Merge pull request #26 from ALPHA-g-Experiment/ext

Allow filename pattern selection

CHANGELOG.md

@@ -14,6 +14,11 @@ changes will be documented in this file.
  This contains all the information from the sequencer events, but also includes
  the counts during each spill for all Chronobox channels.
 - Download tabs now have a small spinner to indicate that processing is ongoing.
+- It is now possible to specify (via the `--pattern` flag) a filename pattern
+ regex for the MIDAS files given a run number. This is useful in the case that
+ there are both `*.mid` and `*.mid.lz4` files in the data directory. The
+ default pattern will match any `runXXXXXsubYYY.mid*` file. To, for example,
+ match only `*.mid` files, add a `$` anchor to the end of the default pattern.
 
 ## [0.1.1] - 2024-08-24
 

Cargo.toml

@@ -18,6 +18,7 @@ futures = "0.3.30"
 indicatif = "0.17.8"
 jsonwebtoken = "9.3.0"
 rand = "0.8.5"
+regex = "1.10.6"
 semver = "1.0.23"
 serde = { version = "1.0.203", features = ["derive"] }
 serde_json = "1.0.117"

README.md

@@ -75,7 +75,7 @@ After=network.target
 
 [Service]
 Environment="AG_JWT_SECRET=a_random_shared_secret"
-ExecStart=/home/my_user/.cargo/bin/alpha-g-data-handler serve -a 0.0.0.0:8080 -d /path/to/midas/files
+ExecStart=/home/my_user/.cargo/bin/alpha-g-data-handler serve -a 0.0.0.0:8080 -d /path/to/midas/files -p "^run0*(?<run_number>\d+)sub\d+\.mid\.lz4$"
 WorkingDirectory=/home/my_user
 User=my_user
 Restart=always

src/core_command.rs

@@ -37,16 +37,17 @@ pub(super) fn install_core_binaries() -> Result<()> {
  Ok(())
 }
 
-// This is set (only once) at the beginning of the program based on the CLI
+// These are set (only once) at the beginning of the program based on the CLI
 // arguments.
-// Throughout the module it should be assumed to be set.
+// Throughout the module they should be assumed to be set.
 pub(super) static MIDAS_DATA_PATH: OnceLock<PathBuf> = OnceLock::new();
+pub(super) static FILENAME_PATTERN: OnceLock<regex::Regex> = OnceLock::new();
 
 // Get all the MIDAS files for a given run number.
 // If the returned value is OK, the vector is guaranteed to be non-empty.
 async fn midas_files(run_number: u32) -> Result<Vec<PathBuf>> {
  let dir = MIDAS_DATA_PATH.get().unwrap();
- let prefix = format!("run{run_number:05}sub");
+ let re = FILENAME_PATTERN.get().unwrap();
 
  let mut files = Vec::new();
  let mut entries = fs::read_dir(dir)
@@ -57,8 +58,14 @@ async fn midas_files(run_number: u32) -> Result<Vec<PathBuf>> {
  .await
  .with_context(|| format!("failed to iterate over `{}`", dir.display()))?
  {
- if let Some(filename) = entry.file_name().to_str() {
- if filename.starts_with(&prefix) {
+ if let Some(n) = entry
+ .file_name()
+ .to_str()
+ .and_then(|filename| re.captures(filename))
+ .and_then(|captures| captures.name("run_number"))
+ .and_then(|n| n.as_str().parse::<u32>().ok())
+ {
+ if n == run_number {
  files.push(entry.path());
  }
  }

src/main.rs

@@ -46,6 +46,9 @@ enum Commands {
  /// Path to the MIDAS data directory
  #[arg(short, long, default_value = ".")]
  data_dir: std::path::PathBuf,
+ /// Filename pattern
+ #[arg(short, long, default_value = r"^run0*(?<run_number>\d+)sub\d+\.mid")]
+ pattern: regex::Regex,
  },
 }
 
@@ -97,10 +100,17 @@ async fn main() -> Result<(), anyhow::Error> {
 
  spinner.finish_and_clear();
  }
- Commands::Serve { address, data_dir } => {
+ Commands::Serve {
+ address,
+ data_dir,
+ pattern,
+ } => {
  core_command::MIDAS_DATA_PATH
  .set(data_dir)
  .expect("failed to set MIDAS_DATA_PATH");
+ core_command::FILENAME_PATTERN
+ .set(pattern)
+ .expect("failed to set FILENAME_PATTERN");
 
  let app_state = Arc::new(AppState::default());
  let app = Router::new()