((size,), None)?;
//}
results[part_number] = headers;
}
Ok(Err(py_err)) => {
return Err(py_err);
}
Err(err) => {
return Err(Error::new_err(format!(
"Error occurred while uploading: {err}"
)));
}
}
}
Ok(results)
}
async fn upload_chunk(
client: &reqwest::Client,
url: &str,
path: &str,
start: u64,
chunk_size: u64,
) -> Result<HashMap<String, String>, Error> {
let mut options = OpenOptions::new();
let mut file = options.read(true).open(path).await?;
let file_size = file.metadata().await?.len();
let bytes_transferred = std::cmp::min(file_size - start, chunk_size);
file.seek(SeekFrom::Start(start)).await?;
let chunk = file.take(chunk_size);
let response = client
.put(url)
.header(CONTENT_LENGTH, bytes_transferred)
.body(reqwest::Body::wrap_stream(FramedRead::new(
chunk,
BytesCodec::new(),
)))
.send()
.await?;
let response = response.error_for_status()?;
let mut headers = HashMap::new();
for (name, value) in response.headers().into_iter() {
headers.insert(name.to_string(), value.to_str()?.to_owned());
}
Ok(headers)
}
@@ -0,0 +1,15 @@
//! Implemented from https://github.com/huggingface/hf_transfer/blob/main/src/lib.rs
use rand::{thread_rng, Rng};
use std::fmt::Display;
pub const BASE_WAIT_TIME: usize = 300;
pub const MAX_WAIT_TIME: usize = 10_000;
fn jitter() -> usize {
thread_rng().gen_range(0..=500)
}
pub fn exponential_backoff(base_wait_time: usize, n: usize, max: usize) -> usize {
(base_wait_time + n.pow(2) + jitter()).min(max)
}
@@ -212,7 +212,7 @@ export class RealTimeDrawEngine {
// Add mouse events for preview canvas copying
this.previewCopyListener();
//this.listenToServerEvents();
this.listenToServerEvents();
}