Bump versions

.editorconfig

@@ -0,0 +1,13 @@
+root = true
+
+[*]
+indent_style = tab
+indent_size = 4
+end_of_line = lf
+charset = utf-8
+trim_trailing_whitespace = true
+insert_final_newline = true
+
+[*.yml]
+indent_style = space
+indent_size = 2

Cargo.lock

@@ -2481,7 +2481,7 @@ checksum = "9b4f627cb1b25917193a259e49bdad08f671f8d9708acfd5fe0a8c1455d87220"
 
 [[package]]
 name = "pile"
-version = "0.0.1"
+version = "0.0.2"
 dependencies = [
  "anstyle",
  "anyhow",
@@ -2504,7 +2504,7 @@ dependencies = [
 
 [[package]]
 name = "pile-config"
-version = "0.0.1"
+version = "0.0.2"
 dependencies = [
  "serde",
  "smartstring",
@@ -2514,7 +2514,7 @@ dependencies = [
 
 [[package]]
 name = "pile-dataset"
-version = "0.0.1"
+version = "0.0.2"
 dependencies = [
  "axum",
  "chrono",
@@ -2535,7 +2535,7 @@ dependencies = [
 
 [[package]]
 name = "pile-flac"
-version = "0.0.1"
+version = "0.0.2"
 dependencies = [
  "base64",
  "itertools 0.14.0",
@@ -2549,7 +2549,7 @@ dependencies = [
 
 [[package]]
 name = "pile-toolbox"
-version = "0.0.1"
+version = "0.0.2"
 dependencies = [
  "thiserror",
  "tokio",
@@ -2557,7 +2557,7 @@ dependencies = [
 
 [[package]]
 name = "pile-value"
-version = "0.0.1"
+version = "0.0.2"
 dependencies = [
  "async-trait",
  "aws-sdk-s3",

Cargo.toml

@@ -5,7 +5,7 @@ resolver = "2"
 [workspace.package]
 rust-version = "1.94.0"
 edition = "2024"
-version = "0.0.1"
+version = "0.0.2"
 
 [workspace.lints.rust]
 unused_import_braces = "deny"
@@ -73,15 +73,15 @@ pile-value = { path = "crates/pile-value" }
 tantivy = "0.25.0"
 axum = { version = "0.8.8", features = ["macros", "multipart"] }
 utoipa = { version = "5.4.0", features = [
-    "axum_extras",
-    "chrono",
-    "url",
-    "uuid",
+	"axum_extras",
+	"chrono",
+	"url",
+	"uuid",
 ] }
 utoipa-swagger-ui = { version = "9.0.2", features = [
-    "axum",
-    "debug-embed",
-    "vendored",
+	"axum",
+	"debug-embed",
+	"vendored",
 ] }
 
 # Async & Parallelism

bump-version.sh

@@ -0,0 +1,26 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+CARGO_TOML="$(dirname "$0")/Cargo.toml"
+DEFAULT_NIX="$(dirname "$0")/default.nix"
+
+# Read current version from workspace Cargo.toml
+current=$(grep '^version = ' "$CARGO_TOML" | head -1 | sed 's/version = "\(.*\)"/\1/')
+
+echo "Current version: $current"
+read -rp "New version: " new
+
+if [[ -z "$new" ]]; then
+	echo "No version entered. Aborting." >&2
+	exit 1
+fi
+
+# Update Cargo.toml workspace version
+sed -i "s/^version = \"$current\"/version = \"$new\"/" "$CARGO_TOML"
+
+# Update default.nix version field
+sed -i "s/version = \"$current\";/version = \"$new\";/" "$DEFAULT_NIX"
+
+echo "Bumped $current -> $new in:"
+echo "  $CARGO_TOML"
+echo "  $DEFAULT_NIX"

crates/pile-value/build.rs

@@ -23,6 +23,13 @@ fn main() {
 		.expect("unexpected OUT_DIR structure")
 		.to_path_buf();
 
+	// If PDFIUM_LIB_DIR is set (e.g. by Nix), use the pre-installed library directly.
+	if let Ok(lib_dir) = env::var("PDFIUM_LIB_DIR") {
+		println!("cargo:rustc-link-search=native={lib_dir}");
+		println!("cargo:rustc-link-lib=dylib=pdfium");
+		return;
+	}
+
 	let lib_path = profile_dir.join("libpdfium.so");
 
 	if !lib_path.exists() {

crates/pile-value/src/extract/string.rs

@@ -4,6 +4,20 @@ use std::sync::Arc;
 
 use crate::{extract::traits::ObjectExtractor, value::PileValue};
 
