mnemo-engine 0.4.0

//! Smoke tests for the `mnemo` CLI exploration commands (`get`, `list`,
//! `recall`).
//!
//! These tests build a tiny database with the library, then invoke the
//! freshly built `mnemo` binary via [`std::process::Command`], asserting on
//! exit code and standard-output substrings. They are deliberately light:
//! the library API is exercised by `tests/integration.rs`; this file only
//! verifies that the CLI wiring around `db.get` / `db.memories` / `db.recall`
//! parses arguments and renders output as expected.

use std::process::Command;

use mnemo::{KdfParams, Memory, MemoryType, Mnemo, MnemoConfig};
use tempfile::tempdir;

/// Path to the `mnemo` binary built by Cargo for these tests.
fn bin() -> &'static str {
    env!("CARGO_BIN_EXE_mnemo")
}

/// Cheap KDF so the tests are fast.
fn fast_cfg(dimensions: usize) -> MnemoConfig {
    MnemoConfig {
        dimensions,
        kdf: KdfParams::fast(),
        ..Default::default()
    }
}

/// Create a tiny test database with three memories and return (path,
/// passphrase, id-of-the-semantic-memory).
fn seed_db() -> (tempfile::TempDir, std::path::PathBuf, &'static str, String) {
    let dir = tempdir().unwrap();
    let path = dir.path().join("smoke.mnemo");
    let pp = "smoke-pw";

    let mut db = Mnemo::create(&path, pp, fast_cfg(4)).unwrap();
    let semantic_id = db
        .remember(
            Memory::new(
                "the user prefers tea",
                MemoryType::Semantic,
                vec![1.0, 0.0, 0.0, 0.0],
            )
            .with_agent("assistant")
            .with_importance(0.9),
        )
        .unwrap();
    db.remember(
        Memory::new(
            "ran the deploy script",
            MemoryType::Episodic,
            vec![0.0, 1.0, 0.0, 0.0],
        )
        .with_agent("assistant")
        .with_importance(0.4),
    )
    .unwrap();
    db.remember(
        Memory::new(
            "escalate via pager",
            MemoryType::Procedural,
            vec![0.0, 0.0, 1.0, 0.0],
        )
        .with_agent("ops")
        .with_importance(0.6),
    )
    .unwrap();
    db.flush().unwrap();
    db.close().unwrap();

    (dir, path, pp, semantic_id.to_string())
}

/// Run the CLI with `MNEMO_PASSPHRASE` and return (status, stdout, stderr).
fn run_cli(args: &[&str], pp: &str) -> (std::process::ExitStatus, String, String) {
    let out = Command::new(bin())
        .args(args)
        .env("MNEMO_PASSPHRASE", pp)
        .output()
        .expect("failed to invoke mnemo binary");
    let stdout = String::from_utf8_lossy(&out.stdout).into_owned();
    let stderr = String::from_utf8_lossy(&out.stderr).into_owned();
    (out.status, stdout, stderr)
}

#[test]
fn list_table_shows_three_memories() {
    let (_dir, path, pp, _) = seed_db();
    let (status, stdout, stderr) =
        run_cli(&["list", path.to_str().unwrap(), "--format", "table"], pp);
    assert!(status.success(), "list failed: {stderr}");
    assert!(
        stdout.contains("showing 3 of 3 memories"),
        "header missing: {stdout}"
    );
    assert!(
        stdout.contains("the user prefers tea"),
        "content missing: {stdout}"
    );
    assert!(stdout.contains("[semantic]"), "type tag missing: {stdout}");
}

#[test]
fn list_filters_by_type_and_agent() {
    let (_dir, path, pp, _) = seed_db();
    let (status, stdout, _) = run_cli(
        &[
            "list",
            path.to_str().unwrap(),
            "--type",
            "procedural",
            "--agent",
            "ops",
        ],
        pp,
    );
    assert!(status.success());
    assert!(stdout.contains("escalate via pager"));
    assert!(!stdout.contains("prefers tea"));
    assert!(!stdout.contains("deploy script"));
}

#[test]
fn list_json_emits_total_and_count() {
    let (_dir, path, pp, _) = seed_db();
    let (status, stdout, _) = run_cli(
        &[
            "list",
            path.to_str().unwrap(),
            "--format",
            "json",
            "--limit",
            "2",
        ],
        pp,
    );
    assert!(status.success());
    let v: serde_json::Value = serde_json::from_str(&stdout).expect("json output");
    assert_eq!(v["total"], 3);
    assert_eq!(v["count"], 2);
    let memories = v["memories"].as_array().expect("memories array");
    assert_eq!(memories.len(), 2);
    // Vector is stripped by default.
    assert!(
        memories[0].get("vector").is_none(),
        "vector should be omitted"
    );
}

