feat: Add data server for market data loading
This commit introduces a new data server module (`src/ast/data_server`) designed for high-performance, thread-safe loading and caching of market data. Key components: - `DataServer`: Manages symbol indexing and delegates loading/caching. - `SymbolIndex`: Scans the data directory and maintains a sorted index of data files per symbol. - `FileCache`: Implements a thread-safe cache using `RwLock` and a loading guard to prevent duplicate work. - `loader`: Handles ZIP decompression and binary record parsing from `.bin` files. - `records`: Defines raw and parsed data structures for M1 and tick data. - `SymbolChunkIter`: Provides an iterator over pre-parsed data chunks, prefetching subsequent files. This architecture allows multiple VM threads to access market data concurrently without redundant I/O, mirroring the strategy used in the original Delphi `TDataServer`.
This commit is contained in:
@@ -0,0 +1,518 @@
|
||||
//! High-performance data server for binary market data files.
|
||||
//!
|
||||
//! Provides thread-safe (`Arc`-based) loading and caching of tick data,
|
||||
//! designed to serve hundreds of single-threaded VMs in parallel without
|
||||
//! redundant I/O. Each file is read and parsed once, then shared across
|
||||
//! all consumers via `Arc<[T]>` chunk references.
|
||||
//!
|
||||
//! # Architecture
|
||||
//!
|
||||
//! ```text
|
||||
//! Arc<DataServer>
|
||||
//! ┌──────────────────────────┐
|
||||
//! │ SymbolIndex FileCache │
|
||||
//! │ (HashMap) (RwLock) │
|
||||
//! └─────┬────────────┬───────┘
|
||||
//! │ │
|
||||
//! VM Thread 1 │ VM Thread N
|
||||
//! Generator ──┘ Generator ──┘
|
||||
//! ```
|
||||
|
||||
pub mod cache;
|
||||
pub mod loader;
|
||||
pub mod records;
|
||||
|
||||
use cache::{ChunkVec, FileCache, FileKey};
|
||||
use records::{DataFormat, M1Parsed, TickParsed};
|
||||
|
||||
use regex::Regex;
|
||||
use std::collections::HashMap;
|
||||
use std::path::{Path, PathBuf};
|
||||
use std::sync::{Arc, OnceLock};
|
||||
|
||||
/// Global data server instance, initialized once before any VM threads start.
|
||||
static DATA_SERVER: OnceLock<Arc<DataServer>> = OnceLock::new();
|
||||
|
||||
/// Initializes the global data server for the given base path.
|
||||
///
|
||||
/// Scans the directory for data files and builds the symbol index.
|
||||
/// Must be called once before any `get_data_server()` calls (typically
|
||||
/// at program startup). Returns the `Arc` for optional local use.
|
||||
///
|
||||
/// # Panics
|
||||
///
|
||||
/// Panics if called more than once (the `OnceLock` is already set).
|
||||
pub fn init_data_server(base_path: impl AsRef<Path>) -> Arc<DataServer> {
|
||||
let server = Arc::new(DataServer::new(base_path));
|
||||
DATA_SERVER
|
||||
.set(server.clone())
|
||||
.expect("DataServer already initialized");
|
||||
server
|
||||
}
|
||||
|
||||
/// Returns the global data server, or `None` if not yet initialized.
|
||||
pub fn get_data_server() -> Option<Arc<DataServer>> {
|
||||
DATA_SERVER.get().cloned()
|
||||
}
|
||||
|
||||
/// Pre-scanned directory index mapping symbols to their sorted data files.
|
||||
struct SymbolIndex {
|
||||
base_path: PathBuf,
|
||||
/// Symbol name (e.g. "EURUSD") → sorted list of file keys (chronological).
|
||||
symbols: HashMap<Arc<str>, Vec<FileKey>>,
|
||||
}
|
||||
|
||||
impl SymbolIndex {
|
||||
/// Scans `base_path` for files matching `SYMBOL_YYYY_MM.(m1|tick)` and
|
||||
/// builds a grouped, chronologically sorted index.
|
||||
///
|
||||
/// Mirrors Delphi's `TDataServer.UpdateSymbols`.
