czkawka/czkawka_core/src/duplicate.rs

use std::collections::BTreeMap;
#[cfg(target_family = "unix")]
use std::collections::HashSet;
use std::fs::{File, Metadata, OpenOptions};
use std::hash::Hasher;
use std::io::prelude::*;
use std::io::{self, Error, ErrorKind};
use std::io::{BufReader, BufWriter};
#[cfg(target_family = "unix")]
use std::os::unix::fs::MetadataExt;
use std::path::{Path, PathBuf};
use std::sync::atomic::{AtomicBool, AtomicUsize, Ordering};
use std::sync::Arc;
use std::thread::sleep;
use std::time::{Duration, SystemTime, UNIX_EPOCH};
use std::{fs, mem, thread};

use crossbeam_channel::Receiver;
use directories_next::ProjectDirs;
use humansize::{file_size_opts as options, FileSize};
use rayon::prelude::*;

use crate::common::Common;
use crate::common_directory::Directories;
use crate::common_extensions::Extensions;
use crate::common_items::ExcludedItems;
use crate::common_messages::Messages;
use crate::common_traits::*;
use crate::fl;
use crate::localizer::generate_translation_hashmap;

#[derive(Debug)]
pub struct ProgressData {
    pub checking_method: CheckingMethod,
    pub current_stage: u8,
    pub max_stage: u8,
    pub files_checked: usize,
    pub files_to_check: usize,
}

#[derive(PartialEq, Eq, Clone, Debug, Copy)]
pub enum CheckingMethod {
    None,
    Name,
    Size,
    Hash,
}

#[derive(PartialEq, Eq, Clone, Debug, Copy)]
pub enum HashType {
    Blake3,
    Crc32,
    Xxh3,
}

impl HashType {
    fn hasher(self: &HashType) -> Box<dyn MyHasher> {
        match self {
            HashType::Blake3 => Box::new(blake3::Hasher::new()),
            HashType::Crc32 => Box::new(crc32fast::Hasher::new()),
            HashType::Xxh3 => Box::new(xxhash_rust::xxh3::Xxh3::new()),
        }
    }
}

#[derive(Eq, PartialEq, Clone, Debug)]
pub enum DeleteMethod {
    None,
    AllExceptNewest,
    AllExceptOldest,
    OneOldest,
    OneNewest,
    HardLink,
}

#[derive(Clone, Debug, PartialEq, Default)]
pub struct FileEntry {
    pub path: PathBuf,
    pub size: u64,
    pub modified_date: u64,
    pub hash: String,
}

/// Info struck with helpful information's about results
#[derive(Default)]
pub struct Info {
    pub number_of_groups_by_size: usize,
    pub number_of_duplicated_files_by_size: usize,
    pub number_of_groups_by_hash: usize,
    pub number_of_duplicated_files_by_hash: usize,
    pub number_of_groups_by_name: usize,
    pub number_of_duplicated_files_by_name: usize,
    pub lost_space_by_size: u64,
    pub lost_space_by_hash: u64,
    pub gained_space: u64,
}

impl Info {
    pub fn new() -> Self {
        Default::default()
    }
}

/// Struct with required information's to work
pub struct DuplicateFinder {
    text_messages: Messages,
    information: Info,
    files_with_identical_names: BTreeMap<String, Vec<FileEntry>>,    // File Size, File Entry
    files_with_identical_size: BTreeMap<u64, Vec<FileEntry>>,        // File Size, File Entry
    files_with_identical_hashes: BTreeMap<u64, Vec<Vec<FileEntry>>>, // File Size, File Entry
    directories: Directories,
    allowed_extensions: Extensions,
    excluded_items: ExcludedItems,
    recursive_search: bool,
    minimal_file_size: u64,
    maximal_file_size: u64,
    check_method: CheckingMethod,
    delete_method: DeleteMethod,
    hash_type: HashType,
    ignore_hard_links: bool,
    dryrun: bool,
    stopped_search: bool,
    use_cache: bool,
    use_prehash_cache: bool,
    minimal_cache_file_size: u64,
    minimal_prehash_cache_file_size: u64,
    delete_outdated_cache: bool,
}

