alex/src/backup/mod.rs

mod delta;
mod manager;
mod path;

use delta::Delta;
pub use manager::Manager;

use chrono::Utc;
use flate2::write::GzEncoder;
use flate2::Compression;
use path::PathExt;
use serde::{Deserialize, Serialize};
use std::collections::{HashMap, HashSet};
use std::fs::File;
use std::io;
use std::path::{Path, PathBuf};

#[derive(Debug, PartialEq, Serialize, Deserialize)]
pub enum BackupType {
    Full,
    Incremental,
}

/// Represents a successful backup
#[derive(Serialize, Deserialize)]
pub struct Backup<T: Clone> {
    /// When the backup was started (also corresponds to the name)
    pub start_time: chrono::DateTime<Utc>,
    /// Type of the backup
    pub type_: BackupType,
    pub delta: Delta,
    /// Additional metadata that can be associated with a given backup
    pub metadata: Option<T>,
}

impl Backup<()> {
    /// Return the path to a backup file by properly formatting the data.
    pub fn path<P: AsRef<Path>>(backup_dir: P, start_time: chrono::DateTime<Utc>) -> PathBuf {
        let backup_dir = backup_dir.as_ref();

        let filename = format!("{}", start_time.format(Self::FILENAME_FORMAT));
        backup_dir.join(filename)
    }
}

impl<T: Clone> Backup<T> {
    const FILENAME_FORMAT: &str = "%Y-%m-%d_%H-%M-%S.tar.gz";

    pub fn set_metadata(&mut self, metadata: T) {
        self.metadata = Some(metadata);
    }

    /// Resolve the state of the list of backups by applying their deltas in-order to an initially
    /// empty state.
    pub fn state(backups: &Vec<Self>) -> HashMap<PathBuf, HashSet<PathBuf>> {
        let mut state: HashMap<PathBuf, HashSet<PathBuf>> = HashMap::new();

        for backup in backups {
            backup.delta.apply(&mut state);
        }

        state
    }

    /// Create a new Full backup, populated with the given directories.
    ///
    /// # Arguments
    ///
    /// * `backup_dir` - Directory to store archive in
    /// * `dirs` - list of tuples `(path_in_tar, src_dir)` with `path_in_tar` the directory name
    ///     under which `src_dir`'s contents should be stored in the archive
    ///
    /// # Returns
    ///
    /// The `Backup` instance describing this new backup.
    pub fn create<P: AsRef<Path>>(
        backup_dir: P,
        dirs: Vec<(PathBuf, PathBuf)>,
    ) -> io::Result<Self> {
        let start_time = chrono::offset::Utc::now();

        let path = Backup::path(backup_dir, start_time);
        let tar_gz = File::create(path)?;
        let enc = GzEncoder::new(tar_gz, Compression::default());
        let mut ar = tar::Builder::new(enc);

        let mut delta = Delta::new();

        for (dir_in_tar, src_dir) in dirs {
            let mut added_files: HashSet<PathBuf> = HashSet::new();

            for entry in src_dir.read_dir_recursive()?.ignored("cache").files() {
                let path = entry?.path();
                let stripped = path.strip_prefix(&src_dir).unwrap();

                ar.append_path_with_name(&path, dir_in_tar.join(stripped))?;
                added_files.insert(stripped.to_path_buf());
            }

            delta.added.insert(dir_in_tar, added_files);
        }

        Ok(Backup {
            type_: BackupType::Full,
            start_time,
            delta,
            metadata: None,
        })
    }

    /// Create a new incremental backup from a given previous backup
    pub fn create_from<P: AsRef<Path>>(
        previous_state: HashMap<PathBuf, HashSet<PathBuf>>,
        previous_start_time: chrono::DateTime<Utc>,
        backup_dir: P,
        dirs: Vec<(PathBuf, PathBuf)>,
    ) -> io::Result<Self> {
        let start_time = chrono::offset::Utc::now();

        let path = Backup::path(backup_dir, start_time);
        let tar_gz = File::create(path)?;
        let enc = GzEncoder::new(tar_gz, Compression::default());
        let mut ar = tar::Builder::new(enc);

        let mut delta = Delta::new();

        for (dir_in_tar, src_dir) in dirs {
            let mut all_files: HashSet<PathBuf> = HashSet::new();
            let mut added_files: HashSet<PathBuf> = HashSet::new();

            for entry in src_dir.read_dir_recursive()?.ignored("cache").files() {
                let path = entry?.path();
                let stripped = path.strip_prefix(&src_dir).unwrap();

                if !path.not_modified_since(previous_start_time) {
                    ar.append_path_with_name(&path, dir_in_tar.join(stripped))?;
                    added_files.insert(stripped.to_path_buf());
                }

                all_files.insert(stripped.to_path_buf());
            }

            delta.added.insert(dir_in_tar.clone(), added_files);

            if let Some(previous_files) = previous_state.get(&dir_in_tar) {
                delta.removed.insert(
                    dir_in_tar,
                    previous_files.difference(&all_files).cloned().collect(),
                );
            }
        }

        Ok(Backup {
            type_: BackupType::Incremental,
            start_time,
            delta,
            metadata: None,
        })
    }
}