version 0.0.5

Fix some more unhandled errors

Cargo.lock

@@ -269,7 +269,7 @@ dependencies = [
 
 [[package]]
 name = "deduplicator"
-version = "0.0.1"
+version = "0.0.5"
 dependencies = [
  "anyhow",
  "chrono",

Cargo.toml

@@ -1,9 +1,10 @@
 [package]
 name = "deduplicator"
-version = "0.0.2"
+version = "0.0.5"
 edition = "2021"
 description = "find,filter,delete Duplicates"
 license = "MIT"
+authors = ["Sreedev Kodichath <sreedevpadmakumar@gmail.com>", "Valentin Bersier <vbersier@gmail.com>", "Dhruva Sagar <dhruva.sagar@gmail.com>"]
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 

src/app/event_handler.rs

@@ -1,7 +1,8 @@
 use std::time::Duration;
 
-use crossterm::event::{self, KeyCode, KeyEvent};
 use anyhow::Result;
+use crossterm::event::{self, KeyCode, KeyEvent};
+
 use super::events;
 
 pub struct EventHandler;
@@ -21,8 +22,7 @@ impl EventHandler {
     fn handle_keypress(keyevent: KeyEvent) -> Result<events::Event> {
         match keyevent.code {
             KeyCode::Char('q') => Ok(events::Event::Exit),
-            _ => Ok(events::Event::Noop)
+            _ => Ok(events::Event::Noop),
         }
     }
 }
-

src/app/events.rs

@@ -1,4 +1,4 @@
 pub enum Event {
     Exit,
-    Noop
+    Noop,
 }

src/app/mod.rs

@@ -1,18 +1,13 @@
 mod event_handler;
 mod events;
-mod ui;
 mod formatter;
+mod ui;
+
+use std::{io, thread, time::Duration};
 
-use crate::database;
-use crate::output;
-use crate::params::Params;
-use crate::scanner;
 use anyhow::{anyhow, Result};
 use crossterm::{event, execute, terminal};
 use event_handler::EventHandler;
-use std::io;
-use std::thread;
-use std::time::Duration;
 use tui::{
     backend::CrosstermBackend,
     widgets::{Block, Borders, Widget},
@@ -20,19 +15,24 @@ use tui::{
 };
 use ui::Ui;
 
+use crate::database;
+use crate::output;
+use crate::params::Params;
+use crate::scanner;
+
 pub struct App;
 
 impl App {
     pub fn init(app_args: &Params) -> Result<()> {
         // let mut term = Self::init_terminal()?;
 
-        let connection = database::get_connection(&app_args)?;
+        let connection = database::get_connection(app_args)?;
-        let duplicates = scanner::duplicates(&app_args, &connection)?;
+        let duplicates = scanner::duplicates(app_args, &connection)?;
 
         // Self::init_render_loop(&mut term)?;
         // Self::cleanup(&mut term)?;
 
-        output::print(duplicates, &app_args); /* TODO: APP TUI INIT FUNCTION */
+        output::print(duplicates, app_args); /* TODO: APP TUI INIT FUNCTION */
         Ok(())
     }
 
@@ -56,6 +56,8 @@ impl App {
     }
 
     fn init_render_loop(term: &mut Terminal<CrosstermBackend<io::Stdout>>) -> Result<()> {
+        // this could be simplified with a `while Self::render_cycle(term).is_ok() {}` in the current state, but maybe
+        // it's good to keep it to handle errors in the future
         loop {
             match Self::render_cycle(term) {
                 Ok(_) => continue,

src/app/ui.rs

@@ -1,5 +1,6 @@
-use anyhow::Result;
 use std::io;
+
+use anyhow::Result;
 use tui::{
     backend::{Backend, CrosstermBackend},
     layout::{Constraint, Direction, Layout, Rect},

src/database.rs

@@ -1,4 +1,7 @@
+use std::env::temp_dir;
+
 use anyhow::Result;
+
 use crate::params::Params;
 