impl DuplicateFinder {
    pub fn new() -> Self {
        Self {
            text_messages: Messages::new(),
            information: Info::new(),
            files_with_identical_names: Default::default(),
            files_with_identical_size: Default::default(),
            files_with_identical_hashes: Default::default(),
            recursive_search: true,
            allowed_extensions: Extensions::new(),
            check_method: CheckingMethod::None,
            delete_method: DeleteMethod::None,
            minimal_file_size: 8192,
            maximal_file_size: u64::MAX,
            directories: Directories::new(),
            excluded_items: ExcludedItems::new(),
            stopped_search: false,
            ignore_hard_links: true,
            hash_type: HashType::Blake3,
            dryrun: false,
            use_cache: true,
            use_prehash_cache: true,
            minimal_cache_file_size: 1024 * 1024 / 4, // By default cache only >= 256 KB files
            minimal_prehash_cache_file_size: 0,
            delete_outdated_cache: true,
        }
    }

    pub fn find_duplicates(&mut self, stop_receiver: Option<&Receiver<()>>, progress_sender: Option<&futures::channel::mpsc::UnboundedSender<ProgressData>>) {
        self.directories.optimize_directories(self.recursive_search, &mut self.text_messages);

        match self.check_method {
            CheckingMethod::Name => {
                if !self.check_files_name(stop_receiver, progress_sender) {
                    self.stopped_search = true;
                    return;
                }
            }
            CheckingMethod::Size => {
                if !self.check_files_size(stop_receiver, progress_sender) {
                    self.stopped_search = true;
                    return;
                }
            }
            CheckingMethod::Hash => {
                if !self.check_files_size(stop_receiver, progress_sender) {
                    self.stopped_search = true;
                    return;
                }
                if !self.check_files_hash(stop_receiver, progress_sender) {
                    self.stopped_search = true;
                    return;
                }
            }
            CheckingMethod::None => {
                panic!();
            }
        }
        self.delete_files();
        self.debug_print();
    }

    pub fn set_delete_outdated_cache(&mut self, delete_outdated_cache: bool) {
        self.delete_outdated_cache = delete_outdated_cache;
    }

    pub const fn get_check_method(&self) -> &CheckingMethod {
        &self.check_method
    }

    pub fn get_stopped_search(&self) -> bool {
        self.stopped_search
    }

    pub fn set_minimal_cache_file_size(&mut self, minimal_cache_file_size: u64) {
        self.minimal_cache_file_size = minimal_cache_file_size;
    }

    pub fn set_minimal_prehash_cache_file_size(&mut self, minimal_prehash_cache_file_size: u64) {
        self.minimal_prehash_cache_file_size = minimal_prehash_cache_file_size;
    }

    pub const fn get_files_sorted_by_names(&self) -> &BTreeMap<String, Vec<FileEntry>> {
        &self.files_with_identical_names
    }

    pub fn set_use_cache(&mut self, use_cache: bool) {
        self.use_cache = use_cache;
    }

    pub fn set_use_prehash_cache(&mut self, use_prehash_cache: bool) {
        self.use_prehash_cache = use_prehash_cache;
    }

    pub const fn get_files_sorted_by_size(&self) -> &BTreeMap<u64, Vec<FileEntry>> {
        &self.files_with_identical_size
    }

    pub const fn get_files_sorted_by_hash(&self) -> &BTreeMap<u64, Vec<Vec<FileEntry>>> {
        &self.files_with_identical_hashes
    }
    pub fn set_maximal_file_size(&mut self, maximal_file_size: u64) {
        self.maximal_file_size = match maximal_file_size {
            0 => 1,
            t => t,
        };
    }

    pub const fn get_text_messages(&self) -> &Messages {
        &self.text_messages
    }

    pub const fn get_information(&self) -> &Info {
        &self.information
    }

    pub fn set_hash_type(&mut self, hash_type: HashType) {
        self.hash_type = hash_type;
    }

    pub fn set_ignore_hard_links(&mut self, ignore_hard_links: bool) {
        self.ignore_hard_links = ignore_hard_links;
    }

    pub fn set_dryrun(&mut self, dryrun: bool) {
        self.dryrun = dryrun;
    }

    pub fn set_check_method(&mut self, check_method: CheckingMethod) {
        self.check_method = check_method;
    }

    pub fn set_delete_method(&mut self, delete_method: DeleteMethod) {
        self.delete_method = delete_method;
    }

    pub fn set_minimal_file_size(&mut self, minimal_file_size: u64) {
        self.minimal_file_size = match minimal_file_size {
            0 => 1,
            t => t,
        };
    }

    pub fn set_recursive_search(&mut self, recursive_search: bool) {
        self.recursive_search = recursive_search;
    }