#[test]
fn list_jsonl_with_vector_includes_embedding() {
    let (_dir, path, pp, _) = seed_db();
    let (status, stdout, _) = run_cli(
        &[
            "list",
            path.to_str().unwrap(),
            "--format",
            "jsonl",
            "--vector",
        ],
        pp,
    );
    assert!(status.success());
    let lines: Vec<&str> = stdout.lines().filter(|l| !l.is_empty()).collect();
    assert_eq!(lines.len(), 3);
    for line in &lines {
        let v: serde_json::Value = serde_json::from_str(line).expect("jsonl line");
        assert!(v["vector"].as_array().is_some(), "missing vector: {line}");
    }
}

#[test]
fn get_by_id_table_and_json() {
    let (_dir, path, pp, semantic_id) = seed_db();

    let (status, stdout, _) = run_cli(
        &["get", path.to_str().unwrap(), &semantic_id, "--verbose"],
        pp,
    );
    assert!(status.success());
    assert!(
        stdout.contains(&semantic_id),
        "id missing in table output: {stdout}"
    );
    assert!(stdout.contains("the user prefers tea"));
    assert!(stdout.contains("imp=0.90"));
    assert!(stdout.contains("scope"));

    let (status, stdout, _) = run_cli(
        &[
            "get",
            path.to_str().unwrap(),
            &semantic_id,
            "--format",
            "json",
        ],
        pp,
    );
    assert!(status.success());
    let v: serde_json::Value = serde_json::from_str(&stdout).expect("json from get");
    assert_eq!(v["content"], "the user prefers tea");
    assert_eq!(v["memory_type"], "Semantic");
    assert!(
        v.get("vector").is_none(),
        "vector should be omitted by default"
    );
}

#[test]
fn get_missing_id_fails_cleanly() {
    let (_dir, path, pp, _) = seed_db();
    // A valid ULID that isn't in the DB.
    let (status, _stdout, stderr) = run_cli(
        &["get", path.to_str().unwrap(), "01ARZ3NDEKTSV4RRFFQ69G5FAV"],
        pp,
    );
    assert!(!status.success(), "missing id should exit non-zero");
    assert!(
        stderr.contains("not found"),
        "expected NotFound message, got: {stderr}"
    );
}

#[test]
fn recall_returns_query_nearest_first() {
    let (_dir, path, pp, _) = seed_db();
    let (status, stdout, _) = run_cli(
        &[
            "recall",
            path.to_str().unwrap(),
            "--query",
            "0.9,0.1,0.0,0.0",
            "--top-k",
            "2",
        ],
        pp,
    );
    assert!(status.success());
    let first_line = stdout.lines().next().unwrap_or("");
    assert!(
        first_line.contains("the user prefers tea"),
        "expected tea first: {stdout}"
    );
    assert!(first_line.starts_with("score="));
}

#[test]
fn recall_json_carries_score_and_similarity() {
    let (_dir, path, pp, _) = seed_db();
    let (status, stdout, _) = run_cli(
        &[
            "recall",
            path.to_str().unwrap(),
            "--query",
            "0.9,0.1,0.0,0.0",
            "--top-k",
            "1",
            "--format",
            "json",
        ],
        pp,
    );
    assert!(status.success());
    let v: serde_json::Value = serde_json::from_str(&stdout).expect("json from recall");
    assert_eq!(v["count"], 1);
    let hits = v["hits"].as_array().expect("hits array");
    assert!(hits[0]["score"].is_number());
    assert!(hits[0]["similarity"].is_number());
    assert_eq!(hits[0]["memory"]["content"], "the user prefers tea");
}

#[test]
fn recall_filters_by_type_excludes_other_kinds() {
    let (_dir, path, pp, _) = seed_db();
    let (status, stdout, _) = run_cli(
        &[
            "recall",
            path.to_str().unwrap(),
            "--query",
            "1.0,0.0,0.0,0.0",
            "--type",
            "procedural",
        ],
        pp,
    );
    assert!(status.success());
    // Only the procedural memory should appear.
    assert!(stdout.contains("escalate via pager"));
    assert!(!stdout.contains("prefers tea"));
    assert!(!stdout.contains("deploy script"));
}

#[test]
fn recall_query_file_json_works() {
    // --query-file accepts a JSON array of floats from a file path, as an
    // alternative to inline --query (which is unusable at high dims).
    let (dir, path, pp, _) = seed_db();
    let qf = dir.path().join("q.json");
    std::fs::write(&qf, "[1.0, 0.0, 0.0, 0.0]").unwrap();
    let (status, stdout, stderr) = run_cli(
        &[
            "recall",
            path.to_str().unwrap(),
            "--query-file",
            qf.to_str().unwrap(),
            "--top-k",
            "1",
        ],
        pp,
    );
    assert!(status.success(), "recall failed: {stderr}");
    assert!(stdout.contains("the user prefers tea"), "got: {stdout}");
}