|
||||
fn scan(base_path: &Path) -> Self {
|
||||
let re = Regex::new(r"^(.+)_(\d{4})_(\d{2})\.(m1|tick)$").unwrap();
|
||||
let mut map: HashMap<Arc<str>, Vec<FileKey>> = HashMap::new();
|
||||
|
||||
if let Ok(entries) = std::fs::read_dir(base_path) {
|
||||
for entry in entries.flatten() {
|
||||
let file_name = entry.file_name();
|
||||
let name = file_name.to_string_lossy();
|
||||
|
||||
if let Some(caps) = re.captures(&name) {
|
||||
let symbol: Arc<str> = Arc::from(&caps[1]);
|
||||
let year: u16 = caps[2].parse().unwrap_or(0);
|
||||
let month: u8 = caps[3].parse().unwrap_or(0);
|
||||
let format = match &caps[4] {
|
||||
"m1" => DataFormat::M1,
|
||||
"tick" => DataFormat::Tick,
|
||||
_ => continue,
|
||||
};
|
||||
|
||||
if year > 0 && (1..=12).contains(&month) {
|
||||
let key = FileKey {
|
||||
symbol: symbol.clone(),
|
||||
year,
|
||||
month,
|
||||
format,
|
||||
};
|
||||
map.entry(symbol).or_default().push(key);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
// Sort each symbol's files chronologically (year, month, format)
|
||||
for files in map.values_mut() {
|
||||
files.sort_by(|a, b| {
|
||||
(a.year, a.month, a.format as u8).cmp(&(b.year, b.month, b.format as u8))
|
||||
});
|
||||
}
|
||||
|
||||
Self {
|
||||
base_path: base_path.to_path_buf(),
|
||||
symbols: map,
|
||||
}
|
||||
}
|
||||
|
||||
/// Returns the filesystem path for a given file key.
|
||||
fn file_path(&self, key: &FileKey) -> PathBuf {
|
||||
let ext = match key.format {
|
||||
DataFormat::M1 => "m1",
|
||||
DataFormat::Tick => "tick",
|
||||
};
|
||||
self.base_path
|
||||
.join(format!("{}_{:04}_{:02}.{}", key.symbol, key.year, key.month, ext))
|
||||
}
|
||||
}
|
||||
|
||||
/// High-performance data server for binary market data files.
|
||||
///
|
||||
/// Thread-safe (`Send + Sync`). Designed to be wrapped in `Arc` and shared
|
||||
/// across hundreds of VM threads. Each file is loaded and parsed once,
|
||||
/// then served from the in-memory cache.
|
||||
pub struct DataServer {
|
||||
index: SymbolIndex,
|
||||
cache: FileCache,
|
||||
}
|
||||
|
||||
impl std::fmt::Debug for DataServer {
|
||||
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
|
||||
write!(
|
||||
f,
|
||||
"DataServer({:?}, {} symbols)",
|
||||
self.index.base_path,
|
||||
self.index.symbols.len()
|
||||
)
|
||||
}
|
||||
}
|
||||
|
||||
impl DataServer {
|
||||
/// Creates a new data server, scanning the base path for data files.
|
||||
pub fn new(base_path: impl AsRef<Path>) -> Self {
|
||||
Self {
|
||||
index: SymbolIndex::scan(base_path.as_ref()),
|
||||
cache: FileCache::new(),
|
||||
}
|
||||
}
|
||||
|
||||
/// Returns a sorted list of all known symbol names.
|
||||
pub fn symbols(&self) -> Vec<Arc<str>> {
|
||||
let mut syms: Vec<_> = self.index.symbols.keys().cloned().collect();
|
||||
syms.sort();
|
||||
syms
|
||||
}
|
||||
|
||||
/// Returns the number of data files for a given symbol and format,
|
||||
/// or `None` if the symbol is unknown.
|
||||
pub fn file_count(&self, symbol: &str, format: DataFormat) -> Option<usize> {
|
||||
self.index.symbols.get(symbol).map(|files| {
|
||||
files.iter().filter(|f| f.format == format).count()
|
||||
})
|
||||
}
|
||||
|
||||
/// Returns a chunk iterator over all M1 data for a symbol.
|
||||
///
|
||||
/// The iterator yields `Arc<[M1Parsed]>` chunks in chronological order,
|
||||
/// prefetching the next file while the current one is consumed.
|
||||
pub fn stream_m1(self: &Arc<Self>, symbol: &str) -> Option<SymbolChunkIter<M1Parsed>> {
|
||||
let files: Vec<FileKey> = self
|
||||
.index
|
||||
.symbols
|
||||
.get(symbol)?