    pub fn set_included_directory(&mut self, included_directory: Vec<PathBuf>) -> bool {
        self.directories.set_included_directory(included_directory, &mut self.text_messages)
    }

    pub fn set_excluded_directory(&mut self, excluded_directory: Vec<PathBuf>) {
        self.directories.set_excluded_directory(excluded_directory, &mut self.text_messages);
    }
    pub fn set_allowed_extensions(&mut self, allowed_extensions: String) {
        self.allowed_extensions.set_allowed_extensions(allowed_extensions, &mut self.text_messages);
    }

    pub fn set_excluded_items(&mut self, excluded_items: Vec<String>) {
        self.excluded_items.set_excluded_items(excluded_items, &mut self.text_messages);
    }

    fn check_files_name(&mut self, stop_receiver: Option<&Receiver<()>>, progress_sender: Option<&futures::channel::mpsc::UnboundedSender<ProgressData>>) -> bool {
        let start_time: SystemTime = SystemTime::now();
        let mut folders_to_check: Vec<PathBuf> = Vec::with_capacity(1024 * 2); // This should be small enough too not see to big difference and big enough to store most of paths without needing to resize vector

        // Add root folders for finding
        for id in &self.directories.included_directories {
            folders_to_check.push(id.clone());
        }

        //// PROGRESS THREAD START
        const LOOP_DURATION: u32 = 200; //in ms
        let progress_thread_run = Arc::new(AtomicBool::new(true));

        let atomic_file_counter = Arc::new(AtomicUsize::new(0));

        let progress_thread_handle = if let Some(progress_sender) = progress_sender {
            let progress_send = progress_sender.clone();
            let progress_thread_run = progress_thread_run.clone();
            let atomic_file_counter = atomic_file_counter.clone();
            thread::spawn(move || loop {
                progress_send
                    .unbounded_send(ProgressData {
                        checking_method: CheckingMethod::Name,
                        current_stage: 0,
                        max_stage: 0,
                        files_checked: atomic_file_counter.load(Ordering::Relaxed) as usize,
                        files_to_check: 0,
                    })
                    .unwrap();
                if !progress_thread_run.load(Ordering::Relaxed) {
                    break;
                }
                sleep(Duration::from_millis(LOOP_DURATION as u64));
            })
        } else {
            thread::spawn(|| {})
        };

        //// PROGRESS THREAD END

        while !folders_to_check.is_empty() {
            if stop_receiver.is_some() && stop_receiver.unwrap().try_recv().is_ok() {
                // End thread which send info to gui
                progress_thread_run.store(false, Ordering::Relaxed);
                progress_thread_handle.join().unwrap();
                return false;
            }

            let segments: Vec<_> = folders_to_check
                .par_iter()
                .map(|current_folder| {
                    let mut dir_result = vec![];
                    let mut warnings = vec![];
                    let mut fe_result = vec![];
                    // Read current dir childrens
                    let read_dir = match fs::read_dir(&current_folder) {
                        Ok(t) => t,
                        Err(e) => {
                            warnings.push(fl!(
                                "core_cannot_open_dir",
                                generate_translation_hashmap(vec![("dir", current_folder.display().to_string()), ("reason", e.to_string())])
                            ));
                            return (dir_result, warnings, fe_result);
                        }
                    };

                    // Check every sub folder/file/link etc.
                    'dir: for entry in read_dir {
                        let entry_data = match entry {
                            Ok(t) => t,
                            Err(e) => {
                                warnings.push(fl!(
                                    "core_cannot_read_entry_dir",
                                    generate_translation_hashmap(vec![("dir", current_folder.display().to_string()), ("reason", e.to_string())])
                                ));
                                continue 'dir;
                            }
                        };
                        let metadata: Metadata = match entry_data.metadata() {
                            Ok(t) => t,
                            Err(e) => {
                                warnings.push(fl!(
                                    "core_cannot_read_metadata_dir",
                                    generate_translation_hashmap(vec![("dir", current_folder.display().to_string()), ("reason", e.to_string())])
                                ));
                                continue 'dir;
                            }
                        };
                        if metadata.is_dir() {
                            if !self.recursive_search {
                                continue 'dir;
                            }

                            let next_folder = current_folder.join(entry_data.file_name());
                            if self.directories.is_excluded(&next_folder) {
                                continue 'dir;
                            }

                            if self.excluded_items.is_excluded(&next_folder) {
                                continue 'dir;
                            }

                            dir_result.push(next_folder);
                        } else if metadata.is_file() {
                            atomic_file_counter.fetch_add(1, Ordering::Relaxed);

