jan-cli 0.10.0

mod builtins;
mod config;
mod cron;
mod deps;
mod inputs;
mod inspect;
mod runner;
mod spec_load;
mod yaml_closure;

pub use config::{load_user_config, UserConfig};
pub use runner::run_jan;
pub use spec_load::HostPlatform;

use std::collections::BTreeMap;
use std::ffi::OsString;
use std::path::{Path, PathBuf};
use std::process::Command;

use anyhow::{bail, Context, Result};
use rusqlite::Connection;
use serde::Deserialize;
use serde::de::{self, Deserializer, Visitor};
use std::fmt;

#[derive(Debug, Deserialize)]
pub struct RootSpec {
    pub metadata: Option<Metadata>,
    #[serde(default)]
    pub commands: BTreeMap<String, CommandNode>,
}

#[derive(Debug, Deserialize)]
pub struct Metadata {
    pub name: Option<String>,
    pub description: Option<String>,
}

/// Child-process environment declaration for a command node.
///
/// Two YAML shapes are accepted:
///
/// ```yaml
/// # Legacy / shorthand — all keys are public assignments
/// env:
///   FOO: bar
///
/// # Explicit sections
/// env:
///   public:
///     FOO: bar
///   private:
///     - GH_TOKEN
/// ```
///
/// `public` values are taken from the YAML. `private` names must already exist in
/// jan's own environment; their values are copied into the child and never stored
/// in the spec. When either section is non-empty, the child runs with a cleared
/// environment containing only those variables plus a small essential allowlist
/// (PATH, HOME, …).
#[derive(Debug, Default, Clone, PartialEq, Eq)]
pub struct EnvSpec {
    pub public: BTreeMap<String, String>,
    pub private: Vec<String>,
}

impl EnvSpec {
    pub fn is_empty(&self) -> bool {
        self.public.is_empty() && self.private.is_empty()
    }

    /// True when the child should not inherit the full parent environment.
    pub fn restricts_child_env(&self) -> bool {
        !self.is_empty()
    }

    pub fn merge_from(&mut self, other: EnvSpec) {
        for (k, v) in other.public {
            self.public.insert(k, v);
        }
        for name in other.private {
            if !self.private.iter().any(|p| p == &name) {
                self.private.push(name);
            }
        }
    }
}

impl<'de> Deserialize<'de> for EnvSpec {
    fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
    where
        D: Deserializer<'de>,
    {
        #[derive(Deserialize)]
        struct Structured {
            #[serde(default)]
            public: BTreeMap<String, String>,
            #[serde(default, deserialize_with = "deserialize_string_or_seq")]
            private: Vec<String>,
        }

        #[derive(Deserialize)]
        #[serde(untagged)]
        enum EnvDe {
            Flat(BTreeMap<String, String>),
            Sections(Structured),
        }

        Ok(match EnvDe::deserialize(deserializer)? {
            EnvDe::Flat(public) => Self {
                public,
                private: Vec::new(),
            },
            EnvDe::Sections(s) => Self {
                public: s.public,
                private: s.private,
            },
        })
    }
}

#[derive(Debug, Deserialize, Default, Clone)]
pub struct CommandNode {
    /// If non-empty, this command and its subtree are only offered on these
    /// platforms (`linux`, `macos`, `windows`, …). `darwin` is accepted as an alias for `macos`.
    #[serde(default)]
    pub os: Vec<String>,
    #[serde(default)]
    pub about: String,
    /// Directory prepended to PATH when this script (or a descendant leaf) runs.
    pub path: Option<String>,
    /// Other script names whose `path` directories are prepended before this one runs.
    #[serde(default)]
    pub dependencies: Vec<String>,
    /// External binaries that must be on PATH (e.g. `fzf`, `jq`) before the leaf runs.
    #[serde(default)]
    pub requires: Vec<String>,
    /// Public assignments and/or private names required from the host environment.
    #[serde(default)]
    pub env: EnvSpec,
    /// Named CLI inputs (`--name value`) available as `${{ inputs.name }}` in env/argv.
    #[serde(default)]
    pub inputs: BTreeMap<String, crate::inputs::InputDef>,
    /// Optional crontab schedule(s). When set, `jan cron` runs this script's `run`
    /// leaf whenever the local time matches any expression.
    #[serde(default, deserialize_with = "deserialize_string_or_seq")]
    pub cron: Vec<String>,
    #[serde(default)]
    pub commands: BTreeMap<String, CommandNode>,
    pub exec: Option<ExecSpec>,
}

