rqmd 0.1.1

//! Shared harness for the E2E CLI tests (`tests/cli.rs`).
//!
//! Rust port of the `runQmd()` helper + `beforeAll` fixtures in qmd's
//! `test/cli.test.ts`. Each test gets its own [`Env`] (a fresh `TempDir` with
//! the same markdown fixtures qmd creates) and spawns the built `rqmd` binary
//! with fully isolated state:
//!
//! * `RQMD_INDEX_PATH`  — points the SQLite index at the temp dir
//!   (honoured before the production-mode gate, `store/path.rs`).
//! * `RQMD_CONFIG_DIR`  — points the YAML config (`index.yml`) at the temp dir.
//! * `PWD`              — rqmd's `pwd()` prefers this over `current_dir()`.
//! * `--index index`    — prepended so `IndexState::new` skips the `.rqmd/`
//!   local-config walk up the ancestors of the cwd (which, on Windows, live
//!   under the user profile alongside the OS temp dir). This keeps every run
//!   hermetic regardless of the host's `~/.rqmd` / `~/.config/rqmd`.
//! * `NO_COLOR=1`       — strip ANSI so substring assertions are clean.
//! * `CI=1`             — make any accidental LLM/model path fail fast instead
//!   of hitting the network (none of the ported commands need a model).

#![allow(dead_code)]

use std::net::TcpListener;
use std::path::{Path, PathBuf};
use std::process::{Child, Command, Stdio};
use std::time::{Duration, Instant};

use assert_cmd::cargo::CommandCargoExt;

/// Result of one `rqmd` invocation.
pub struct Out {
    pub stdout: String,
    pub stderr: String,
    pub code: i32,
}

impl Out {
    /// Assert exit code 0, surfacing stderr on failure.
    pub fn assert_ok(&self) -> &Self {
        assert_eq!(
            self.code, 0,
            "expected exit 0\n--- stdout ---\n{}\n--- stderr ---\n{}",
            self.stdout, self.stderr
        );
        self
    }

    /// Assert a non-zero exit code (clap parse errors use 2; app errors use 1).
    pub fn assert_err(&self) -> &Self {
        assert_ne!(
            self.code, 0,
            "expected non-zero exit\n--- stdout ---\n{}\n--- stderr ---\n{}",
            self.stdout, self.stderr
        );
        self
    }

    /// Assert exit code 1 (an application error routed through `main`'s
    /// `eprintln!("error: …"); exit(1)`).
    pub fn assert_code(&self, code: i32) -> &Self {
        assert_eq!(
            self.code, code,
            "expected exit {code}\n--- stdout ---\n{}\n--- stderr ---\n{}",
            self.stdout, self.stderr
        );
        self
    }
}

/// One isolated test environment: a temp dir holding the markdown fixtures, a
/// `config/` dir for `index.yml`, and a default index db path.
pub struct Env {
    /// Kept alive for the duration of the test; dropping it removes the dir.
    _root: tempfile::TempDir,
    pub root: PathBuf,
    pub fixtures: PathBuf,
    pub config_dir: PathBuf,
    pub db: PathBuf,
}

impl Env {
    /// Spawn `rqmd <args>` from the fixtures dir with the default isolated db
    /// + config. `--index index` is prepended (see module docs).
    pub fn run(&self, args: &[&str]) -> Out {
        self.run_in(&self.fixtures, args)
    }

    /// Like [`run`](Self::run) but from an explicit working directory.
    pub fn run_in(&self, cwd: &Path, args: &[&str]) -> Out {
        let mut full: Vec<&str> = vec!["--index", "index"];
        full.extend_from_slice(args);
        spawn(cwd, &self.db, &self.config_dir, &full, &[])
    }

    /// Run from an explicit cwd with extra env overrides and *no* `--index`
    /// prepend. Used by `rqmd init` tests, which are index-agnostic and need to
    /// control `HOME`/`USERPROFILE` to exercise the `$HOME` guard.
    pub fn run_in_env(&self, cwd: &Path, args: &[&str], extra: &[(&str, &str)]) -> Out {
        spawn(cwd, &self.db, &self.config_dir, args, extra)
    }