                            let file_name_lowercase: String = match entry_data.file_name().into_string() {
                                Ok(t) => t,
                                Err(_inspected) => {
                                    warnings.push(fl!(
                                        "core_file_not_utf8_name",
                                        generate_translation_hashmap(vec![("name", entry_data.path().display().to_string())])
                                    ));
                                    continue 'dir;
                                }
                            }
                            .to_lowercase();

                            if !self.allowed_extensions.matches_filename(&file_name_lowercase) {
                                continue 'dir;
                            }

                            if (self.minimal_file_size..=self.maximal_file_size).contains(&metadata.len()) {
                                let current_file_name = current_folder.join(entry_data.file_name());
                                if self.excluded_items.is_excluded(&current_file_name) {
                                    continue 'dir;
                                }

                                let fe: FileEntry = FileEntry {
                                    path: current_file_name.clone(),
                                    size: metadata.len(),
                                    modified_date: match metadata.modified() {
                                        Ok(t) => match t.duration_since(UNIX_EPOCH) {
                                            Ok(d) => d.as_secs(),
                                            Err(_inspected) => {
                                                warnings.push(fl!(
                                                    "core_file_modified_before_epoch",
                                                    generate_translation_hashmap(vec![("name", current_file_name.display().to_string())])
                                                ));
                                                0
                                            }
                                        },
                                        Err(e) => {
                                            warnings.push(fl!(
                                                "core_file_no_modification_date",
                                                generate_translation_hashmap(vec![("name", current_file_name.display().to_string()), ("reason", e.to_string())])
                                            ));
                                            0
                                        }
                                    },
                                    hash: "".to_string(),
                                };

                                fe_result.push((entry_data.file_name().to_string_lossy().to_string(), fe));
                            }
                        }
                    }
                    (dir_result, warnings, fe_result)
                })
                .collect();

            // Advance the frontier
            folders_to_check.clear();

            // Process collected data
            for (segment, warnings, fe_result) in segments {
                folders_to_check.extend(segment);
                self.text_messages.warnings.extend(warnings);
                for (name, fe) in fe_result {
                    self.files_with_identical_names.entry(name.clone()).or_insert_with(Vec::new);
                    self.files_with_identical_names.get_mut(&name).unwrap().push(fe);
                }
            }
        }

        // End thread which send info to gui
        progress_thread_run.store(false, Ordering::Relaxed);
        progress_thread_handle.join().unwrap();

        // Create new BTreeMap without single size entries(files have not duplicates)
        let mut new_map: BTreeMap<String, Vec<FileEntry>> = Default::default();

        for (name, vector) in &self.files_with_identical_names {
            if vector.len() > 1 {
                self.information.number_of_duplicated_files_by_name += vector.len() - 1;
                self.information.number_of_groups_by_name += 1;
                new_map.insert(name.clone(), vector.clone());
            }
        }
        self.files_with_identical_names = new_map;

        Common::print_time(start_time, SystemTime::now(), "check_files_name".to_string());
        true
    }

    /// Read file length and puts it to different boxes(each for different lengths)
    /// If in box is only 1 result, then it is removed
    fn check_files_size(&mut self, stop_receiver: Option<&Receiver<()>>, progress_sender: Option<&futures::channel::mpsc::UnboundedSender<ProgressData>>) -> bool {
        let start_time: SystemTime = SystemTime::now();
        let mut folders_to_check: Vec<PathBuf> = Vec::with_capacity(1024 * 2); // This should be small enough too not see to big difference and big enough to store most of paths without needing to resize vector

        // Add root folders for finding
        for id in &self.directories.included_directories {
            folders_to_check.push(id.clone());
        }

        //// PROGRESS THREAD START
        const LOOP_DURATION: u32 = 200; //in ms
        let progress_thread_run = Arc::new(AtomicBool::new(true));

        let atomic_file_counter = Arc::new(AtomicUsize::new(0));

        let progress_thread_handle = if let Some(progress_sender) = progress_sender {
            let progress_send = progress_sender.clone();
            let progress_thread_run = progress_thread_run.clone();
            let atomic_file_counter = atomic_file_counter.clone();
            let checking_method = self.check_method;
            let max_stage = match self.check_method {
                CheckingMethod::Size => 0,
                CheckingMethod::Hash => 2,
                _ => panic!(),
            };
            thread::spawn(move || loop {
                progress_send
                    .unbounded_send(ProgressData {
                        checking_method,
                        current_stage: 0,
                        max_stage,
                        files_checked: atomic_file_counter.load(Ordering::Relaxed) as usize,
                        files_to_check: 0,
                    })
                    .unwrap();
                if !progress_thread_run.load(Ordering::Relaxed) {
                    break;
                }
                sleep(Duration::from_millis(LOOP_DURATION as u64));
            })
        } else {
            thread::spawn(|| {})
        };