|
||||
.iter()
|
||||
.filter(|f| f.format == DataFormat::M1)
|
||||
.cloned()
|
||||
.collect();
|
||||
|
||||
if files.is_empty() {
|
||||
return None;
|
||||
}
|
||||
|
||||
// Eagerly load the first file and prefetch the second
|
||||
self.ensure_m1_loaded(&files[0]);
|
||||
if files.len() > 1 {
|
||||
self.prefetch_m1(&files[1]);
|
||||
}
|
||||
|
||||
Some(SymbolChunkIter {
|
||||
server: Arc::clone(self),
|
||||
files,
|
||||
file_idx: 0,
|
||||
chunk_idx: 0,
|
||||
current_file_chunks: None,
|
||||
})
|
||||
}
|
||||
|
||||
/// Returns a chunk iterator over all tick data for a symbol.
|
||||
pub fn stream_tick(self: &Arc<Self>, symbol: &str) -> Option<SymbolChunkIter<TickParsed>> {
|
||||
let files: Vec<FileKey> = self
|
||||
.index
|
||||
.symbols
|
||||
.get(symbol)?
|
||||
.iter()
|
||||
.filter(|f| f.format == DataFormat::Tick)
|
||||
.cloned()
|
||||
.collect();
|
||||
|
||||
if files.is_empty() {
|
||||
return None;
|
||||
}
|
||||
|
||||
self.ensure_tick_loaded(&files[0]);
|
||||
if files.len() > 1 {
|
||||
self.prefetch_tick(&files[1]);
|
||||
}
|
||||
|
||||
Some(SymbolChunkIter {
|
||||
server: Arc::clone(self),
|
||||
files,
|
||||
file_idx: 0,
|
||||
chunk_idx: 0,
|
||||
current_file_chunks: None,
|
||||
})
|
||||
}
|
||||
|
||||
// -- M1 loading ----------------------------------------------------------
|
||||
|
||||
/// Ensures the given M1 file is loaded into the cache (blocking).
|
||||
fn ensure_m1_loaded(&self, key: &FileKey) {
|
||||
if self.cache.is_cached(key) {
|
||||
return;
|
||||
}
|
||||
if self.cache.is_loading(key) {
|
||||
self.cache.wait_for_load(key);
|
||||
if self.cache.is_cached(key) {
|
||||
return;
|
||||
}
|
||||
}
|
||||
self.load_m1_sync(key);
|
||||
}
|
||||
|
||||
/// Synchronously loads an M1 file into the cache.
|
||||
fn load_m1_sync(&self, key: &FileKey) {
|
||||
if !self.cache.try_claim_loading(key) {
|
||||
// Another thread is loading or already cached — wait for it
|
||||
self.cache.wait_for_load(key);
|
||||
return;
|
||||
}
|
||||
let path = self.index.file_path(key);
|
||||
match loader::load_m1_file(&path) {
|
||||
Ok(chunks) => self.cache.insert_m1(key.clone(), chunks),
|
||||
Err(e) => {
|
||||
eprintln!("Warning: failed to load {}: {}", path.display(), e);
|
||||
// Insert empty to unblock waiters and prevent retries
|
||||
self.cache.insert_m1(key.clone(), vec![]);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// Prefetches an M1 file in a background thread.
|
||||
///
|
||||
/// Requires `Arc<Self>` to safely share the server with the spawned thread.
|
||||
fn prefetch_m1(self: &Arc<Self>, key: &FileKey) {
|
||||
if self.cache.is_cached(key) || !self.cache.try_claim_loading(key) {
|
||||
return;
|
||||
}
|
||||
let path = self.index.file_path(key);
|
||||
let cache_key = key.clone();
|
||||
let server = Arc::clone(self);
|
||||
std::thread::spawn(move || {
|
||||
match loader::load_m1_file(&path) {
|
||||
Ok(chunks) => server.cache.insert_m1(cache_key, chunks),
|
||||
Err(e) => {
|
||||
eprintln!("Warning: prefetch failed for {}: {}", path.display(), e);
|
||||
server.cache.insert_m1(cache_key, vec![]);
|
||||
}
|
||||
}
|
||||
});
|
||||
}
|
||||
|
||||
// -- Tick loading --------------------------------------------------------
|
||||
|
||||
/// Ensures the given tick file is loaded into the cache (blocking).
|
||||
fn ensure_tick_loaded(&self, key: &FileKey) {
|
||||
if self.cache.is_cached(key) {
|
||||
return;
|
||||
}
|
||||
if self.cache.is_loading(key) {
|
||||
self.cache.wait_for_load(key);
|
||||
if self.cache.is_cached(key) {
|
||||
return;
|
||||
}
|
||||
}
|
||||
self.load_tick_sync(key);
|
||||
}
|
||||
|
||||
/// Synchronously loads a tick file into the cache.
