netsky 0.2.0

//! `netsky session salvage <session-id|latest>` — retroactive notes
//! from a claude session JSONL.
//!
//! When agent0 dies mid-session before `/notes`, the session transcript
//! at `~/.claude/projects/-Users-cody-netsky/<session-id>.jsonl`
//! persists on disk. This subcommand reads that transcript, extracts
//! the forensic shape (user directives, clones dispatched, commits,
//! tail of unfinished work), and appends a retroactive session block
//! to `notes/<YYYY>/<MM>/<DD>/agent0.md` keyed on the session's first
//! timestamp.
//!
//! The 2026-04-17 outage lost 9 hours of mid-session context across 3
//! crashes because no /notes fired between /up and death. Salvage is
//! the post-hoc recovery path: run it once per dead session and the
//! owner (and the next session's /up) gets back the directive trail.

use std::fs;
use std::io::Write;
use std::path::{Path, PathBuf};

use chrono::Utc;
use netsky_core::Result;
use serde_json::{Value, json};

/// Root under `~/.claude/projects/` where netsky-cwd sessions land.
/// Claude Code derives it from the cwd by replacing `/` with `-`.
const SESSION_PROJECT_DIR: &str = "-Users-cody-netsky";

/// Hard cap on how many extracted lines any single bucket (directives,
/// dispatches, commits) can carry into the notes block. A 9-hour
/// session with thousands of tool calls must not produce a 5000-line
/// notes append — the retro is a lossy summary, not a transcript.
const MAX_PER_BUCKET: usize = 50;

/// Character cap per extracted line. Prevents a single huge bash
/// command or dispatch body from dominating the block.
const MAX_LINE_CHARS: usize = 200;

pub fn salvage(target: &str, json_out: bool) -> Result<()> {
    let jsonl = resolve_session_path(target)?;
    let salvage = read_salvage(&jsonl)?;
    let notes_path = notes_path_for(&salvage.first_ts);
    let appended = append_retro(&notes_path, &salvage)?;
    if json_out {
        let envelope = json!({
            "command": "session",
            "status": "green",
            "summary": format!(
                "salvaged {} line(s) from {}", appended.lines, jsonl.display()
            ),
            "generated_at": Utc::now().format("%Y-%m-%dT%H:%M:%SZ").to_string(),
            "data": {
                "subcommand": "salvage",
                "session_id": salvage.session_id,
                "first_ts": salvage.first_ts,
                "last_ts": salvage.last_ts,
                "source_jsonl": jsonl.display().to_string(),
                "notes_path": notes_path.display().to_string(),
                "appended_lines": appended.lines,
                "directives": salvage.directives.len(),
                "dispatches": salvage.dispatches.len(),
                "commits": salvage.commits.len(),
            },
        });
        println!("{}", serde_json::to_string_pretty(&envelope)?);
    } else {
        println!(
            "salvaged {} lines from {} ({} directives, {} dispatches, {} commits) -> {}",
            appended.lines,
            jsonl.display(),
            salvage.directives.len(),
            salvage.dispatches.len(),
            salvage.commits.len(),
            notes_path.display()
        );
    }
    Ok(())
}

/// Resolve `latest` or a bare UUID into the on-disk jsonl path.
/// Accepts a full path too for manual re-runs against archived
/// sessions.
fn resolve_session_path(target: &str) -> Result<PathBuf> {
    if target.contains('/') {
        let p = PathBuf::from(target);
        if !p.exists() {
            netsky_core::bail!("session file not found: {}", p.display());
        }
        return Ok(p);
    }
    let dir = session_dir()?;
    if target == "latest" {
        let latest = most_recent_jsonl(&dir)?;
        return latest.ok_or_else(|| {
            netsky_core::Error::Message(format!("no *.jsonl sessions under {}", dir.display()))
        });
    }
    let candidate = dir.join(format!("{target}.jsonl"));
    if !candidate.exists() {
        netsky_core::bail!("session {target} not found at {}", candidate.display());
    }
    Ok(candidate)
}

fn session_dir() -> Result<PathBuf> {
    let home = netsky_core::paths::home();
    Ok(home
        .join(".claude")
        .join("projects")
        .join(SESSION_PROJECT_DIR))
}