pub(crate) fn deserialize_string_or_seq<'de, D>(deserializer: D) -> Result<Vec<String>, D::Error>
where
    D: Deserializer<'de>,
{
    struct StringOrSeq;

    impl<'de> Visitor<'de> for StringOrSeq {
        type Value = Vec<String>;

        fn expecting(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
            formatter.write_str("a string or a sequence of strings")
        }

        fn visit_str<E>(self, value: &str) -> Result<Self::Value, E>
        where
            E: de::Error,
        {
            if value.trim().is_empty() {
                Ok(Vec::new())
            } else {
                Ok(vec![value.to_string()])
            }
        }

        fn visit_string<E>(self, value: String) -> Result<Self::Value, E>
        where
            E: de::Error,
        {
            self.visit_str(&value)
        }

        fn visit_seq<A>(self, mut seq: A) -> Result<Self::Value, A::Error>
        where
            A: de::SeqAccess<'de>,
        {
            let mut out = Vec::new();
            while let Some(s) = seq.next_element::<String>()? {
                if !s.trim().is_empty() {
                    out.push(s);
                }
            }
            Ok(out)
        }

        fn visit_none<E>(self) -> Result<Self::Value, E>
        where
            E: de::Error,
        {
            Ok(Vec::new())
        }

        fn visit_unit<E>(self) -> Result<Self::Value, E>
        where
            E: de::Error,
        {
            Ok(Vec::new())
        }
    }

    deserializer.deserialize_any(StringOrSeq)
}

#[derive(Debug, Deserialize, Clone)]
pub struct ExecSpec {
    /// Full argument vector; first element is the program.
    pub argv: Vec<String>,
    /// Append extra CLI arguments after those from `argv`.
    #[serde(default)]
    pub passthrough: bool,
}

impl CommandNode {
    pub fn is_leaf_exec(&self) -> bool {
        self.exec.is_some()
    }

    pub fn validate(&self, path: &str) -> Result<()> {
        if self.exec.is_some() && !self.commands.is_empty() {
            bail!("command '{path}' cannot define both `exec` and nested `commands`");
        }
        if let Some(ref e) = self.exec {
            if e.argv.is_empty() {
                bail!("command '{path}': exec.argv must not be empty");
            }
        }
        for name in self.inputs.keys() {
            inputs::InputDef::validate_name(name)
                .map_err(|e| anyhow::anyhow!("command '{path}': {e}"))?;
        }
        for (name, child) in &self.commands {
            let p = if path.is_empty() {
                name.clone()
            } else {
                format!("{path} {name}")
            };
            child.validate(&p)?;
        }
        Ok(())
    }
}

/// Deep-merge `overlay.commands` into `base`, letting included YAML fragments
/// add or replace leaves and extend nested groups.
pub fn merge_specs_into(base: &mut RootSpec, overlay: RootSpec) -> Result<()> {
    for (name, node) in overlay.commands {
        match base.commands.get_mut(&name) {
            Some(existing) => merge_command_node(existing, node)?,
            None => {
                base.commands.insert(name, node);
            }
        }
    }
    Ok(())
}

fn merge_command_node(dst: &mut CommandNode, src: CommandNode) -> Result<()> {
    if src.exec.is_some() && !src.commands.is_empty() {
        bail!("merge overlay: command cannot define both `exec` and nested `commands`");
    }
    if !src.os.is_empty() {
        dst.os = src.os;
    }
    if !src.about.trim().is_empty() {
        dst.about = src.about;
    }
    if src.path.is_some() {
        dst.path = src.path;
    }
    if !src.dependencies.is_empty() {
        dst.dependencies = src.dependencies;
    }
    if !src.requires.is_empty() {
        dst.requires = src.requires;
    }
    if !src.cron.is_empty() {
        dst.cron = src.cron;
    }
    if !src.env.is_empty() {
        dst.env.merge_from(src.env);
    }
    for (k, v) in src.inputs {
        dst.inputs.insert(k, v);
    }
    if let Some(exec) = src.exec {
        dst.exec = Some(exec);
        dst.commands.clear();
        return Ok(());
    }
    if !src.commands.is_empty() {
        dst.exec = None;
        for (k, child) in src.commands {
            match dst.commands.get_mut(&k) {
                Some(existing) => merge_command_node(existing, child)?,
                None => {
                    dst.commands.insert(k, child);
                }
            }
        }
    }
    Ok(())
}