|
||||
fn load_tick_sync(&self, key: &FileKey) {
|
||||
if !self.cache.try_claim_loading(key) {
|
||||
self.cache.wait_for_load(key);
|
||||
return;
|
||||
}
|
||||
let path = self.index.file_path(key);
|
||||
match loader::load_tick_file(&path) {
|
||||
Ok(chunks) => self.cache.insert_tick(key.clone(), chunks),
|
||||
Err(e) => {
|
||||
eprintln!("Warning: failed to load {}: {}", path.display(), e);
|
||||
self.cache.insert_tick(key.clone(), vec![]);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/// Prefetches a tick file in a background thread.
|
||||
fn prefetch_tick(self: &Arc<Self>, key: &FileKey) {
|
||||
if self.cache.is_cached(key) || !self.cache.try_claim_loading(key) {
|
||||
return;
|
||||
}
|
||||
let path = self.index.file_path(key);
|
||||
let cache_key = key.clone();
|
||||
let server = Arc::clone(self);
|
||||
std::thread::spawn(move || {
|
||||
match loader::load_tick_file(&path) {
|
||||
Ok(chunks) => server.cache.insert_tick(cache_key, chunks),
|
||||
Err(e) => {
|
||||
eprintln!("Warning: prefetch failed for {}: {}", path.display(), e);
|
||||
server.cache.insert_tick(cache_key, vec![]);
|
||||
}
|
||||
}
|
||||
});
|
||||
}
|
||||
}
|
||||
|
||||
/// Iterator yielding chunks of parsed records for a single symbol.
|
||||
///
|
||||
/// Traverses all files for the symbol in chronological order, yielding
|
||||
/// each 1024-record chunk as an `Arc<[T]>`. When advancing to the next
|
||||
/// file, the file after that is prefetched in the background.
|
||||
///
|
||||
/// This is the Rust equivalent of Delphi's `ProcessFile` chain with its
|
||||
/// `LoadDataFile(nextFileInfo)` prefetch call.
|
||||
pub struct SymbolChunkIter<T> {
|
||||
server: Arc<DataServer>,
|
||||
files: Vec<FileKey>,
|
||||
file_idx: usize,
|
||||
chunk_idx: usize,
|
||||
/// Cached reference to the current file's chunks to avoid repeated
|
||||
/// cache lookups within the same file.
|
||||
current_file_chunks: Option<ChunkVec<T>>,
|
||||
}
|
||||
|
||||
impl SymbolChunkIter<M1Parsed> {
|
||||
/// Returns the next chunk of M1 records, or `None` when exhausted.
|
||||
pub fn next_chunk(&mut self) -> Option<Arc<[M1Parsed]>> {
|
||||
loop {
|
||||
// Try to get chunks for the current file
|
||||
if self.current_file_chunks.is_none() && self.file_idx < self.files.len() {
|
||||
let key = &self.files[self.file_idx];
|
||||
self.server.ensure_m1_loaded(key);
|
||||
self.current_file_chunks = self.server.cache.get_m1(key);
|
||||
self.chunk_idx = 0;
|
||||
}
|
||||
|
||||
let chunks = self.current_file_chunks.as_ref()?;
|
||||
|
||||
if self.chunk_idx < chunks.len() {
|
||||
let chunk = chunks[self.chunk_idx].clone();
|
||||
self.chunk_idx += 1;
|
||||
return Some(chunk);
|
||||
}
|
||||
|
||||
// Current file exhausted — advance to next file
|
||||
self.file_idx += 1;
|
||||
self.current_file_chunks = None;
|
||||
|
||||
if self.file_idx >= self.files.len() {
|
||||
return None; // All files exhausted
|
||||
}
|
||||
|
||||
// Prefetch the file AFTER the next one (N+2), since N+1 was
|
||||
// already prefetched when we started N.
|
||||
if self.file_idx + 1 < self.files.len() {
|
||||
self.server.prefetch_m1(&self.files[self.file_idx + 1]);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
impl SymbolChunkIter<TickParsed> {
|
||||
/// Returns the next chunk of tick records, or `None` when exhausted.