fn most_recent_jsonl(dir: &Path) -> Result<Option<PathBuf>> {
    let mut best: Option<(std::time::SystemTime, PathBuf)> = None;
    let entries = match fs::read_dir(dir) {
        Ok(iter) => iter,
        Err(e) if e.kind() == std::io::ErrorKind::NotFound => return Ok(None),
        Err(e) => return Err(e.into()),
    };
    for entry in entries.flatten() {
        let path = entry.path();
        if path.extension().and_then(|s| s.to_str()) != Some("jsonl") {
            continue;
        }
        let Ok(meta) = entry.metadata() else { continue };
        let Ok(mtime) = meta.modified() else { continue };
        if best.as_ref().is_none_or(|(ts, _)| mtime > *ts) {
            best = Some((mtime, path));
        }
    }
    Ok(best.map(|(_, p)| p))
}

#[derive(Debug, Default)]
pub struct Salvage {
    pub session_id: String,
    pub first_ts: String,
    pub last_ts: String,
    pub directives: Vec<String>,
    pub dispatches: Vec<String>,
    pub commits: Vec<String>,
    pub tail: Vec<String>,
}

fn read_salvage(path: &Path) -> Result<Salvage> {
    let data = fs::read_to_string(path)?;
    Ok(extract_from_lines(&data))
}

pub fn extract_from_lines(data: &str) -> Salvage {
    let mut s = Salvage::default();
    let mut tail_ring: Vec<String> = Vec::new();

    for raw in data.lines() {
        let Ok(v) = serde_json::from_str::<Value>(raw) else {
            continue;
        };
        let kind = v.get("type").and_then(|x| x.as_str()).unwrap_or("");
        let ts = v.get("timestamp").and_then(|x| x.as_str()).unwrap_or("");
        if !ts.is_empty() {
            if s.first_ts.is_empty() {
                s.first_ts = ts.to_string();
            }
            s.last_ts = ts.to_string();
        }
        if s.session_id.is_empty()
            && let Some(sid) = v.get("sessionId").and_then(|x| x.as_str())
        {
            s.session_id = sid.to_string();
        }
        match kind {
            "user" => collect_user(&v, &mut s, &mut tail_ring),
            "assistant" => collect_assistant(&v, &mut s, &mut tail_ring),
            _ => {}
        }
    }
    s.tail = tail_ring;
    cap_buckets(&mut s);
    s
}

fn collect_user(v: &Value, s: &mut Salvage, tail: &mut Vec<String>) {
    let Some(content) = v.pointer("/message/content") else {
        return;
    };
    let text = content_to_text(content);
    let trimmed = text.trim();
    if trimmed.is_empty() {
        return;
    }
    // Skip command invocations and skill scaffolding: the retro cares
    // about owner intent, not the /up body or /notes skill preamble.
    if trimmed.starts_with("<command-message>")
        || trimmed.starts_with("<command-name>")
        || trimmed.starts_with("Base directory for this skill:")
        || trimmed.starts_with("<system-reminder>")
        || trimmed.starts_with("[SYSTEM NOTIFICATION")
    {
        return;
    }
    // Tool results come back as `user` entries with array content; skip.
    if trimmed.starts_with("[{") || trimmed.starts_with("{\"tool_use_id") {
        return;
    }
    s.directives.push(truncate(trimmed, MAX_LINE_CHARS));
    push_tail(tail, format!("user: {}", truncate(trimmed, MAX_LINE_CHARS)));
}

fn collect_assistant(v: &Value, s: &mut Salvage, tail: &mut Vec<String>) {
    let Some(content) = v.get("message").and_then(|m| m.get("content")) else {
        return;
    };
    let Some(arr) = content.as_array() else {
        return;
    };
    for block in arr {
        match block.get("type").and_then(|x| x.as_str()) {
            Some("tool_use") => collect_tool_use(block, s, tail),
            Some("text") => {
                if let Some(t) = block.get("text").and_then(|x| x.as_str()) {
                    let trimmed = t.trim();
                    if !trimmed.is_empty() {
                        push_tail(
                            tail,
                            format!("agent: {}", truncate(trimmed, MAX_LINE_CHARS)),
                        );
                    }
                }
            }
            _ => {}
        }
    }
}