+fn parse_name(s: &str) -> (&str, Option<&str>) {
+	match s.find('(') {
+		None => (s, None),
+		Some(i) => {
+			let name = &s[..i];
+			let rest = &s[i + 1..];
+			match rest.strip_suffix(')') {
+				Some(args) => (name, Some(args)),
+				None => (name, None),
+			}
+		}
+	}
+}
+
 pub struct StringExtractor {
 	item: Arc<SmartString<LazyCompact>>,
 }
@@ -17,24 +31,49 @@ impl StringExtractor {
 #[async_trait::async_trait]
 impl ObjectExtractor for StringExtractor {
 	async fn field(&self, name: &Label) -> Result<Option<PileValue>, std::io::Error> {
-		Ok(match name.as_str() {
-			"trim" => Some(PileValue::String(Arc::new(
+		let (name, args) = parse_name(name.as_str());
+		Ok(match (name, args) {
+			("trim", None) => Some(PileValue::String(Arc::new(
 				self.item.as_str().trim().into(),
 			))),
 
-			"upper" => Some(PileValue::String(Arc::new(
+			("upper", None) => Some(PileValue::String(Arc::new(
 				self.item.as_str().to_lowercase().into(),
 			))),
 
-			"lower" => Some(PileValue::String(Arc::new(
+			("lower", None) => Some(PileValue::String(Arc::new(
 				self.item.as_str().to_uppercase().into(),
 			))),
 
-			"nonempty" => Some(match self.item.is_empty() {
+			("nonempty", None) => Some(match self.item.is_empty() {
 				true => PileValue::Null,
 				false => PileValue::String(self.item.clone()),
 			}),
 
+			("trimprefix", Some(prefix)) => Some(PileValue::String(Arc::new(
+				self.item
+					.as_str()
+					.strip_prefix(prefix)
+					.unwrap_or(self.item.as_str())
+					.into(),
+			))),
+
+			("trimsuffix", Some(suffix)) => Some(PileValue::String(Arc::new(
+				self.item
+					.as_str()
+					.strip_suffix(suffix)
+					.unwrap_or(self.item.as_str())
+					.into(),
+			))),
+
+			("split", Some(by)) => Some(PileValue::Array(Arc::new(
+				self.item
+					.as_str()
+					.split(by)
+					.map(|s| PileValue::String(Arc::new(s.into())))
+					.collect(),
+			))),
+
 			_ => None,
 		})
 	}
@@ -49,3 +88,134 @@ impl ObjectExtractor for StringExtractor {
 		]);
 	}
 }
+
+#[cfg(test)]
+mod tests {
+	use super::*;
+
+	fn extractor(s: &str) -> StringExtractor {
+		StringExtractor::new(&Arc::new(s.into()))
+	}
+
+	#[expect(clippy::unwrap_used)]
+	async fn field(ext: &StringExtractor, name: &str) -> Option<PileValue> {
+		ext.field(&Label::new(name).unwrap()).await.unwrap()
+	}
+
+	fn string(v: Option<PileValue>) -> Option<String> {
+		match v? {
+			PileValue::String(s) => Some(s.as_str().to_owned()),
+			_ => panic!("expected string"),
+		}
+	}
+
+	fn array(v: Option<PileValue>) -> Vec<String> {
+		match v.expect("expected Some") {
+			PileValue::Array(arr) => arr
+				.iter()
+				.map(|v| match v {
+					PileValue::String(s) => s.as_str().to_owned(),
+					_ => panic!("expected string element"),
+				})
+				.collect(),
+			_ => panic!("expected array"),
+		}
+	}
+
+	#[tokio::test]
+	async fn trim() {
+		assert_eq!(
+			string(field(&extractor("  hi  "), "trim").await),
+			Some("hi".into())
+		);
+	}
+
+	#[tokio::test]
+	async fn trim_no_args() {
+		assert!(field(&extractor("x"), "trim(foo)").await.is_none());
+	}
+
+	#[tokio::test]
+	async fn nonempty_with_content() {
+		assert!(matches!(
+			field(&extractor("hello"), "nonempty").await,
+			Some(PileValue::String(_))
+		));
+	}
+
+	#[tokio::test]
+	async fn nonempty_empty_string() {
+		assert!(matches!(
+			field(&extractor(""), "nonempty").await,
+			Some(PileValue::Null)
+		));
+	}
+
+	#[tokio::test]
+	async fn trimprefix_present() {
+		assert_eq!(
+			string(field(&extractor("foobar"), "trimprefix(foo)").await),
+			Some("bar".into())
+		);
+	}
+
+	#[tokio::test]
+	async fn trimprefix_absent() {
+		assert_eq!(
+			string(field(&extractor("foobar"), "trimprefix(baz)").await),
+			Some("foobar".into())
+		);
+	}
+
+	#[tokio::test]
+	async fn trimprefix_no_args() {
+		assert!(field(&extractor("foobar"), "trimprefix").await.is_none());
+	}
+
+	#[tokio::test]
+	async fn trimsuffix_present() {
+		assert_eq!(
+			string(field(&extractor("foobar"), "trimsuffix(bar)").await),
+			Some("foo".into())
+		);
+	}
+
+	#[tokio::test]
+	async fn trimsuffix_absent() {
+		assert_eq!(
+			string(field(&extractor("foobar"), "trimsuffix(baz)").await),
+			Some("foobar".into())
+		);
+	}
+
+	#[tokio::test]
+	async fn split_basic() {
+		assert_eq!(
+			array(field(&extractor("a,b,c"), "split(,)").await),
+			vec!["a", "b", "c"]
+		);
+	}
+
+	#[tokio::test]
+	async fn split_no_match() {
+		assert_eq!(
+			array(field(&extractor("abc"), "split(,)").await),
+			vec!["abc"]
+		);
+	}
+
+	#[tokio::test]
+	async fn split_no_args() {
+		assert!(field(&extractor("abc"), "split").await.is_none());
+	}
+
+	#[tokio::test]
+	async fn split_unclosed_paren() {
+		assert!(field(&extractor("abc"), "split(,").await.is_none());
+	}
+
+	#[tokio::test]
+	async fn unknown_field() {
+		assert!(field(&extractor("abc"), "bogus").await.is_none());
+	}
+}

