boxmaker/docs/plans/M5a/files/crates/memoryd/tests/index.rs

//! The index: what is indexed from a session log and `notes.md`, each record's label, catching up,
//! `max_class`, and a rebuild equal to indexing as it went (M5a spec, section 6). Do not edit.

#[path = "support/logs.rs"]
mod logs;
#[path = "support/tmp.rs"]
mod tmp;

use logs::{append, assistant, result, user, write_log};
use memoryd::index::{Index, rebuild};
use proto::DataClass;
use tmp::TempDir;

/// Every indexed item, in order: (kind, role, text, class, untrusted, record).
fn items(index: &Index) -> Vec<(String, String, String, String, bool, i64)> {
    let mut stmt = index
        .conn()
        .prepare("SELECT s.kind, s.role, h.text, s.class, s.untrusted, s.record FROM sources s JOIN hits h ON h.rowid = s.rowid ORDER BY s.rowid")
        .unwrap();
    stmt.query_map([], |r| {
        Ok((
            r.get(0)?,
            r.get(1)?,
            r.get(2)?,
            r.get(3)?,
            r.get(4)?,
            r.get(5)?,
        ))
    })
    .unwrap()
    .map(|r| r.unwrap())
    .collect()
}

fn item(
    kind: &str,
    role: &str,
    text: &str,
    class: &str,
    untrusted: bool,
    record: i64,
) -> (String, String, String, String, bool, i64) {
    (
        kind.to_string(),
        role.to_string(),
        text.to_string(),
        class.to_string(),
        untrusted,
        record,
    )
}

#[test]
fn owner_assistant_and_calls_are_indexed_with_the_label_in_force() {
    let dir = TempDir::new("index-log");
    let log = write_log(
        dir.path(),
        "mm-a",
        &[
            user(1, "What is the launch window?"),
            assistant(
                2,
                None,
                &[("http_fetch", r#"{"url":"https://example.org"}"#)],
            ),
            result(3, "PAGE TEXT never indexed", DataClass::Private, true),
            assistant(4, Some("The page says Thursday."), &[]),
            user(5, "   "),
        ],
    );
    let index = Index::open(&dir.path().join("index/memory.sqlite3"), DataClass::Private).unwrap();
    assert_eq!(index.index_log(&log, "mm-a", 0), Ok(5));
    assert_eq!(
        items(&index),
        [
            item(
                "session",
                "owner",
                "What is the launch window?",
                "private",
                false,
                0
            ),
            item(
                "session",
                "call",
                r#"http_fetch {"url":"https://example.org"}"#,
                "private",
                false,
                1
            ),
            item(
                "session",
                "assistant",
                "The page says Thursday.",
                "private",
                true,
                3
            ),
        ]
    );
}

#[test]
fn catching_up_reads_only_new_records_and_keeps_the_label() {
    let dir = TempDir::new("index-catch");
    let log = write_log(
        dir.path(),
        "mm-b",
        &[user(1, "first"), result(2, "x", DataClass::Private, true)],
    );
    let index = Index::open(&dir.path().join("index/memory.sqlite3"), DataClass::Private).unwrap();
    assert_eq!(index.index_log(&log, "mm-b", 0), Ok(2));
    append(&log, &user(3, "second"));
    assert_eq!(index.index_log(&log, "mm-b", 0), Ok(1));
    assert_eq!(index.index_log(&log, "mm-b", 0), Ok(0));
    let got = items(&index);
    assert_eq!(got.len(), 2);
    assert_eq!(
        got[1],
        item("session", "owner", "second", "private", true, 2),
        "tainted by the earlier result"
    );
}