        //// PROGRESS THREAD END

        while !folders_to_check.is_empty() {
            if stop_receiver.is_some() && stop_receiver.unwrap().try_recv().is_ok() {
                // End thread which send info to gui
                progress_thread_run.store(false, Ordering::Relaxed);
                progress_thread_handle.join().unwrap();
                return false;
            }

            let segments: Vec<_> = folders_to_check
                .par_iter()
                .map(|current_folder| {
                    let mut dir_result = vec![];
                    let mut warnings = vec![];
                    let mut fe_result = vec![];
                    // Read current dir childrens
                    let read_dir = match fs::read_dir(&current_folder) {
                        Ok(t) => t,
                        Err(e) => {
                            warnings.push(fl!(
                                "core_cannot_open_dir",
                                generate_translation_hashmap(vec![("dir", current_folder.display().to_string()), ("reason", e.to_string())])
                            ));
                            return (dir_result, warnings, fe_result);
                        }
                    };

                    // Check every sub folder/file/link etc.
                    'dir: for entry in read_dir {
                        let entry_data = match entry {
                            Ok(t) => t,
                            Err(e) => {
                                warnings.push(fl!(
                                    "core_cannot_read_entry_dir",
                                    generate_translation_hashmap(vec![("dir", current_folder.display().to_string()), ("reason", e.to_string())])
                                ));
                                continue 'dir;
                            }
                        };
                        let metadata: Metadata = match entry_data.metadata() {
                            Ok(t) => t,
                            Err(e) => {
                                warnings.push(fl!(
                                    "core_cannot_read_metadata_dir",
                                    generate_translation_hashmap(vec![("dir", current_folder.display().to_string()), ("reason", e.to_string())])
                                ));
                                continue 'dir;
                            }
                        };
                        if metadata.is_dir() {
                            if !self.recursive_search {
                                continue 'dir;
                            }

                            let next_folder = current_folder.join(entry_data.file_name());
                            if self.directories.is_excluded(&next_folder) {
                                continue 'dir;
                            }

                            if self.excluded_items.is_excluded(&next_folder) {
                                continue 'dir;
                            }

                            dir_result.push(next_folder);
                        } else if metadata.is_file() {
                            atomic_file_counter.fetch_add(1, Ordering::Relaxed);

                            let file_name_lowercase: String = match entry_data.file_name().into_string() {
                                Ok(t) => t,
                                Err(_inspected) => {
                                    warnings.push(fl!(
                                        "core_file_not_utf8_name",
                                        generate_translation_hashmap(vec![("name", entry_data.path().display().to_string())])
                                    ));
                                    continue 'dir;
                                }
                            }
                            .to_lowercase();

                            if !self.allowed_extensions.matches_filename(&file_name_lowercase) {
                                continue 'dir;
                            }

                            if (self.minimal_file_size..=self.maximal_file_size).contains(&metadata.len()) {
                                let current_file_name = current_folder.join(entry_data.file_name());
                                if self.excluded_items.is_excluded(&current_file_name) {
                                    continue 'dir;
                                }

                                // Creating new file entry
                                let fe: FileEntry = FileEntry {
                                    path: current_file_name.clone(),
                                    size: metadata.len(),
                                    modified_date: match metadata.modified() {
                                        Ok(t) => match t.duration_since(UNIX_EPOCH) {
                                            Ok(d) => d.as_secs(),
                                            Err(_inspected) => {
                                                warnings.push(fl!(
                                                    "core_file_modified_before_epoch",
                                                    generate_translation_hashmap(vec![("name", current_file_name.display().to_string())])
                                                ));
                                                0
                                            }
                                        },
                                        Err(e) => {
                                            warnings.push(fl!(
                                                "core_file_no_modification_date",
                                                generate_translation_hashmap(vec![("name", current_file_name.display().to_string()), ("reason", e.to_string())])
                                            ));
                                            0
                                        }
                                    },
                                    hash: "".to_string(),
                                };

                                fe_result.push(fe);
                            }
                        }
                    }