/// Validate every command in the tree (after merges or programmatic edits).
pub fn validate_spec(spec: &RootSpec) -> Result<()> {
    for (name, node) in &spec.commands {
        node.validate(name)?;
    }
    Ok(())
}

/// Parse YAML from memory. Use `include_base` when the document uses `include:` (root or subtree).
pub fn load_spec_from_str(raw: &str, include_base: Option<&Path>) -> Result<RootSpec> {
    spec_load::load_spec_from_str(raw, include_base, HostPlatform::detect())
}

pub fn load_spec(path: &Path) -> Result<RootSpec> {
    spec_load::load_spec_from_path(path, HostPlatform::detect())
}

pub fn resolve_git_branch(cwd: &Path, override_branch: Option<&str>) -> String {
    if let Some(b) = override_branch {
        if !b.is_empty() {
            return b.to_string();
        }
    }
    if let Ok(v) = std::env::var("JAN_BRANCH") {
        if !v.is_empty() {
            return v;
        }
    }
    let output = Command::new("git")
        .args(["rev-parse", "--abbrev-ref", "HEAD"])
        .current_dir(cwd)
        .output();
    match output {
        Ok(o) if o.status.success() => String::from_utf8_lossy(&o.stdout).trim().to_string(),
        _ => "(no-git)".to_string(),
    }
}

fn first_line(s: &str) -> String {
    s.lines().next().unwrap_or("").trim().to_string()
}

pub fn format_help(spec: &RootSpec, chain: &[String], node: Option<&CommandNode>) -> String {
    let mut out = String::new();
    let bin = spec
        .metadata
        .as_ref()
        .and_then(|m| m.name.as_deref())
        .unwrap_or("jan");
    let full_cmd = if chain.is_empty() {
        bin.to_string()
    } else {
        format!("{} {}", bin, chain.join(" "))
    };

    let (about, children, exec) = match node {
        Some(n) => (n.about.as_str(), &n.commands, n.exec.as_ref()),
        None => ("", &spec.commands, None),
    };

    if chain.is_empty() {
        if let Some(meta) = &spec.metadata {
            if let Some(desc) = &meta.description {
                out.push_str(desc.trim());
                out.push_str("\n\n");
            }
        }
    }

    if !about.is_empty() {
        out.push_str(about.trim());
        out.push_str("\n\n");
    }

    if exec.is_some() && children.is_empty() {
        out.push_str("This command runs an external program (see spec `exec.argv`).\n");
        let defs = inputs::collect_chain_inputs(chain, spec);
        if !defs.is_empty() {
            out.push('\n');
            out.push_str(&inputs::format_inputs_help(&defs));
        }
        return out;
    }

    if !children.is_empty() {
        out.push_str("Subcommands:\n");
        for (name, child) in children {
            let line = if child.about.is_empty() {
                format!("  {name}\n")
            } else {
                format!("  {name} — {}\n", first_line(&child.about))
            };
            out.push_str(&line);
        }
        out.push('\n');
        out.push_str(&format!(
            "Use `{} --help` for more about a subcommand.\n",
            full_cmd
        ));
        let defs = inputs::collect_chain_inputs(chain, spec);
        if !defs.is_empty() {
            out.push('\n');
            out.push_str(&inputs::format_inputs_help(&defs));
        }
    } else if exec.is_none() {
        out.push_str("(No subcommands defined.)\n");
    }
    if chain.is_empty() && node.is_none() {
        out.push_str(
            "\nPlace `--help` or `-h` right after the subcommand prefix you want. Built-ins: `use`, `bundle`, `alias`, `list`, `search`, `show`, `validate`, `audit`, `cron`.\n",
        );
    }
    out
}