    /// Like [`run`](Self::run) but with extra environment overrides applied
    /// last (so they win over the defaults — used to point at a custom
    /// `RQMD_INDEX_PATH`).
    pub fn run_env(&self, args: &[&str], extra: &[(&str, &str)]) -> Out {
        let mut full: Vec<&str> = vec!["--index", "index"];
        full.extend_from_slice(args);
        spawn(&self.fixtures, &self.db, &self.config_dir, &full, extra)
    }

    /// Spawn `rqmd <args>` verbatim (no `--index` prepend). Used for the
    /// help / no-args parser tests that don't touch the index.
    pub fn run_bare(&self, args: &[&str]) -> Out {
        spawn(&self.fixtures, &self.db, &self.config_dir, args, &[])
    }

    /// Overwrite `<config_dir>/index.yml` (the global `beforeEach` reset, and
    /// the ignore-pattern tests that hand-write a collections config).
    pub fn write_config(&self, yaml: &str) {
        std::fs::write(self.config_dir.join("index.yml"), yaml).expect("write index.yml");
    }

    /// Forward-slash form of an absolute path under `root`, safe to embed in a
    /// double-quoted YAML scalar on Windows (`C:/Users/...`).
    pub fn yaml_path(&self, p: &Path) -> String {
        p.to_string_lossy().replace('\\', "/")
    }
}

/// Create a fresh isolated [`Env`], mirroring qmd's `beforeAll` fixtures.
/// The fixtures subdir is literally named `fixtures` so `collection add .`
/// derives the collection name `fixtures` (qmd relies on the same basename).
pub fn env() -> Env {
    let root = tempfile::tempdir().expect("mkdtemp");
    let root_path = root.path().to_path_buf();
    let fixtures = root_path.join("fixtures");
    let config_dir = root_path.join("config");
    let db = root_path.join("index.sqlite");

    std::fs::create_dir_all(fixtures.join("notes")).unwrap();
    std::fs::create_dir_all(fixtures.join("docs")).unwrap();
    std::fs::create_dir_all(&config_dir).unwrap();
    std::fs::write(config_dir.join("index.yml"), "collections: {}\n").unwrap();

    write_fixtures(&fixtures);

    Env {
        _root: root,
        root: root_path,
        fixtures,
        config_dir,
        db,
    }
}

fn write_fixtures(fixtures: &Path) {
    std::fs::write(
        fixtures.join("README.md"),
        "# Test Project\n\n\
         This is a test project for QMD CLI testing.\n\n\
         ## Features\n\n\
         - Full-text search with BM25\n\
         - Vector similarity search\n\
         - Hybrid search with reranking\n",
    )
    .unwrap();

    std::fs::write(
        fixtures.join("notes").join("meeting.md"),
        "# Team Meeting Notes\n\n\
         Date: 2024-01-15\n\n\
         ## Attendees\n\
         - Alice\n\
         - Bob\n\
         - Charlie\n\n\
         ## Discussion Topics\n\
         - Project timeline review\n\
         - Resource allocation\n\
         - Technical debt prioritization\n\n\
         ## Action Items\n\
         1. Alice to update documentation\n\
         2. Bob to fix authentication bug\n\
         3. Charlie to review pull requests\n",
    )
    .unwrap();

    std::fs::write(
        fixtures.join("notes").join("ideas.md"),
        "# Product Ideas\n\n\
         ## Feature Requests\n\
         - Dark mode support\n\
         - Keyboard shortcuts\n\
         - Export to PDF\n\n\
         ## Technical Improvements\n\
         - Improve search performance\n\
         - Add caching layer\n\
         - Optimize database queries\n",
    )
    .unwrap();

    std::fs::write(
        fixtures.join("docs").join("api.md"),
        "# API Documentation\n\n\
         ## Endpoints\n\n\
         ### GET /search\n\
         Search for documents.\n\n\
         Parameters:\n\
         - q: Search query (required)\n\
         - limit: Max results (default: 10)\n\n\
         ### GET /document/:id\n\
         Retrieve a specific document.\n\n\
         ### POST /index\n\
         Index new documents.\n",
    )
    .unwrap();