#[test]
fn recall_query_file_whitespace_works() {
    // --query-file also accepts whitespace/newline-separated floats — handy
    // for piping from awk/jq without JSON serialization.
    let (dir, path, pp, _) = seed_db();
    let qf = dir.path().join("q.txt");
    std::fs::write(&qf, "1.0\n0.0\n0.0\n0.0\n").unwrap();
    let (status, stdout, stderr) = run_cli(
        &[
            "recall",
            path.to_str().unwrap(),
            "--query-file",
            qf.to_str().unwrap(),
            "--top-k",
            "1",
        ],
        pp,
    );
    assert!(status.success(), "recall failed: {stderr}");
    assert!(stdout.contains("the user prefers tea"), "got: {stdout}");
}

#[test]
fn recall_requires_query_source() {
    // Neither --query nor --query-file should fail loudly rather than silently
    // running with no query.
    let (_dir, path, pp, _) = seed_db();
    let (status, _stdout, stderr) =
        run_cli(&["recall", path.to_str().unwrap(), "--top-k", "1"], pp);
    assert!(!status.success(), "recall without query should fail");
    assert!(
        stderr.contains("--query") || stderr.contains("query"),
        "stderr should mention missing query, got: {stderr}"
    );
}

// === MCP stdio server ====================================================

/// Drive the MCP server end-to-end: spawn `mnemo serve --mcp <path>`,
/// pipe a scripted JSON-RPC session through stdin, capture stdout, and
/// assert the responses. Uses one long-lived child process rather than
/// one per request so we exercise the actual line-oriented dispatch loop.
#[test]
fn mcp_serve_initialize_list_and_call_tools() {
    use std::io::Write;
    use std::process::{Command, Stdio};

    // Seed the file with a *scoped* writer so its exclusive lock is
    // released before the server subprocess tries to acquire its own.
    // The temp dir stays alive for the whole test so the file survives.
    let dir = tempdir().unwrap();
    let path = dir.path().join("mcp.mnemo");
    let pp = "smoke-pw";
    {
        let mut db = Mnemo::create(&path, pp, fast_cfg(4)).unwrap();
        db.remember(
            Memory::new(
                "the user prefers tea",
                MemoryType::Semantic,
                vec![1.0, 0.0, 0.0, 0.0],
            )
            .with_importance(0.9),
        )
        .unwrap();
        db.flush().unwrap();
    }

    // ---- Race note (Linux only, observed on Rust 1.75 MSRV CI, 2026-08-29) ----
    // cargo runs `integration.rs` and `cli_smoke.rs` in parallel; several
    // integration tests fork subprocesses of their own via `Command::spawn`.
    // On Rust 1.75 Linux, `Command::spawn` still uses fork+exec, which
    // briefly clones ALL parent FDs into the sibling child before exec's
    // CLOEXEC pass drops them. If a sibling test forks while our own
    // scoped seed writer above is mid-drop, an inherited copy of the
    // writer's FD can keep the flock on `mcp.mnemo` alive for a few
    // milliseconds after our writer's `db` has dropped. Our `mnemo serve
    // --mcp` subprocess then loses the race for `try_lock_exclusive` and
    // exits with `MnemoError::Locked`. Stable Rust switched `spawn` to
    // `posix_spawn`, which skips the FD-clone window, so this reproduces
    // only on the MSRV Linux job.
    //
    // Fix scope: **test-only retry**. We re-spawn the whole session up
    // to 5 times with 100/200/400/800 ms backoff, but ONLY when the
    // child stderr carries the specific `file already locked` marker.
    // Any other non-zero exit falls through to the normal assertion
    // below with the same panic message it used to have. Do not weaken
    // the engine — the exclusive lock is a correctness invariant for
    // Phase 1.1; the race is inherent to fork+exec-plus-parallel-tests
    // on 1.75 and cannot recur once the process is stable.
    // -------------------------------------------------------------------------
    let mut out_opt = None;
    let mut last_stderr = String::new();
    for attempt in 0..5u32 {
        if attempt > 0 {
            // 100 ms → 200 → 400 → 800; cumulative worst case ~1.5 s.
            let backoff_ms = 100u64 << (attempt - 1);
            std::thread::sleep(std::time::Duration::from_millis(backoff_ms));
        }
        let mut child = Command::new(bin())
            .args(["serve", path.to_str().unwrap(), "--mcp"])
            .env("MNEMO_PASSPHRASE", pp)
            .stdin(Stdio::piped())
            .stdout(Stdio::piped())
            .stderr(Stdio::piped())
            .spawn()
            .expect("failed to spawn mnemo serve --mcp");