/// Stable identity for where a YAML spec tree is rooted (directory of linked fragments + entry file).
#[derive(Debug, Clone)]
pub struct SpecRootIdentity {
    /// Canonical directory containing top-level YAML fragments.
    pub spec_dir: String,
    /// Entry YAML file name relative to `spec_dir`.
    pub root_yaml: String,
}

/// Resolve the preferred jan directory saved by `jan use`.
pub fn resolve_preferred_spec() -> Result<(PathBuf, SpecRootIdentity)> {
    let cfg = config::load_user_config().context("load user config")?;
    let Some(dir_s) = cfg
        .jan_dir
        .as_ref()
        .map(|s| s.trim())
        .filter(|s| !s.is_empty())
    else {
        bail!(
            "no preferred jan directory configured\n\
             Run `jan use <DIR>` to save a YAML command tree (see docs/PORTABLE_SCRIPTS.md)"
        );
    };
    let dir = PathBuf::from(dir_s);
    if !dir.is_dir() {
        bail!(
            "preferred jan directory does not exist: {}\n\
             Fix the path or run `jan use <DIR>` again (config: {})",
            dir.display(),
            config::config_path().display()
        );
    }
    let root = cfg
        .spec_root
        .as_deref()
        .map(str::trim)
        .filter(|s| !s.is_empty())
        .unwrap_or("scripts.spec.yaml");
    resolve_spec_dir_entry(&dir, root)
}

/// Resolve a jan directory + entry file name into an absolute spec path and identity.
pub fn resolve_spec_dir_entry(
    spec_dir: &Path,
    root_yaml: &str,
) -> Result<(PathBuf, SpecRootIdentity)> {
    let rel = Path::new(root_yaml);
    if rel.is_absolute() {
        bail!("entry YAML must be a relative file name, not an absolute path");
    }
    if rel
        .components()
        .any(|c| matches!(c, std::path::Component::ParentDir))
    {
        bail!("entry YAML must not contain `..`");
    }
    let normal_only = rel
        .components()
        .all(|c| matches!(c, std::path::Component::Normal(_)));
    let n = rel
        .components()
        .filter(|c| matches!(c, std::path::Component::Normal(_)))
        .count();
    if !normal_only || n != 1 {
        bail!("entry YAML must be a single file name inside the jan directory");
    }
    let dir = spec_dir
        .canonicalize()
        .with_context(|| format!("canonicalize jan directory {}", spec_dir.display()))?;
    if !dir.is_dir() {
        bail!("not a directory: {}", dir.display());
    }
    let spec_path = dir.join(rel);
    if !spec_path.is_file() {
        bail!(
            "spec entry not found: {} (under {})\n\
             Expected a properly formatted jan directory (e.g. scripts.spec.yaml)",
            spec_path.display(),
            dir.display()
        );
    }
    let identity = SpecRootIdentity {
        spec_dir: dir.to_string_lossy().into_owned(),
        root_yaml: rel
            .file_name()
            .expect("relative root has file_name")
            .to_string_lossy()
            .into_owned(),
    };
    Ok((spec_path, identity))
}

pub struct RunContext<'a> {
    pub cwd: &'a Path,
    pub db_path: Option<&'a Path>,
    pub branch: String,
    pub no_log: bool,
    pub spec_root: &'a SpecRootIdentity,
}

pub fn run_matched(
    spec: &RootSpec,
    chain: &[String],
    node: &CommandNode,
    trailing: &[OsString],
    ctx: &RunContext<'_>,
) -> Result<i32> {
    let exec = match &node.exec {
        Some(e) => e,
        None => {
            let help = format_help(spec, chain, Some(node));
            print!("{help}");
            bail!("missing subcommand");
        }
    };
    if exec.argv.is_empty() {
        bail!("exec.argv must not be empty");
    }

    let input_defs = inputs::collect_chain_inputs(chain, spec);
    let (input_vals, rest) = inputs::resolve_inputs(&input_defs, trailing)?;