fn collect_tool_use(block: &Value, s: &mut Salvage, tail: &mut Vec<String>) {
    let name = block.get("name").and_then(|x| x.as_str()).unwrap_or("");
    let input = block.get("input").cloned().unwrap_or(Value::Null);
    match name {
        "Bash" => {
            let cmd = input
                .get("command")
                .and_then(|x| x.as_str())
                .unwrap_or("")
                .trim();
            if cmd.is_empty() {
                return;
            }
            if is_commit_command(cmd) {
                s.commits.push(truncate(cmd, MAX_LINE_CHARS));
            }
            // Spawns via `netsky agent <N>` count as dispatches.
            if cmd.contains("netsky agent ") {
                s.dispatches
                    .push(truncate(&format!("spawn: {cmd}"), MAX_LINE_CHARS));
            }
            push_tail(tail, truncate(&format!("bash: {cmd}"), MAX_LINE_CHARS));
        }
        "mcp__agent__reply" => {
            let chat_id = input
                .get("chat_id")
                .and_then(|x| x.as_str())
                .unwrap_or("agent0");
            let text = input.get("text").and_then(|x| x.as_str()).unwrap_or("");
            s.dispatches.push(truncate(
                &format!("reply -> {chat_id}: {text}"),
                MAX_LINE_CHARS,
            ));
        }
        _ => {}
    }
}

fn is_commit_command(cmd: &str) -> bool {
    // Match `git commit`, `git commit -m ...`, pipelines, and newline
    // continuations. Excludes `git commit --help`/`-h`.
    let cleaned = cmd.replace('\n', " ");
    if cleaned.contains("git commit --help") || cleaned.contains("git commit -h ") {
        return false;
    }
    cleaned.contains("git commit ")
        || cleaned.ends_with("git commit")
        || cleaned.starts_with("git commit")
}

fn content_to_text(v: &Value) -> String {
    if let Some(s) = v.as_str() {
        return s.to_string();
    }
    if let Some(arr) = v.as_array() {
        let mut buf = String::new();
        for item in arr {
            if let Some(t) = item.get("text").and_then(|x| x.as_str()) {
                if !buf.is_empty() {
                    buf.push('\n');
                }
                buf.push_str(t);
            }
        }
        return buf;
    }
    String::new()
}

fn push_tail(tail: &mut Vec<String>, line: String) {
    const TAIL_SIZE: usize = 20;
    tail.push(line);
    if tail.len() > TAIL_SIZE {
        tail.remove(0);
    }
}

fn cap_buckets(s: &mut Salvage) {
    if s.directives.len() > MAX_PER_BUCKET {
        s.directives.drain(0..s.directives.len() - MAX_PER_BUCKET);
    }
    if s.dispatches.len() > MAX_PER_BUCKET {
        s.dispatches.drain(0..s.dispatches.len() - MAX_PER_BUCKET);
    }
    if s.commits.len() > MAX_PER_BUCKET {
        s.commits.drain(0..s.commits.len() - MAX_PER_BUCKET);
    }
}

fn truncate(s: &str, n: usize) -> String {
    if s.chars().count() <= n {
        return s.replace('\n', " ").trim().to_string();
    }
    let head: String = s.chars().take(n).collect();
    format!("{}...", head.trim().replace('\n', " "))
}

fn notes_path_for(first_ts: &str) -> PathBuf {
    let (y, m, d) = date_parts(first_ts);
    PathBuf::from("notes")
        .join(y)
        .join(m)
        .join(d)
        .join("agent0.md")
}

fn date_parts(ts: &str) -> (String, String, String) {
    // Format is RFC3339 — take the date part and split.
    let date_part = ts.split('T').next().unwrap_or(ts);
    let mut split = date_part.split('-');
    let y = split.next().unwrap_or("unknown").to_string();
    let m = split.next().unwrap_or("unknown").to_string();
    let d = split.next().unwrap_or("unknown").to_string();
    (y, m, d)
}

