feat(media): Implement directory listing

Work on tokenizing file names for TMDB API calls in future.

Cargo.lock

@@ -94,6 +94,12 @@ version = "3.14.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7f30e7476521f6f8af1a1c4c0b8cc94f0bee37d91763d0ca2665f299b6cd8aec"
 
+[[package]]
+name = "byteorder"
+version = "1.5.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1fd0f2584146f6f2ef48085050886acf353beff7305ebd1ae69500e27c67f64b"
+
 [[package]]
 name = "cc"
 version = "1.0.83"
@@ -103,6 +109,17 @@ dependencies = [
  "libc",
 ]
 
+[[package]]
+name = "cfb"
+version = "0.7.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d38f2da7a0a2c4ccf0065be06397cc26a81f4e528be095826eee9d4adbb8c60f"
+dependencies = [
+ "byteorder",
+ "fnv",
+ "uuid",
+]
+
 [[package]]
 name = "cfg-if"
 version = "1.0.0"
@@ -206,6 +223,12 @@ version = "1.0.14"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "23d2f3407d9a573d666de4b5bdf10569d73ca9478087346697dcbae6244bfbcd"
 
+[[package]]
+name = "fnv"
+version = "1.0.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "3f9eec918d3f24069decb9af1554cad7c880e2da24a9afd88aca000531ab82c1"
+
 [[package]]
 name = "heck"
 version = "0.4.1"
@@ -244,6 +267,15 @@ dependencies = [
  "cc",
 ]
 
+[[package]]
+name = "infer"
+version = "0.15.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cb33622da908807a06f9513c19b3c1ad50fab3e4137d82a78107d502075aa199"
+dependencies = [
+ "cfb",
+]
+
 [[package]]
 name = "inquire"
 version = "0.6.2"
@@ -367,6 +399,7 @@ name = "plex-media-ingest"
 version = "0.1.0"
 dependencies = [
  "clap",
+ "infer",
  "inquire",
  "log",
  "serde",
@@ -583,6 +616,12 @@ version = "0.2.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "711b9620af191e0cdc7468a8d14e709c3dcdb115b36f838e601583af800a370a"
 
+[[package]]
+name = "uuid"
+version = "1.5.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "88ad59a7560b41a70d191093a945f0b87bc1deeda46fb237479708a1d6b6cdfc"
+
 [[package]]
 name = "walkdir"
 version = "2.4.0"

Cargo.toml

@@ -7,6 +7,7 @@ edition = "2021"
 
 [dependencies]
 clap = { version = "4.4.7", features = ["derive"] }
+infer = "0.15.0"
 inquire = "0.6.2"
 log = "0.4.20"
 serde = { version = "1.0.190", features = ["derive"] }

src/directory.rs

@@ -1,16 +1,16 @@
-use std::path::PathBuf;
+use std::{path::PathBuf, fs::{self, DirEntry}, error::Error};
 
-use walkdir::{DirEntry, WalkDir};
+use crate::media::handle_media;
 
-fn is_not_hidden(entry: &DirEntry) -> bool {
+/*fn is_not_hidden(entry: &DirEntry) -> bool {
     entry
         .file_name()
         .to_str()
-        .map(|s| entry.depth() == 0 || !s.starts_with("."))
+        .map(|s| entry.depth() == 0 || (!s.starts_with(".") && !s.starts_with("@"))) // todo!: Allow ignored chars to be configured, here, @ is QNAP special folders
         .unwrap_or(false)
 }
 
-pub fn list_files(path: PathBuf) -> Vec<PathBuf>{
+pub fn walk_path(path: PathBuf) -> Vec<PathBuf> {
     let mut entries: Vec<PathBuf> = vec![];
     WalkDir::new(path)
         .into_iter()
@@ -18,4 +18,52 @@ pub fn list_files(path: PathBuf) -> Vec<PathBuf>{
         .filter_map(|v| v.ok())
         .for_each(|x| entries.push(x.into_path()));
     entries
+}*/
+
+pub fn search_path(path: PathBuf) -> Result<(), Box<dyn Error>> {
+    let entries = fs::read_dir(path)?;
+    let mut files: Vec<DirEntry> = Vec::new();
+    let mut dirs: Vec<DirEntry> = Vec::new();
+
+    // Put all files and folders in corresponding vectors
+    for entry in entries {
+        if let Ok(entry) = entry {
+            if let Ok(file_type) = entry.file_type() {
+                if file_type.is_dir() {
+                    dirs.push(entry);
+                } else if file_type.is_file() {
+                    files.push(entry);
+                }
+            }
+        }
+    }
+
+    if dirs.len() == 0 {
+        // No folders present, assuming there are only distinct media files
+        for file in files {
+            handle_media(file);
+        }
+    }
+
+    Ok(())
 }
+
+/*
+Look at current directory:
+    Only directories, no media files ->
+        Media must be in subfolders, look at name of folder (in case media file has cryptic name) and traverse into it, look at media files
+
+    Media files present, folders as well ->
+        Treat media as media to add, traverse into subfolders and look for eventual extra content
+
+    Media file(s), but no folders present ->
+        Treat media as media to add
+*/
+
+/*
+Use folder/file name as name to look up on tmdb (replace . with ' ' till first occurence of non alphanumeric symbol ([]())) (For shows, look for SxxEyy or similar tokens, if single file assume movie by default)
+If there is a token with only 4 digits (and maybe parantheses), assume this is a year, add it to tmdb search, retry search without 'year' if result is empty
+Show user file with title(s) found on tmdb, make user select one
+Remember selection and look in current folder for extra content (deleted scenes, trailers, featurettes) -> show user what files were found and have them select which files they want
+For each selection show file name plus try to match to extras category, show user selection of which kind of extra it is, then allow them to enter a arbitary name (prefill from file if possible)
+*/