    std::fs::write(
        fixtures.join("test1.md"),
        "# Test Document 1\n\n\
         This is the first test document.\n\n\
         It has multiple lines for testing line numbers.\n\
         Line 6 is here.\n\
         Line 7 is here.\n",
    )
    .unwrap();

    std::fs::write(
        fixtures.join("test2.md"),
        "# Test Document 2\n\n\
         This is the second test document.\n",
    )
    .unwrap();
}

fn spawn(cwd: &Path, db: &Path, cfg: &Path, args: &[&str], extra: &[(&str, &str)]) -> Out {
    let mut cmd = Command::cargo_bin("rqmd").expect("rqmd binary is built by cargo test");
    cmd.current_dir(cwd)
        .env_remove("XDG_CACHE_HOME")
        .env_remove("XDG_CONFIG_HOME")
        .env_remove("RQMD_CACHE_DIR")
        .env("NO_COLOR", "1")
        .env("CI", "1")
        .env("PWD", cwd)
        .env("RQMD_INDEX_PATH", db)
        .env("RQMD_CONFIG_DIR", cfg)
        .args(args);
    for (k, v) in extra {
        cmd.env(k, v);
    }
    let out = cmd.output().expect("spawn rqmd");
    Out {
        stdout: String::from_utf8_lossy(&out.stdout).into_owned(),
        stderr: String::from_utf8_lossy(&out.stderr).into_owned(),
        code: out.status.code().unwrap_or(-1),
    }
}

/// Spawn `rqmd` against a named-index *cache directory* instead of a single
/// pinned `RQMD_INDEX_PATH`, so `--index <name>` resolves to
/// `<cache>/<name>.sqlite` (and a link's `?index=<name>` does the same). This
/// mirrors qmd's custom-index test env (`XDG_CACHE_HOME` set, `INDEX_PATH`
/// empty). `--index` is NOT prepended — pass it explicitly, or omit it (e.g.
/// for the `get` round-trip that relies on the link's `?index=`). The same
/// `cache` dir must be reused across calls so each index DB persists.
pub fn spawn_cache(cwd: &Path, cache: &Path, cfg: &Path, args: &[&str]) -> Out {
    let mut cmd = Command::cargo_bin("rqmd").expect("rqmd binary is built by cargo test");
    cmd.current_dir(cwd)
        .env_remove("XDG_CACHE_HOME")
        .env_remove("XDG_CONFIG_HOME")
        .env_remove("RQMD_INDEX_PATH")
        .env("NO_COLOR", "1")
        .env("CI", "1")
        .env("PWD", cwd)
        .env("RQMD_CACHE_DIR", cache)
        .env("RQMD_CONFIG_DIR", cfg)
        .args(args);
    let out = cmd.output().expect("spawn rqmd");
    Out {
        stdout: String::from_utf8_lossy(&out.stdout).into_owned(),
        stderr: String::from_utf8_lossy(&out.stderr).into_owned(),
        code: out.status.code().unwrap_or(-1),
    }
}

// ---------------------------------------------------------------------------
// Assertion helpers
// ---------------------------------------------------------------------------

/// True if `s` is exactly 6 lowercase-hex characters (rqmd's docid shape).
pub fn is_hex6(s: &str) -> bool {
    s.len() == 6
        && s.bytes()
            .all(|b| b.is_ascii_hexdigit() && !b.is_ascii_uppercase())
}

/// First non-empty line of `s` (the first data row for csv/files output).
pub fn first_line(s: &str) -> &str {
    s.lines().find(|l| !l.trim().is_empty()).unwrap_or("")
}

// ---------------------------------------------------------------------------
// MCP HTTP server harness (`mcp --http` / `--daemon`)
// ---------------------------------------------------------------------------
//
// These helpers spawn the `rqmd` binary as a *long-running* child (the existing
// `spawn`/`spawn_cache` use `.output()`, which blocks until exit and is useless
// for a server). They use the `spawn_cache` env contract — `RQMD_CACHE_DIR` set,
// `RQMD_INDEX_PATH` + `XDG_CACHE_HOME` removed — because `default_db_path`
// short-circuits on `RQMD_INDEX_PATH` *before* the `--index <name>` path is
// computed, so a pinned index path would make `--index` a no-op. The same
// `cache` dir must back the seeding (`collection add`) and the server so they
// open the same `<cache>/<name>.sqlite`.