 #[derive(Debug, Clone)]
@@ -7,13 +10,18 @@ pub struct File {
     pub hash: String,
 }
 
-pub fn get_connection(args: &Params) -> Result<sqlite::Connection, sqlite::Error> {
+fn db_connection_url(args: &Params) -> String {
-    let connection_url = match args.nocache {
+    match args.nocache {
-        false => "/tmp/deduplicator.db",
+        true => String::from(":memory:"),
-        true => ":memory:"
+        false => {
-    };
+            let temp_dir_path = temp_dir();
+            format!("{}/deduplicator.db", temp_dir_path.display())
+        }
+    }
+}
 
-    sqlite::open(connection_url).and_then(|conn| {
+pub fn get_connection(args: &Params) -> Result<sqlite::Connection, sqlite::Error> {
+    sqlite::open(db_connection_url(args)).and_then(|conn| {
         setup(&conn).ok();
         Ok(conn)
     })
@@ -30,31 +38,28 @@ pub fn put(file: &File, connection: &sqlite::Connection) -> Result<()> {
         "INSERT INTO files (file_identifier, hash) VALUES (\"{}\", \"{}\")",
         file.path, file.hash
     );
-    let result = connection.execute(query)?;
+    connection.execute(query)?;
-
+    Ok(())
-    Ok(result)
 }
 
 pub fn indexed_paths(connection: &sqlite::Connection) -> Result<Vec<File>> {
-    let query = format!(
+    let query = "SELECT * FROM files";
-        "SELECT * FROM files"
-        );
 
     let result: Vec<File> = connection
         .prepare(query)?
         .into_iter()
-        .map(|row_result| row_result.unwrap())
+        .filter_map(|row_result| row_result.ok())
         .map(|row| {
             let path = row.read::<&str, _>("file_identifier").to_string();
             let hash = row.read::<i64, _>("hash").to_string();
             File { path, hash }
         })
-    .collect();
+        .collect();
 
     Ok(result)
 }
 
-pub fn duplicate_hashes(connection: &sqlite::Connection, path: &String) -> Result<Vec<File>> {
+pub fn duplicate_hashes(connection: &sqlite::Connection, path: &str) -> Result<Vec<File>> {
     let query = format!(
         "
             SELECT a.* FROM files a
@@ -65,13 +70,14 @@ pub fn duplicate_hashes(connection: &sqlite::Connection, path: &String) -> Resul
             ON a.hash = b.hash
             WHERE a.file_identifier LIKE \"{}%\"
             ORDER BY a.file_identifier
-        ", path
+        ",
+        path
     );
 
     let result: Vec<File> = connection
         .prepare(query)?
         .into_iter()
-        .map(|row_result| row_result.unwrap())
+        .filter_map(|row_result| row_result.ok())
         .map(|row| {
             let path = row.read::<&str, _>("file_identifier").to_string();
             let hash = row.read::<i64, _>("hash").to_string();

src/main.rs

@@ -1,12 +1,13 @@
-mod params;
+#![allow(unused)] // TODO: remove this once TUI is implemented
+mod app;
 mod database;
 mod output;
+mod params;
 mod scanner;
-mod app;
 
 use anyhow::Result;
-use clap::Parser;
 use app::App;
+use clap::Parser;
 
 #[tokio::main]
 async fn main() -> Result<()> {

src/output.rs

@@ -1,33 +1,40 @@
-use crate::database::File;
+use std::{collections::HashMap, fs};
+
+use anyhow::Result;
 use chrono::offset::Utc;
 use chrono::DateTime;
 use colored::Colorize;
 use humansize::{format_size, DECIMAL};
-use std::{collections::HashMap, fs};
+
+use crate::database::File;
 use crate::params::Params;
 