    let mut argv: Vec<String> = Vec::with_capacity(exec.argv.len());
    for a in &exec.argv {
        argv.push(inputs::interpolate(a, &input_vals)?);
    }
    if exec.passthrough {
        for a in &rest {
            argv.push(a.to_string_lossy().into_owned());
        }
    } else if !rest.is_empty() {
        let preview = rest
            .iter()
            .take(3)
            .map(|s| s.to_string_lossy().into_owned())
            .collect::<Vec<_>>()
            .join(" ");
        bail!(
            "unexpected trailing arguments: {preview}{}",
            if rest.len() > 3 { "…" } else { "" }
        );
    }

    let cmd_path = if chain.is_empty() {
        "(root)".to_string()
    } else {
        chain.join(" ")
    };

    let (_, requires, _) = deps::collect_chain_metadata(chain, spec);
    deps::check_requires(&requires)?;

    let path_dirs = deps::resolve_path_prefixes(spec, chain, ctx)?;
    let program = deps::resolve_program(&argv[0], &path_dirs)?;
    let mut env_spec = deps::collect_chain_env(chain, spec);
    for value in env_spec.public.values_mut() {
        *value = inputs::interpolate(value, &input_vals)?;
    }
    deps::check_private_env(&env_spec.private)?;
    let path_override = if !path_dirs.is_empty() {
        Some(deps::prepend_path_env(&path_dirs)?)
    } else {
        None
    };

    let mut c = Command::new(&program);
    if argv.len() > 1 {
        c.args(&argv[1..]);
    }
    c.current_dir(ctx.cwd);
    deps::apply_process_env(&mut c, &env_spec, path_override)?;

    let status = c.status().with_context(|| format!("spawn `{}`", argv[0]))?;
    let code = status.code().unwrap_or(255);

    if !ctx.no_log {
        if let Some(db) = ctx.db_path {
            log_invocation(
                db,
                &ctx.branch,
                ctx.cwd,
                &cmd_path,
                &argv,
                code,
                ctx.spec_root,
            )?;
        }
    }

    Ok(code)
}

fn ensure_invocations_spec_root_column(conn: &Connection) -> Result<()> {
    let mut stmt = conn.prepare("PRAGMA table_info(invocations)")?;
    let cols: Vec<String> = stmt
        .query_map([], |row| row.get::<_, String>(1))?
        .collect::<std::result::Result<_, _>>()?;
    if !cols.iter().any(|c| c == "spec_root_id") {
        conn.execute(
            "ALTER TABLE invocations ADD COLUMN spec_root_id INTEGER",
            [],
        )?;
    }
    Ok(())
}

fn upsert_spec_root(conn: &Connection, spec: &SpecRootIdentity) -> Result<i64> {
    let ts = unix_ts();
    conn.execute(
        r"INSERT INTO spec_roots (spec_dir, root_yaml, last_used_ts) VALUES (?1, ?2, ?3)
          ON CONFLICT(spec_dir, root_yaml) DO UPDATE SET last_used_ts = excluded.last_used_ts",
        rusqlite::params![&spec.spec_dir, &spec.root_yaml, &ts],
    )?;
    let id: i64 = conn.query_row(
        "SELECT id FROM spec_roots WHERE spec_dir = ?1 AND root_yaml = ?2",
        [&spec.spec_dir, &spec.root_yaml],
        |r| r.get(0),
    )?;
    Ok(id)
}

fn log_invocation(
    db_path: &Path,
    branch: &str,
    cwd: &Path,
    command_path: &str,
    argv: &[String],
    exit_code: i32,
    spec_root: &SpecRootIdentity,
) -> Result<()> {
    if let Some(parent) = db_path.parent() {
        std::fs::create_dir_all(parent).ok();
    }
    let conn = Connection::open(db_path).with_context(|| format!("open {}", db_path.display()))?;
    conn.execute_batch(
        r"
        CREATE TABLE IF NOT EXISTS spec_roots (
            id INTEGER PRIMARY KEY AUTOINCREMENT,
            spec_dir TEXT NOT NULL,
            root_yaml TEXT NOT NULL,
            last_used_ts TEXT NOT NULL,
            UNIQUE(spec_dir, root_yaml)
        );
        CREATE TABLE IF NOT EXISTS invocations (
            id INTEGER PRIMARY KEY AUTOINCREMENT,
            ts TEXT NOT NULL,
            git_branch TEXT NOT NULL,
            cwd TEXT NOT NULL,
            command_path TEXT NOT NULL,
            argv_json TEXT NOT NULL,
            exit_code INTEGER NOT NULL,
            spec_root_id INTEGER
        );
        ",
    )?;
    ensure_invocations_spec_root_column(&conn)?;
    let spec_root_id = upsert_spec_root(&conn, spec_root)?;
    let ts = unix_ts();
    let argv_json = serde_json::to_string(argv).unwrap_or_else(|_| "[]".to_string());
    let cwd_s = cwd.to_string_lossy();
    conn.execute(
        "INSERT INTO invocations (ts, git_branch, cwd, command_path, argv_json, exit_code, spec_root_id)
         VALUES (?1, ?2, ?3, ?4, ?5, ?6, ?7)",
        rusqlite::params![
            ts,
            branch,
            cwd_s.as_ref(),
            command_path,
            argv_json,
            exit_code,
            spec_root_id
        ],
    )?;
    Ok(())
}