/// A spawned long-running `rqmd` server child plus the port it was told to bind.
/// Killed (and reaped) on drop so a failing test never leaks the process.
pub struct ServerChild {
    pub child: Child,
    pub port: u16,
}

impl Drop for ServerChild {
    fn drop(&mut self) {
        let _ = self.child.kill();
        let _ = self.child.wait();
    }
}

/// Best-effort teardown for `mcp --daemon` tests: on drop, runs `rqmd mcp stop`
/// (which kills the detached daemon child and removes its PID file). The daemon
/// is a *detached* process not tracked by any `Child` handle, so a child-kill
/// guard cannot reach it — `mcp stop` (reading the PID file) is the only handle.
/// Mirrors qmd's `afterAll` PID-file cleanup (cli.test.ts:1630-1646).
pub struct DaemonGuard {
    pub cwd: PathBuf,
    pub cache: PathBuf,
    pub cfg: PathBuf,
}

impl Drop for DaemonGuard {
    fn drop(&mut self) {
        let _ = spawn_cache(&self.cwd, &self.cache, &self.cfg, &["mcp", "stop"]);
    }
}

/// Pick an ephemeral free TCP port by binding `:0` and immediately releasing it.
/// (Small TOCTOU window before the server rebinds, acceptable for tests; mirrors
/// qmd's random-port approach.)
pub fn free_port() -> u16 {
    TcpListener::bind("127.0.0.1:0")
        .expect("bind ephemeral port")
        .local_addr()
        .expect("local_addr")
        .port()
}

/// Spawn `rqmd <args>` as a detached long-running child under the `spawn_cache`
/// env (see module note). stdio is nulled. The caller polls for readiness.
pub fn spawn_cache_child(cwd: &Path, cache: &Path, cfg: &Path, args: &[&str]) -> Child {
    let mut cmd = Command::cargo_bin("rqmd").expect("rqmd binary is built by cargo test");
    cmd.current_dir(cwd)
        .env_remove("XDG_CACHE_HOME")
        .env_remove("XDG_CONFIG_HOME")
        .env_remove("RQMD_INDEX_PATH")
        .env("NO_COLOR", "1")
        .env("CI", "1")
        .env("PWD", cwd)
        .env("RQMD_CACHE_DIR", cache)
        .env("RQMD_CONFIG_DIR", cfg)
        .stdin(Stdio::null())
        .stdout(Stdio::null())
        .stderr(Stdio::null())
        .args(args);
    cmd.spawn().expect("spawn rqmd server")
}

/// Poll `http://127.0.0.1:<port>/health` until it returns 200 (parsed JSON
/// body) or the deadline elapses. `None` on timeout.
pub fn wait_for_health(port: u16, timeout: Duration) -> Option<serde_json::Value> {
    let url = format!("http://127.0.0.1:{port}/health");
    let deadline = Instant::now() + timeout;
    while Instant::now() < deadline {
        match ureq::get(&url).call() {
            Ok(resp) if resp.status() == 200 => {
                if let Ok(v) = resp.into_json::<serde_json::Value>() {
                    return Some(v);
                }
            }
            _ => {}
        }
        std::thread::sleep(Duration::from_millis(150));
    }
    None
}

/// POST a JSON body to `http://127.0.0.1:<port>/query` and return the parsed
/// response. Panics on transport error so tests fail loudly.
pub fn post_query(port: u16, body: serde_json::Value) -> serde_json::Value {
    let url = format!("http://127.0.0.1:{port}/query");
    ureq::post(&url)
        .send_json(body)
        .expect("POST /query")
        .into_json::<serde_json::Value>()
        .expect("query response json")
}