Extractor rewrite

crates/pile-value/src/source/misc.rs

@@ -0,0 +1,121 @@
+use chrono::{DateTime, Utc};
+use std::fs;
+use std::path::Path;
+
+/// Returns the age of a path as a [DateTime].
+/// - If the path doesn't exist, returns [None]
+/// - If it's a file, returns the modified time
+/// - If it's a directory, returns the LATEST modified time of all files within
+pub fn path_ts_latest(path: impl AsRef<Path>) -> Result<Option<DateTime<Utc>>, std::io::Error> {
+	let path = path.as_ref();
+	if !path.exists() {
+		return Ok(None);
+	}
+
+	let metadata = fs::metadata(path)?;
+
+	if metadata.is_file() {
+		let modified = metadata.modified()?;
+		Ok(Some(modified.into()))
+	} else if metadata.is_dir() {
+		find_latest_modified(path)
+	} else {
+		Ok(None)
+	}
+}
+
+/// Returns the age of a path as a [DateTime].
+/// - If the path doesn't exist, returns [None]
+/// - If it's a file, returns the modified time
+/// - If it's a directory, returns the EARLIEST modified time of all files within
+pub fn path_ts_earliest(path: impl AsRef<Path>) -> Result<Option<DateTime<Utc>>, std::io::Error> {
+	let path = path.as_ref();
+	if !path.exists() {
+		return Ok(None);
+	}
+
+	let metadata = fs::metadata(path)?;
+
+	if metadata.is_file() {
+		let modified = metadata.modified()?;
+		Ok(Some(modified.into()))
+	} else if metadata.is_dir() {
+		find_earliest_modified(path)
+	} else {
+		Ok(None)
+	}
+}
+
+fn find_latest_modified(dir: &Path) -> Result<Option<DateTime<Utc>>, std::io::Error> {
+	let mut latest: Option<DateTime<Utc>> = None;
+
+	// Include the directory's own modification time
+	let dir_metadata = fs::metadata(dir)?;
+	if let Ok(modified) = dir_metadata.modified() {
+		let dt: DateTime<Utc> = modified.into();
+		latest = Some(dt);
+	}
+
+	let entries = fs::read_dir(dir)?;
+
+	for entry in entries.flatten() {
+		let path = entry.path();
+		let metadata = entry.metadata()?;
+
+		if metadata.is_file() {
+			if let Ok(modified) = metadata.modified() {
+				let dt: DateTime<Utc> = modified.into();
+				latest = Some(match latest {
+					Some(prev) if prev > dt => prev,
+					_ => dt,
+				});
+			}
+		} else if metadata.is_dir()
+			&& let Some(dir_latest) = find_latest_modified(&path)?
+		{
+			latest = Some(match latest {
+				Some(prev) if prev > dir_latest => prev,
+				_ => dir_latest,
+			});
+		}
+	}
+
+	return Ok(latest);
+}
+
+fn find_earliest_modified(dir: &Path) -> Result<Option<DateTime<Utc>>, std::io::Error> {
+	let mut earliest: Option<DateTime<Utc>> = None;
+
+	// Include the directory's own modification time
+	let dir_metadata = fs::metadata(dir)?;
+	if let Ok(modified) = dir_metadata.modified() {
+		let dt: DateTime<Utc> = modified.into();
+		earliest = Some(dt);
+	}
+
+	let entries = fs::read_dir(dir)?;
+
+	for entry in entries.flatten() {
+		let path = entry.path();
+		let metadata = entry.metadata()?;
+
+		if metadata.is_file() {
+			if let Ok(modified) = metadata.modified() {
+				let dt: DateTime<Utc> = modified.into();
+				earliest = Some(match earliest {
+					Some(prev) if prev < dt => prev,
+					_ => dt,
+				});
+			}
+		} else if metadata.is_dir()
+			&& let Some(dir_earliest) = find_earliest_modified(&path)?
+		{
+			earliest = Some(match earliest {
+				Some(prev) if prev < dir_earliest => prev,
+				_ => dir_earliest,
+			});
+		}
+	}
+
+	return Ok(earliest);
+}