struct AppendStats {
    lines: usize,
}

fn append_retro(path: &Path, s: &Salvage) -> Result<AppendStats> {
    if let Some(parent) = path.parent() {
        fs::create_dir_all(parent)?;
    }
    let block = render_block(s);
    let mut f = fs::OpenOptions::new()
        .create(true)
        .append(true)
        .open(path)?;
    if path.metadata().map(|m| m.len()).unwrap_or(0) == 0 {
        writeln!(f, "# notes")?;
        writeln!(f)?;
    }
    f.write_all(block.as_bytes())?;
    Ok(AppendStats {
        lines: block.lines().count(),
    })
}

pub fn render_block(s: &Salvage) -> String {
    let mut out = String::new();
    out.push_str(&format!(
        "## session salvage ({} -> {} UTC)\n\n",
        short_ts(&s.first_ts),
        short_ts(&s.last_ts)
    ));
    out.push_str(&format!("- session id: {}\n", s.session_id));
    out.push_str("- source: post-hoc from claude JSONL (agent0 died before /notes)\n\n");

    if !s.directives.is_empty() {
        out.push_str("### directives\n\n");
        for d in &s.directives {
            out.push_str(&format!("- {d}\n"));
        }
        out.push('\n');
    }

    if !s.dispatches.is_empty() {
        out.push_str("### dispatches\n\n");
        for d in &s.dispatches {
            out.push_str(&format!("- {d}\n"));
        }
        out.push('\n');
    }

    if !s.commits.is_empty() {
        out.push_str("### commits\n\n");
        for c in &s.commits {
            out.push_str(&format!("- {c}\n"));
        }
        out.push('\n');
    }

    if !s.tail.is_empty() {
        out.push_str("### tail (last 20 events, unfinished work)\n\n");
        for t in &s.tail {
            out.push_str(&format!("- {t}\n"));
        }
        out.push('\n');
    }

    out
}

fn short_ts(ts: &str) -> &str {
    // Trim sub-second precision from `2026-04-17T19:14:05.123Z` to
    // `2026-04-17T19:14:05` for readability.
    ts.split('.').next().unwrap_or(ts)
}

#[cfg(test)]
mod tests {
    use super::*;
    use tempfile::tempdir;

    fn sample_jsonl() -> String {
        [
            r#"{"type":"user","message":{"role":"user","content":"<command-message>up</command-message>\n<command-name>/up</command-name>"},"timestamp":"2026-04-17T20:00:00.000Z","sessionId":"abc-123"}"#,
            r#"{"type":"user","message":{"role":"user","content":"please salvage item 3 — the session-salvage subcommand"},"timestamp":"2026-04-17T20:01:00.000Z","sessionId":"abc-123"}"#,
            r#"{"type":"assistant","message":{"content":[{"type":"tool_use","name":"Bash","input":{"command":"git commit -m 'wip'"}}]},"timestamp":"2026-04-17T20:02:00.000Z","sessionId":"abc-123"}"#,
            r#"{"type":"assistant","message":{"content":[{"type":"tool_use","name":"mcp__agent__reply","input":{"chat_id":"agent0","text":"done: SHA eb91d34"}}]},"timestamp":"2026-04-17T20:03:00.000Z","sessionId":"abc-123"}"#,
            r#"{"type":"assistant","message":{"content":[{"type":"tool_use","name":"Bash","input":{"command":"netsky agent 11 --type codex --prompt briefs/x.md"}}]},"timestamp":"2026-04-17T20:04:00.000Z","sessionId":"abc-123"}"#,
            r#"{"type":"assistant","message":{"content":[{"type":"text","text":"all done — wave landed"}]},"timestamp":"2026-04-17T20:05:00.000Z","sessionId":"abc-123"}"#,
        ]
        .join("\n")
    }