src/main.rs

@@ -1,5 +1,6 @@
 mod config;
 mod directory;
+mod media;
 
 use log::*;
 use clap::Parser;
@@ -62,10 +63,13 @@ fn main() {
         args.path.unwrap()
     };
 
-    let files = directory::list_files(search_path);
+    //let files = directory::walk_path(search_path);
+    directory::search_path(search_path).unwrap();
 
-    for file in files {
+    /*for file in files.clone() {
         info!("Found: {}", file.to_str().unwrap());
-    }
+    }*/
+
+    //search_media(files).unwrap();
 
 }

src/media.rs

@@ -0,0 +1,109 @@
+use std::{path::PathBuf, error::Error, io::Read, fs::{File, DirEntry}, cmp};
+use infer;
+use log::{info, warn, error, trace, debug};
+
+#[derive(Debug)]
+struct MediaName {
+    name: String,
+    year: String
+}
+
+fn get_file_header(path: PathBuf) -> Result<Vec<u8>, Box<dyn Error>> {
+    let f = File::open(path)?;
+
+    let limit = f
+        .metadata()
+        .map(|m| cmp::min(m.len(), 8192) as usize + 1)
+        .unwrap_or(0);
+    let mut bytes = Vec::with_capacity(limit);
+    f.take(8192).read_to_end(&mut bytes)?;
+    Ok(bytes)
+}
+
+fn token_year_likely(t: &&str) -> bool {
+    if t.len() == 6 &&
+        (t.starts_with('[') && t.ends_with(']')) ||
+        (t.starts_with('(') && t.ends_with(')')) ||
+        (t.starts_with('{') && t.ends_with('}'))   
+    {
+        return true;    
+    }
+    return false
+}
+
+fn token_valid(t: &&str) -> bool {
+    if
+        t.eq_ignore_ascii_case("dvd") ||
+        t.eq_ignore_ascii_case("bluray") ||
+        t.eq_ignore_ascii_case("webrip") ||
+        t.eq_ignore_ascii_case("web") ||
+        t.eq_ignore_ascii_case("uhd") ||
+        t.eq_ignore_ascii_case("hd") ||
+        t.eq_ignore_ascii_case("tv") ||
+        t.eq_ignore_ascii_case("tvrip") ||
+        t.eq_ignore_ascii_case("1080p") ||
+        t.eq_ignore_ascii_case("1080i") ||
+        t.eq_ignore_ascii_case("2160p") ||
+        (t.len() != 6 && t.starts_with('[') && t.ends_with(']')) ||
+        (t.len() != 6 && t.starts_with('(') && t.ends_with(')')) ||
+        (t.len() != 6 && t.starts_with('{') && t.ends_with('}'))
+    {
+        return false;
+    }
+    true
+}
+
+fn find_media_name(file_name: String) -> MediaName {
+    let mut tokens: Vec<&str> = file_name.split(&['-', ' ', ':', '@', '.'][..]).filter(|t| token_valid(t)).collect();
+    trace!("Tokens are: {:#?}", tokens);
+
+    // Remove last token (file ext)
+    _ = tokens.pop();
+
+    let mut year = String::new();
+    let mut name = String::new();
+
+    for token in tokens {
+        if token_year_likely(&token)  {
+            year = token.strip_prefix(['(', '[', '{']).unwrap().strip_suffix([')', ']', '}']).unwrap().to_string();
+        } else if token.len() != 0 {
+            name.push_str(token);
+            name.push(' ');
+        }
+    }
+
+    // Remove last added space
+    name.pop();
+
+    let media_name = MediaName { name: name, year: year };
+    debug!("Name is now: {:#?}", media_name);
+    media_name
+}
+
+fn video_file_handler(entry: DirEntry) {
+    let path = entry.path();
+    info!("Found video file: {:#?}", path);
+
+    let file_name = path.file_name().unwrap_or_default();
+    trace!("File name is: {:#?}", file_name);
+
+    let name = find_media_name(file_name.to_str().unwrap_or_default().to_string());
+    todo!("Do TMDB API calls");
+}
+
+pub fn handle_media(entry: DirEntry) {
+    if entry.file_type().is_ok_and(|t| t.is_dir()) {
+        warn!("Directory passed to handle_media, {:#?} will be skipped", entry);
+        return
+    }
+
+    match get_file_header(entry.path()) {
+        Ok(header) => {
+            // Handle video files
+            if infer::is_video(&header) {
+                video_file_handler(entry);
+            }
+        },
+        Err(error) => error!("Can not get file header for {:#?}, Error: {:#?}", entry, error),
+    }    
+}