default.nix

@@ -0,0 +1,52 @@
+let
+  rustOverlay = import (
+    builtins.fetchTarball {
+      url = "https://github.com/oxalica/rust-overlay/archive/master.tar.gz";
+      sha256 = "0qgrkgc695a7gja83dngxrcx4gdg9056gvg5325i5yyjxg0ni6c9";
+    }
+  );
+  pkgsDefault = import <nixpkgs> { overlays = [ rustOverlay ]; };
+  rustToolchain = pkgsDefault.rust-bin.fromRustupToolchainFile ./rust-toolchain.toml;
+  rustPlatformDefault = pkgsDefault.makeRustPlatform {
+    cargo = rustToolchain;
+    rustc = rustToolchain;
+  };
+in
+{
+  pkgs ? pkgsDefault,
+  pileRustPlatform ? rustPlatformDefault,
+}:
+
+pileRustPlatform.buildRustPackage {
+  pname = "pile";
+  version = "0.0.2";
+  src = ./.;
+
+  cargoLock.lockFile = ./Cargo.lock;
+
+  PDFIUM_LIB_DIR = "${pkgs.pdfium-binaries}/lib";
+
+  buildInputs = [
+    pkgs.pdfium-binaries
+    pkgs.openssl
+  ]
+  ++ pkgs.lib.optionals pkgs.stdenv.isDarwin [
+    pkgs.darwin.apple_sdk.frameworks.Security
+    pkgs.darwin.apple_sdk.frameworks.SystemConfiguration
+  ];
+
+  nativeBuildInputs = [
+    pkgs.pkg-config
+    pkgs.makeWrapper
+  ];
+
+  postInstall = ''
+    wrapProgram $out/bin/pile \
+      --prefix LD_LIBRARY_PATH : ${pkgs.lib.makeLibraryPath [ pkgs.pdfium-binaries ]}
+  '';
+
+  meta = {
+    description = "pile - flexible file indexing";
+    mainProgram = "pile";
+  };
+}

flake.nix

@@ -0,0 +1,40 @@
+{
+  description = "pile - personal data indexer";
+
+  inputs = {
+    nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
+    rust-overlay = {
+      url = "github:oxalica/rust-overlay";
+      inputs.nixpkgs.follows = "nixpkgs";
+    };
+    flake-utils.url = "github:numtide/flake-utils";
+  };
+
+  outputs = { self, nixpkgs, rust-overlay, flake-utils }:
+    flake-utils.lib.eachDefaultSystem (system:
+      let
+        overlays = [ (import rust-overlay) ];
+        pkgs = import nixpkgs { inherit system overlays; };
+
+        rustToolchain = pkgs.rust-bin.fromRustupToolchainFile ./rust-toolchain.toml;
+
+        rustPlatform = pkgs.makeRustPlatform {
+          cargo = rustToolchain;
+          rustc = rustToolchain;
+        };
+
+      in {
+        packages.default = import ./default.nix { inherit pkgs; pileRustPlatform = rustPlatform; };
+
+        devShells.default = pkgs.mkShell {
+          buildInputs = [
+            rustToolchain
+            pkgs.pdfium-binaries
+            pkgs.openssl
+            pkgs.pkg-config
+          ];
+
+          LD_LIBRARY_PATH = pkgs.lib.makeLibraryPath [ pkgs.pdfium-binaries ];
+        };
+      });
+}

rust-toolchain.toml

@@ -0,0 +1,2 @@
+[toolchain]
+channel = "1.94.0"