//! Images by hash (oxygenflow#39, OxygenFlow docs/jetstream.md §8): rootfs
//! images kept site-wide for runners to fetch, so nobody copies an image
//! between machines. An image arrives as chunks (Oxygen's FastCDC, sent
//! only when this hub lacks them) and then its manifest, which is accepted
//! only when every chunk is here and the image they make has the hash it
//! is named by. Chunks are kept zstd-compressed under
//! `<data>/images/chunks`, shared by every image that has them.

use crate::db::{now_ms, Pool};
use jet::images::{is_hash, Manifest};
use std::io::Write;
use std::path::{Path, PathBuf};

/// Where this hub keeps images.
#[derive(Clone, Debug)]
pub struct Store {
    root: PathBuf,
}

impl Store {
    pub fn new(data: &Path) -> Store {
        Store { root: data.join("images") }
    }

    fn chunk_path(&self, hash: &str) -> PathBuf {
        self.root.join("chunks").join(&hash[..2]).join(hash)
    }

    fn manifest_path(&self, hash: &str) -> PathBuf {
        self.root.join("manifests").join(format!("{hash}.json"))
    }

    pub fn has_chunk(&self, hash: &str) -> bool {
        is_hash(hash) && self.chunk_path(hash).exists()
    }

    /// Keep chunk `hash`, whose bytes the caller has checked.
    pub fn put_chunk(&self, hash: &str, bytes: &[u8]) -> std::io::Result<()> {
        let path = self.chunk_path(hash);
        if path.exists() {
            return Ok(());
        }
        std::fs::create_dir_all(path.parent().expect("a chunk's directory"))?;
        let tmp = path.with_extension(format!("tmp{}", std::process::id()));
        let mut f = std::fs::File::create(&tmp)?;
        f.write_all(&zstd::encode_all(bytes, 3)?)?;
        f.sync_all()?;
        std::fs::rename(&tmp, &path)
    }

    pub fn chunk(&self, hash: &str) -> std::io::Result<Vec<u8>> {
        if !is_hash(hash) {
            return Err(std::io::Error::new(std::io::ErrorKind::NotFound, "not a chunk name"));
        }
        zstd::decode_all(std::fs::File::open(self.chunk_path(hash))?)
    }

    pub fn manifest(&self, hash: &str) -> Option<Manifest> {
        if !is_hash(hash) {
            return None;
        }
        serde_json::from_slice(&std::fs::read(self.manifest_path(hash)).ok()?).ok()
    }

    /// Accept manifest `m`: well formed, every chunk here, and the image
    /// they make hashing to its name. Reads the whole image once.
    pub fn verify_and_keep(&self, m: &Manifest) -> Result<(), String> {
        m.check().map_err(|e| e.to_string())?;
        let mut whole = blake3::Hasher::new();
        for c in &m.chunks {
            if c.is_zeros() {
                whole.update(&vec![0u8; c.len()]);
                continue;
            }
            let bytes = self.chunk(c.hash()).map_err(|_| format!("chunk {} was not sent; send it first", &c.hash()[..12]))?;
            if bytes.len() != c.len() {
                return Err(format!("chunk {} is {} bytes, not {}", &c.hash()[..12], bytes.len(), c.len()));
            }
            whole.update(&bytes);
        }
        let got = whole.finalize().to_hex().to_string();
        if got != m.hash {
            return Err(format!("the chunks make an image named {}, not {}", &got[..12], &m.hash[..12]));
        }
        let path = self.manifest_path(&m.hash);
        std::fs::create_dir_all(path.parent().expect("the manifests directory")).map_err(|e| e.to_string())?;
        let tmp = path.with_extension("json.tmp");
        std::fs::write(&tmp, serde_json::to_vec(m).map_err(|e| e.to_string())?).map_err(|e| e.to_string())?;
        std::fs::rename(&tmp, &path).map_err(|e| e.to_string())
    }
}

/// Record an image this hub now keeps.
pub async fn record(pool: &Pool, m: &Manifest, uploader: &str) -> sqlx::Result<()> {
    sqlx::query(
        "INSERT INTO jet_images (hash, size, arch, chunks, uploader, created_ms) VALUES (?, ?, ?, ?, ?, ?) \
         ON CONFLICT (hash) DO NOTHING",
    )
    .bind(&m.hash)
    .bind(m.size as i64)
    .bind(m.arch.clone().unwrap_or_default())
    .bind(m.chunks.len() as i64)
    .bind(uploader)
    .bind(now_ms())
    .execute(pool)
    .await?;
    if let Some(a) = m.arch.as_ref().filter(|a| !a.is_empty()) {
        crate::db::flow::record_image_archs(pool, &[(m.hash.clone(), a.clone())].into_iter().collect()).await?;
    }
    Ok(())
}

/// Every image this hub keeps, by hash.
pub async fn held(pool: &Pool) -> sqlx::Result<std::collections::HashSet<String>> {
    Ok(sqlx::query_scalar::<_, String>("SELECT hash FROM jet_images").fetch_all(pool).await?.into_iter().collect())
}

/// An image this hub keeps (`jet_images`).
#[derive(sqlx::FromRow)]
pub struct Kept {
    pub hash: String,
    pub size: i64,
    pub arch: String,
    pub chunks: i64,
    pub uploader: String,
    pub created_ms: i64,
}

pub async fn kept(pool: &Pool) -> sqlx::Result<Vec<Kept>> {
    sqlx::query_as("SELECT hash, size, arch, chunks, uploader, created_ms FROM jet_images").fetch_all(pool).await
}

/// One repository's use of one image: its jobs since `since_ms`, and when
/// it last ran one.
#[derive(sqlx::FromRow)]
pub struct Use {
    pub image: String,
    pub owner: String,
    pub name: String,
    pub jobs: i64,
    pub last_ms: i64,
}

/// What each repository's jobs ran on since `since_ms`, newest first.
pub async fn uses(pool: &Pool, since_ms: i64) -> sqlx::Result<Vec<Use>> {
    sqlx::query_as(
        "SELECT lower(json_extract(j.sandbox, '$.rootfs')) AS image, p.owner, p.name, COUNT(*) AS jobs,
                MAX(COALESCE(j.started_ms, r.created_ms)) AS last_ms
         FROM flow_jobs j JOIN flow_runs r ON r.id = j.run_id JOIN repositories p ON p.id = r.repo_id
         WHERE r.created_ms >= ? AND COALESCE(json_extract(j.sandbox, '$.rootfs'), '') != ''
         GROUP BY image, p.id ORDER BY last_ms DESC",
    )
    .bind(since_ms)
    .fetch_all(pool)
    .await
}

/// The architecture runners reported for each image (`flow_images`).
pub async fn reported_archs(pool: &Pool) -> sqlx::Result<std::collections::HashMap<String, String>> {
    Ok(sqlx::query_as::<_, (String, String)>("SELECT hash, arch FROM flow_images").fetch_all(pool).await?.into_iter().collect())
}