-fn format_path(path: &String, opts: &Params) -> String {
+fn format_path(path: &str, opts: &Params) -> Result<String> {
-    let display_path = path.replace(&opts.get_directory().unwrap(), "");
+    let display_path = path.replace(&opts.get_directory()?, "");
-    let display_range = if display_path.len() > 32 {
+    let text_vec = display_path.chars().collect::<Vec<_>>();
-            &display_path[(display_path.len() - 32)..]
+
+    let display_range = if text_vec.len() > 32 {
+        text_vec[(display_path.len() - 32)..]
+            .iter()
+            .collect::<String>()
     } else {
-        &display_path[..]
+        display_path
     };
 
-    format!("...{}", display_range)
+    Ok(format!("...{}", display_range))
 }
 
-fn file_size(path: &String) -> String {
+fn file_size(path: &String) -> Result<String> {
-    let mdata = fs::metadata(path).unwrap();
+    let mdata = fs::metadata(path)?;
     let formatted_size = format_size(mdata.len(), DECIMAL);
-    format!("{}", formatted_size)
+    Ok(formatted_size)
 }
 
-fn modified_time(path: &String) -> String {
+fn modified_time(path: &String) -> Result<String> {
-    let mdata = fs::metadata(path).unwrap();
+    let mdata = fs::metadata(path)?;
-    let modified_time: DateTime<Utc> = mdata.modified().unwrap().into();
+    let modified_time: DateTime<Utc> = mdata.modified()?.into();
 
-    modified_time.format("%Y-%m-%d %H:%M:%S").to_string()
+    Ok(modified_time.format("%Y-%m-%d %H:%M:%S").to_string())
 }
 
 fn print_divider() {
@@ -48,17 +55,17 @@ pub fn print(duplicates: Vec<File>, opts: &Params) {
         dup_index
             .entry(file.hash.clone())
             .and_modify(|value| value.push(file.clone()))
-            .or_insert(vec![file]);
+            .or_insert_with(|| vec![file]);
     });
 
-    dup_index.into_iter().for_each(|(_, group)| {
+    dup_index.iter().for_each(|(_, group)| {
-        group.into_iter().for_each(|file| {
+        group.iter().for_each(|file| {
             println!(
                 "| {0: <16} | {1: <35} | {2: <16} | {3: <32} |",
                 file.hash.red(),
-                format_path(&file.path, opts).yellow(),
+                format_path(&file.path, opts).unwrap_or_default().yellow(),
-                file_size(&file.path).blue(),
+                file_size(&file.path).unwrap_or_default().blue(),
-                modified_time(&file.path).blue()
+                modified_time(&file.path).unwrap_or_default().blue()
             );
         });
         print_divider();

src/params.rs

@@ -1,7 +1,7 @@
-use std::path::PathBuf;
+use std::{fs, path::PathBuf};
+
+use anyhow::{anyhow, Result};
 use clap::Parser;
-use anyhow::Result;
-use std::fs;
 
 #[derive(Parser, Debug)]
 #[command(author, version, about, long_about = None)]
@@ -19,20 +19,17 @@ pub struct Params {
 
 impl Params {
     pub fn get_directory(&self) -> Result<String> {
-        let dir_string: String = self
+        let dir_pathbuf: PathBuf = self
             .dir
             .clone()
             .unwrap_or(std::env::current_dir()?)
             .as_os_str()
-            .to_str()
+            .into();
-            .unwrap()
-            .to_string();
 
-        let dir_pathbuf = PathBuf::from(&dir_string);
+        let dir = fs::canonicalize(dir_pathbuf)?
-        let dir = fs::canonicalize(&dir_pathbuf)?
             .as_os_str()
             .to_str()
-            .unwrap()
+            .ok_or_else(|| anyhow!("Invalid directory"))?
             .to_string();
 
