feat: add persistent CPU search engine and CLI with lifecycle tests

crates/searchi-core/tests/index_options.rs

@@ -0,0 +1,136 @@
+use anyhow::Result;
+use searchi_core::{IndexOptions, SearchIndex, SearchOptions};
+use std::fs;
+
+#[test]
+fn exclusions_database_and_unchanged_metadata() -> Result<()> {
+    let root = tempfile::tempdir()?;
+    for directory in [".git", "node_modules", "target", "private"] {
+        fs::create_dir(root.path().join(directory))?;
+        fs::write(root.path().join(directory).join("hidden.txt"), "secret")?;
+    }
+    fs::write(root.path().join("visible.txt"), "original content")?;
+    let mut index = SearchIndex::open(root.path().join("index.db"))?;
+    let options = IndexOptions {
+        excluded_directory_names: vec![
+            ".git".into(),
+            "node_modules".into(),
+            "target".into(),
+            "private".into(),
+        ],
+    };
+    assert_eq!(
+        index
+            .index_folder_with_options(root.path(), &options)?
+            .indexed,
+        1
+    );
+    assert_eq!(index.stats()?.files, 1);
+    let unchanged = index.index_folder_with_options(root.path(), &options)?;
+    assert_eq!(unchanged.unchanged, 1);
+    assert_eq!(unchanged.indexed, 0);
+    assert!(index.search("secret", 10)?.is_empty());
+    assert!(index.search("index.db", 10)?.is_empty());
+    fs::write(
+        root.path().join("visible.txt"),
+        "changed and longer content",
+    )?;
+    assert_eq!(
+        index
+            .index_folder_with_options(root.path(), &options)?
+            .indexed,
+        1
+    );
+    assert!(index.search("original", 10)?.is_empty());
+    assert_eq!(index.search("changed", 10)?.len(), 1);
+    // A new exclusion removes previously indexed entries from this root.
+    let mut options = IndexOptions::default();
+    index.index_folder_with_options(root.path(), &options)?;
+    assert_eq!(index.search("secret", 10)?.len(), 1);
+    options.excluded_directory_names.push("private".into());
+    assert_eq!(
+        index
+            .index_folder_with_options(root.path(), &options)?
+            .removed,
+        1
+    );
+    Ok(())
+}
+
+#[test]
+fn filters_apply_before_limit_and_respect_folder_boundaries() -> Result<()> {
+    let root = tempfile::tempdir()?;
+    let docs = root.path().join("docs");
+    let sibling = root.path().join("docs-more");
+    fs::create_dir(&docs)?;
+    fs::create_dir(&sibling)?;
+    fs::write(docs.join("shared.md"), "shared short")?;
+    fs::write(docs.join("shared.txt"), "shared considerably longer")?;
+    fs::write(sibling.join("shared.md"), "shared sibling")?;
+    let mut index = SearchIndex::open(":memory:")?;
+    index.index_folder(root.path())?;
+    let options = SearchOptions {
+        folder: Some(docs),
+        extension: Some(".MD".into()),
+        max_size: Some(15),
+        ..Default::default()
+    };
+    let results = index.search_with_options("shared", 1, &options)?;
+    assert_eq!(results.len(), 1);
+    assert_eq!(results[0].filename, "shared.md");
+    assert!(!results[0].path.contains("docs-more"));
+    let options = SearchOptions {
+        min_size: Some(20),
+        ..Default::default()
+    };
+    assert_eq!(index.search_with_options("shared", 10, &options)?.len(), 1);
+    let options = SearchOptions {
+        modified_before_unix_ms: Some(0),
+        ..Default::default()
+    };
+    assert!(index
+        .search_with_options("shared", 10, &options)?
+        .is_empty());
+    let options = SearchOptions {
+        min_size: Some(20),
+        max_size: Some(10),
+        ..Default::default()
+    };
+    assert!(index.search_with_options("shared", 10, &options).is_err());
+    Ok(())
+}
+
+#[test]
+fn migrates_v1_database_without_losing_search_records() -> Result<()> {
+    let root = tempfile::tempdir()?;
+    let db = root.path().join("legacy.db");
+    let connection = rusqlite::Connection::open(&db)?;
+    connection.execute_batch(
+        "CREATE TABLE files(id INTEGER PRIMARY KEY, path TEXT NOT NULL UNIQUE,
+        root TEXT NOT NULL, filename TEXT NOT NULL, content TEXT NOT NULL);
+        PRAGMA user_version=1;",
+    )?;
+    drop(connection);
+    let mut index = SearchIndex::open(&db)?;
+    fs::write(root.path().join("sample.txt"), "migration evidence")?;
+    index.index_folder(root.path())?;
+    assert_eq!(index.search("evidence", 10)?.len(), 1);
+    Ok(())
+}
+
+#[test]
+fn readers_open_and_search_during_uncommitted_writer() -> Result<()> {
+    let root = tempfile::tempdir()?;
+    let db = root.path().join("index.db");
+    fs::write(root.path().join("notes.txt"), "committed evidence")?;
+    let mut index = SearchIndex::open(&db)?;
+    index.index_folder(root.path())?;
+    let writer = rusqlite::Connection::open(&db)?;
+    writer.execute_batch("BEGIN IMMEDIATE; UPDATE files SET content='uncommitted replacement';")?;
+    // Opening a current schema must not issue DDL or acquire a write lock.
+    let reader = SearchIndex::open(&db)?;
+    assert_eq!(reader.search("evidence", 10)?.len(), 1);
+    assert!(reader.search("replacement", 10)?.is_empty());
+    writer.execute_batch("ROLLBACK;")?;
+    Ok(())
+}