#[test]
fn records_above_max_class_are_not_indexed() {
    let dir = TempDir::new("index-class");
    let log = write_log(
        dir.path(),
        "mm-c",
        &[
            user(1, "before"),
            result(2, "s", DataClass::Secret, false),
            user(3, "after the secret"),
        ],
    );
    let index = Index::open(&dir.path().join("index/memory.sqlite3"), DataClass::Private).unwrap();
    index.index_log(&log, "mm-c", 0).unwrap();
    assert_eq!(
        items(&index),
        [item("session", "owner", "before", "private", false, 0)]
    );
    let open = Index::open(&dir.path().join("index/secret.sqlite3"), DataClass::Secret).unwrap();
    open.index_log(&log, "mm-c", 0).unwrap();
    assert_eq!(
        items(&open)[1],
        item("session", "owner", "after the secret", "secret", false, 2)
    );
}

#[test]
fn notes_are_indexed_with_their_labels_and_only_once() {
    let dir = TempDir::new("index-notes");
    let index = Index::open(&dir.path().join("index/memory.sqlite3"), DataClass::Private).unwrap();
    let mut text = "§ noted 2026-09-24T20:14Z in session mm-a · private · untrusted\nexample.org says March.\n§ moved 2026-09-24T20:15Z from memory.md in session mm-b\nOwner prefers UTC.\n".to_string();
    assert_eq!(index.index_notes(&text), Ok(2));
    text.push_str("§ dropped 2026-09-24T20:16Z from user.md in session mm-b\nLives in Oslo.\n§ noted 2026-09-24T20:17Z in session mm-c · secret\nThe code.\n");
    assert_eq!(index.index_notes(&text), Ok(2));
    assert_eq!(
        items(&index),
        [
            item("noted", "", "example.org says March.", "private", true, 0),
            item(
                "moved",
                "memory.md",
                "Owner prefers UTC.",
                "private",
                false,
                1
            ),
            item("dropped", "user.md", "Lives in Oslo.", "private", false, 2),
        ]
    );
}

#[test]
fn a_rebuild_equals_the_incremental_index_and_skips_damaged_logs() {
    let dir = TempDir::new("index-rebuild");
    let home = dir.path();
    let log = write_log(home, "mm-a", &[user(1, "one")]);
    write_log(
        home,
        "mm-b",
        &[
            user(2, "two"),
            result(3, "r", DataClass::Private, true),
            assistant(4, Some("reply"), &[]),
        ],
    );
    std::fs::create_dir_all(home.join("sessions/mm-bad")).unwrap();
    std::fs::write(home.join("sessions/mm-bad/0.jsonl"), "{not json\n").unwrap();
    let notes = dir.write(
        "memory/notes.md",
        "§ noted 2026-09-24T20:14Z in session mm-a · private\nseen\n",
    );
    let path = home.join("index/memory.sqlite3");
    let index = Index::open(&path, DataClass::Private).unwrap();
    let (counts, problems) = index.catch_up(&home.join("sessions"), &notes).unwrap();
    assert_eq!((counts.logs, counts.records, counts.notes), (2, 4, 1));
    assert_eq!(problems.len(), 1);
    assert!(problems[0].contains("mm-bad"), "{problems:?}");
    append(&log, &user(9, "later"));
    index.catch_up(&home.join("sessions"), &notes).unwrap();
    let incremental = items(&index);
    drop(index);
    let (counts, _) = rebuild(&path, DataClass::Private, &home.join("sessions"), &notes).unwrap();
    assert_eq!((counts.logs, counts.records, counts.notes), (2, 5, 1));
    let rebuilt = Index::open(&path, DataClass::Private).unwrap();
    let mut a = incremental.clone();
    let mut b = items(&rebuilt);
    a.sort();
    b.sort();
    assert_eq!(a, b);
    assert!(!home.join("index/memory.rebuilding").exists());
}

#[test]
fn the_index_file_and_directory_are_private() {
    use std::os::unix::fs::PermissionsExt;
    let dir = TempDir::new("index-mode");
    let path = dir.path().join("index/memory.sqlite3");
    Index::open(&path, DataClass::Private).unwrap();
    let mode = |p: &std::path::Path| std::fs::metadata(p).unwrap().permissions().mode() & 0o777;
    assert_eq!(
        (mode(&dir.path().join("index")), mode(&path)),
        (0o700, 0o600)
    );
}