         Ok(dir)

src/scanner.rs

@@ -1,55 +1,61 @@
-use crate::database;
+use std::{fs, path::PathBuf};
-use crate::{params::Params, database::File};
+
 use anyhow::Result;
+use fxhash::hash32 as hasher;
 use glob::glob;
 use itertools::Itertools;
 use rayon::prelude::*;
-use std::fs;
+
-use std::path::PathBuf;
+use crate::{
-use fxhash::hash32 as hasher;
+    database::{self, File},
+    params::Params,
+};
 
 pub fn duplicates(app_opts: &Params, connection: &sqlite::Connection) -> Result<Vec<File>> {
     let scan_results = scan(app_opts, connection)?;
     let base_path = app_opts.get_directory()?;
 
-    index_files(scan_results, connection);
+    index_files(scan_results, connection)?;
     database::duplicate_hashes(connection, &base_path)
 }
 
-fn get_glob_patterns(opts: &Params, directory: &String) -> Vec<PathBuf> {
+fn get_glob_patterns(opts: &Params, directory: &str) -> Vec<PathBuf> {
     opts.types
         .clone()
-        .unwrap_or(String::from("*"))
+        .unwrap_or_else(|| String::from("*"))
-        .split(",")
+        .split(',')
         .map(|filetype| format!("*.{}", filetype))
         .map(|filetype| {
-            vec![directory.clone(), String::from("**"), filetype]
+            vec![directory.to_owned(), String::from("**"), filetype]
                 .iter()
                 .collect()
         })
         .collect()
 }
 
-fn is_indexed_file(path: &String, indexed: &Vec<File>) -> bool {
+fn is_indexed_file(path: impl Into<String>, indexed: &[File]) -> bool {
     indexed
-        .into_iter()
+        .iter()
         .map(|file| file.path.clone())
-        .contains(path)
+        .contains(&path.into())
 }
 
 fn scan(app_opts: &Params, connection: &sqlite::Connection) -> Result<Vec<String>> {
     let directory = app_opts.get_directory()?;
-    let glob_patterns: Vec<PathBuf> = get_glob_patterns(&app_opts, &directory);
+    let glob_patterns: Vec<PathBuf> = get_glob_patterns(app_opts, &directory);
     let indexed_paths = database::indexed_paths(connection)?;
     let files: Vec<String> = glob_patterns
-        .into_par_iter()
+        .par_iter()
-        .map(|glob_pattern| glob(&glob_pattern.as_os_str().to_str().unwrap()))
+        .filter_map(|glob_pattern| glob(glob_pattern.as_os_str().to_str()?).ok())
-        .map(|glob_result| glob_result.unwrap())
         .flat_map(|file_vec| {
             file_vec
-                .map(|x| x.unwrap().as_os_str().to_str().unwrap().to_string())
+                .filter_map(|x| Some(x.ok()?.as_os_str().to_str()?.to_string()))
                 .filter(|fpath| !is_indexed_file(fpath, &indexed_paths))
-                .filter(|glob_result| fs::metadata(glob_result).unwrap().is_file())
+                .filter(|glob_result| {
+                    fs::metadata(glob_result)
+                        .map(|f| f.is_file())
+                        .unwrap_or(false)
+                })
                 .collect::<Vec<String>>()
         })
         .collect();
@@ -57,18 +63,18 @@ fn scan(app_opts: &Params, connection: &sqlite::Connection) -> Result<Vec<String
     Ok(files)
 }
 
-fn index_files(files: Vec<String>, connection: &sqlite::Connection) {
+fn index_files(files: Vec<String>, connection: &sqlite::Connection) -> Result<()> {
     let hashed: Vec<File> = files
         .into_par_iter()
-        .map(|file| {
+        .filter_map(|file| {
-            let hash = hash_file(&file).unwrap();
+            let hash = hash_file(&file).ok()?;
-            database::File { path: file, hash }
+            Some(database::File { path: file, hash })
         })
         .collect();
 
-    hashed.into_iter().for_each(|file| {
+    hashed
-        database::put(&file, connection).unwrap();
+        .iter()
-    });
+        .try_for_each(|file| database::put(file, connection))
 }
 
 pub fn hash_file(filepath: &str) -> Result<String> {