|
||||
pub fn next_chunk(&mut self) -> Option<Arc<[TickParsed]>> {
|
||||
loop {
|
||||
if self.current_file_chunks.is_none() && self.file_idx < self.files.len() {
|
||||
let key = &self.files[self.file_idx];
|
||||
self.server.ensure_tick_loaded(key);
|
||||
self.current_file_chunks = self.server.cache.get_tick(key);
|
||||
self.chunk_idx = 0;
|
||||
}
|
||||
|
||||
let chunks = self.current_file_chunks.as_ref()?;
|
||||
|
||||
if self.chunk_idx < chunks.len() {
|
||||
let chunk = chunks[self.chunk_idx].clone();
|
||||
self.chunk_idx += 1;
|
||||
return Some(chunk);
|
||||
}
|
||||
|
||||
self.file_idx += 1;
|
||||
self.current_file_chunks = None;
|
||||
|
||||
if self.file_idx >= self.files.len() {
|
||||
return None;
|
||||
}
|
||||
|
||||
if self.file_idx + 1 < self.files.len() {
|
||||
self.server.prefetch_tick(&self.files[self.file_idx + 1]);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
#[cfg(test)]
|
||||
mod tests {
|
||||
use super::*;
|
||||
|
||||
#[test]
|
||||
fn test_symbol_index_scan() {
|
||||
// Create a temp directory with mock files
|
||||
let dir = tempfile::tempdir().unwrap();
|
||||
let names = [
|
||||
"EURUSD_2017_01.m1",
|
||||
"EURUSD_2017_02.m1",
|
||||
"EURUSD_2017_03.m1",
|
||||
"EURUSD_2017_01.tick",
|
||||
"GBPUSD_2020_06.m1",
|
||||
"invalid_file.txt",
|
||||
"AAPL.US_2006_08.m1",
|
||||
];
|
||||
for name in &names {
|
||||
std::fs::write(dir.path().join(name), b"").unwrap();
|
||||
}
|
||||
|
||||
let index = SymbolIndex::scan(dir.path());
|
||||
|
||||
// EURUSD has both m1 and tick files
|
||||
let eur = index.symbols.get("EURUSD").unwrap();
|
||||
assert_eq!(eur.len(), 4); // 3 m1 + 1 tick
|
||||
// Should be sorted chronologically, M1 before Tick within same month
|
||||
assert_eq!(eur[0].month, 1);
|
||||
assert_eq!(eur[0].format, DataFormat::M1);
|
||||
// The tick file for month 1 comes right after the m1 file for month 1
|
||||
assert_eq!(eur[1].month, 1);
|
||||
assert_eq!(eur[1].format, DataFormat::Tick);
|
||||
|
||||
// AAPL.US with dot in symbol name
|
||||
let aapl = index.symbols.get("AAPL.US").unwrap();
|
||||
assert_eq!(aapl.len(), 1);
|
||||
assert_eq!(aapl[0].year, 2006);
|
||||
|
||||
// GBPUSD
|
||||
assert!(index.symbols.contains_key("GBPUSD"));
|
||||
|
||||
// Invalid file should not appear
|
||||
assert!(!index.symbols.contains_key("invalid_file"));
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_file_path_construction() {
|
||||
let index = SymbolIndex {
|
||||
base_path: PathBuf::from("/mnt/tickdata/Pepperstone"),
|
||||
symbols: HashMap::new(),
|
||||
};
|
||||
|
||||
let key = FileKey {
|
||||
symbol: Arc::from("EURUSD"),
|
||||
year: 2017,
|
||||
month: 3,
|
||||
format: DataFormat::M1,
|
||||
};
|
||||
|
||||
assert_eq!(
|
||||
index.file_path(&key),
|
||||
PathBuf::from("/mnt/tickdata/Pepperstone/EURUSD_2017_03.m1")
|
||||
);
|
||||
|
||||
let tick_key = FileKey {
|
||||
symbol: Arc::from("AAPL.US"),
|
||||
year: 2006,
|
||||
month: 8,
|
||||
format: DataFormat::Tick,
|
||||
};
|
||||
|
||||
assert_eq!(
|
||||
index.file_path(&tick_key),
|
||||
PathBuf::from("/mnt/tickdata/Pepperstone/AAPL.US_2006_08.tick")
|
||||
);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_data_server_symbols() {
|
||||
let dir = tempfile::tempdir().unwrap();
|
||||
for name in ["EURUSD_2017_01.m1", "GBPUSD_2020_06.m1"] {
|
||||
std::fs::write(dir.path().join(name), b"").unwrap();
|
||||
}
|
||||
|
||||
let server = DataServer::new(dir.path());
|
||||
let syms = server.symbols();
|
||||
assert_eq!(syms.len(), 2);
|
||||
assert_eq!(syms[0].as_ref(), "EURUSD");
|
||||
assert_eq!(syms[1].as_ref(), "GBPUSD");
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user