    #[test]
    fn extract_captures_all_buckets() {
        let s = extract_from_lines(&sample_jsonl());
        assert_eq!(s.session_id, "abc-123");
        assert_eq!(s.first_ts, "2026-04-17T20:00:00.000Z");
        assert_eq!(s.last_ts, "2026-04-17T20:05:00.000Z");
        assert_eq!(s.directives.len(), 1, "got {:?}", s.directives);
        assert!(s.directives[0].contains("salvage item 3"));
        assert_eq!(s.commits.len(), 1);
        assert!(s.commits[0].contains("git commit"));
        assert_eq!(s.dispatches.len(), 2, "reply + netsky agent spawn");
        assert!(!s.tail.is_empty());
    }

    #[test]
    fn skips_command_scaffolding_and_tool_results() {
        let s = extract_from_lines(&sample_jsonl());
        assert!(
            s.directives.iter().all(|d| !d.contains("command-message")),
            "command scaffolding must be filtered"
        );
    }

    #[test]
    fn render_block_has_stable_shape() {
        let s = extract_from_lines(&sample_jsonl());
        let out = render_block(&s);
        assert!(out.starts_with("## session salvage"));
        assert!(out.contains("### directives"));
        assert!(out.contains("### dispatches"));
        assert!(out.contains("### commits"));
        assert!(out.contains("### tail"));
        assert!(out.contains("abc-123"));
    }

    #[test]
    fn append_retro_creates_notes_path_and_header() {
        let dir = tempdir().unwrap();
        let path = dir.path().join("agent0.md");
        let s = extract_from_lines(&sample_jsonl());
        let stats = append_retro(&path, &s).unwrap();
        assert!(stats.lines > 0);
        let body = fs::read_to_string(&path).unwrap();
        assert!(body.starts_with("# notes\n"), "seeds header on empty file");
        assert!(body.contains("## session salvage"));

        // Second append appends, does not duplicate the header.
        append_retro(&path, &s).unwrap();
        let body2 = fs::read_to_string(&path).unwrap();
        assert_eq!(
            body2.matches("# notes\n").count(),
            1,
            "header seeded once, subsequent appends add below"
        );
        assert_eq!(body2.matches("## session salvage").count(), 2);
    }

    #[test]
    fn notes_path_for_uses_first_ts() {
        let p = notes_path_for("2026-04-17T19:14:05.123Z");
        assert_eq!(p, PathBuf::from("notes/2026/04/17/agent0.md"), "got {p:?}");
    }

    #[test]
    fn commit_detector_matches_real_shapes() {
        assert!(is_commit_command("git commit -m 'x'"));
        assert!(is_commit_command("git add -A && git commit -m 'x'"));
        assert!(is_commit_command("git commit"));
        assert!(!is_commit_command("git log -5"));
        assert!(!is_commit_command("git commit --help"));
    }

    #[test]
    fn truncate_respects_char_cap() {
        let long = "a".repeat(500);
        let out = truncate(&long, MAX_LINE_CHARS);
        assert!(out.len() <= MAX_LINE_CHARS + 3, "len {}", out.len());
        assert!(out.ends_with("..."));
    }

    #[test]
    fn caps_buckets_at_max_per_bucket() {
        let mut data = String::new();
        for i in 0..(MAX_PER_BUCKET + 10) {
            data.push_str(&format!(
                r#"{{"type":"assistant","message":{{"content":[{{"type":"tool_use","name":"Bash","input":{{"command":"git commit -m 'n{i}'"}}}}]}},"timestamp":"2026-04-17T20:00:0{}.000Z","sessionId":"x"}}"#,
                i % 10
            ));
            data.push('\n');
        }
        let s = extract_from_lines(&data);
        assert_eq!(s.commits.len(), MAX_PER_BUCKET);
    }

    #[test]
    fn latest_returns_none_for_empty_dir() {
        let dir = tempdir().unwrap();
        assert!(most_recent_jsonl(dir.path()).unwrap().is_none());
    }

    #[test]
    fn latest_picks_most_recent_mtime() {
        let dir = tempdir().unwrap();
        let a = dir.path().join("a.jsonl");
        let b = dir.path().join("b.jsonl");
        fs::write(&a, "").unwrap();
        std::thread::sleep(std::time::Duration::from_millis(30));
        fs::write(&b, "").unwrap();
        let latest = most_recent_jsonl(dir.path()).unwrap().unwrap();
        assert_eq!(latest, b);
    }
}