        // Scripted client session: initialize, then list tools, then call
        // `stats`. Each line is one JSON-RPC message. Closing stdin ends
        // the loop and the child exits.
        let stdin = child.stdin.as_mut().expect("child stdin missing");
        let script = [
            r#"{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{}}}"#,
            r#"{"jsonrpc":"2.0","method":"notifications/initialized","params":{}}"#,
            r#"{"jsonrpc":"2.0","id":2,"method":"tools/list","params":{}}"#,
            r#"{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"stats","arguments":{}}}"#,
        ];
        for line in script {
            writeln!(stdin, "{line}").unwrap();
        }
        // Close stdin — signals EOF to the server's read loop.
        drop(child.stdin.take());

        // Wait for shutdown (bounded — the server should exit as soon as
        // stdin closes; hanging here would fail the test with a timeout
        // at the test-runner level).
        let run = child.wait_with_output().expect("child wait failed");
        if !run.status.success() {
            let stderr_snapshot = String::from_utf8_lossy(&run.stderr).into_owned();
            if stderr_snapshot.contains("file already locked") {
                // Race lost — retry after backoff.
                last_stderr = stderr_snapshot;
                continue;
            }
            // Any other failure: bubble up unchanged.
            out_opt = Some(run);
            break;
        }
        out_opt = Some(run);
        break;
    }
    let out = out_opt.unwrap_or_else(|| {
        panic!(
            "mnemo serve --mcp lost the fork/flock race on all 5 attempts \
             (bounded retry exhausted); last stderr:\n{last_stderr}"
        )
    });
    let stdout = String::from_utf8_lossy(&out.stdout).into_owned();
    let stderr = String::from_utf8_lossy(&out.stderr).into_owned();
    assert!(
        out.status.success(),
        "server exited non-zero: status={:?} stderr={stderr}",
        out.status
    );

    // Parse the response lines back and assert on structure.
    let lines: Vec<&str> = stdout.lines().filter(|l| !l.trim().is_empty()).collect();
    assert!(
        lines.len() >= 3,
        "expected ≥3 response lines (init, tools/list, stats), got {}: {stdout}",
        lines.len()
    );

    let init: serde_json::Value = serde_json::from_str(lines[0]).expect("init parse");
    assert_eq!(init["jsonrpc"], "2.0");
    assert_eq!(init["id"], 1);
    assert_eq!(init["result"]["protocolVersion"], "2025-06-18");
    assert_eq!(init["result"]["serverInfo"]["name"], "mnemo");

    let list: serde_json::Value = serde_json::from_str(lines[1]).expect("tools/list parse");
    let tools = list["result"]["tools"].as_array().expect("tools array");
    let names: Vec<&str> = tools.iter().filter_map(|t| t["name"].as_str()).collect();
    for expected in [
        "about",
        "remember",
        "recall",
        "forget",
        "list",
        "snapshot_list",
        "stats",
    ] {
        assert!(
            names.contains(&expected),
            "missing tool '{expected}' in {names:?}"
        );
    }

    let stats: serde_json::Value = serde_json::from_str(lines[2]).expect("stats parse");
    let content = &stats["result"]["content"][0]["text"];
    assert!(
        content.is_string(),
        "stats content should be a text block: {stats}"
    );
    let stats_body: serde_json::Value =
        serde_json::from_str(content.as_str().unwrap()).expect("stats body parse");
    assert_eq!(
        stats_body["memories"], 1,
        "seeded exactly one live memory: {stats_body}"
    );
}

#[test]
fn mcp_serve_refuses_without_passphrase() {
    // Env var missing → server errors at startup with a clear message
    // and non-zero exit, never opening the file.
    use std::process::{Command, Stdio};

    let dir = tempdir().unwrap();
    let path = dir.path().join("mcp-nopw.mnemo");
    {
        let mut db = Mnemo::create(&path, "smoke-pw", fast_cfg(4)).unwrap();
        db.remember(Memory::new(
            "x",
            MemoryType::Semantic,
            vec![1.0, 0.0, 0.0, 0.0],
        ))
        .unwrap();
        db.flush().unwrap();
    }

    let out = Command::new(bin())
        .args(["serve", path.to_str().unwrap(), "--mcp"])
        .env_remove("MNEMO_PASSPHRASE")
        .stdin(Stdio::null())
        .stdout(Stdio::piped())
        .stderr(Stdio::piped())
        .output()
        .expect("spawn failed");
    let stderr = String::from_utf8_lossy(&out.stderr).into_owned();
    assert!(
        !out.status.success(),
        "expected non-zero exit without passphrase"
    );
    assert!(
        stderr.contains("MNEMO_PASSPHRASE"),
        "stderr should mention the env var, got: {stderr}"
    );
}