fn unix_ts() -> String {
    use std::time::SystemTime;
    SystemTime::now()
        .duration_since(std::time::UNIX_EPOCH)
        .unwrap_or_default()
        .as_secs()
        .to_string()
}

#[derive(Debug)]
pub struct MatchOutcome<'a> {
    pub chain: Vec<String>,
    pub node: Option<&'a CommandNode>,
    pub trailing: Vec<OsString>,
    pub wants_help: bool,
}

pub fn match_commands<'a>(spec: &'a RootSpec, args: &[OsString]) -> MatchOutcome<'a> {
    let mut chain = Vec::new();
    let mut node: Option<&'a CommandNode> = None;
    let mut map: &BTreeMap<String, CommandNode> = &spec.commands;
    let mut i = 0usize;
    let len = args.len();
    while i < len {
        let raw = &args[i];
        if raw == "--help" || raw == "-h" {
            return MatchOutcome {
                chain,
                node,
                trailing: args[i + 1..].to_vec(),
                wants_help: true,
            };
        }
        let key = raw.to_string_lossy();
        if let Some(next) = map.get(key.as_ref()) {
            chain.push(key.into_owned());
            node = Some(next);
            map = &next.commands;
            i += 1;
            continue;
        }
        break;
    }
    MatchOutcome {
        chain,
        node,
        trailing: args[i..].to_vec(),
        wants_help: false,
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use std::io::Write;

    #[test]
    fn examples_default_spec_validates() {
        let path = PathBuf::from(env!("CARGO_MANIFEST_DIR")).join("examples/default.spec.yaml");
        load_spec(&path).unwrap();
    }

    #[test]
    fn merge_specs_adds_and_replaces_leaves() {
        let mut base = load_spec_from_str(
            r"
commands:
  a:
    about: base
    commands:
      x:
        about: old
        exec:
          argv: [echo, old]
",
            None,
        )
        .unwrap();
        let overlay = load_spec_from_str(
            r"
commands:
  a:
    commands:
      x:
        about: new leaf
        exec:
          argv: [echo, new]
  b:
    about: added top
    exec:
      argv: [echo, b]
",
            None,
        )
        .unwrap();
        merge_specs_into(&mut base, overlay).unwrap();
        base.commands["a"].commands["x"].validate("a x").unwrap();
        assert_eq!(
            base.commands["a"].commands["x"].exec.as_ref().unwrap().argv,
            vec!["echo", "new"]
        );
        assert_eq!(
            base.commands["b"].exec.as_ref().unwrap().argv,
            vec!["echo", "b"]
        );
    }

    #[test]
    fn validate_rejects_exec_with_children() {
        let mut tmp = tempfile::NamedTempFile::with_suffix(".yaml").unwrap();
        write!(
            tmp,
            r"
commands:
  x:
    exec:
      argv: [echo]
    commands:
      child:
        about: nested
"
        )
        .unwrap();
        let err = load_spec(tmp.path()).unwrap_err();
        assert!(err.to_string().contains("cannot define both"));
    }
}

pub fn default_db_path() -> PathBuf {
    if let Ok(p) = std::env::var("JAN_DB") {
        return PathBuf::from(p);
    }
    dirs::data_local_dir()
        .unwrap_or_else(|| PathBuf::from("."))
        .join("jan-cli")
        .join("audit.db")
}