add config file, blacklists, whitelists, resolution cache

.gitignore

@@ -2,3 +2,4 @@
 **/*.rs.bk
 .idea
 *.iml
+rapblock.json

Cargo.lock

@@ -925,10 +925,10 @@ version = "0.1.0"
 dependencies = [
  "env_logger 0.6.0 (registry+https://github.com/rust-lang/crates.io-index)",
  "failure 0.1.5 (registry+https://github.com/rust-lang/crates.io-index)",
- "lazy_static 1.2.0 (registry+https://github.com/rust-lang/crates.io-index)",
  "log 0.4.6 (registry+https://github.com/rust-lang/crates.io-index)",
  "mpris 1.1.1 (registry+https://github.com/rust-lang/crates.io-index)",
  "percent-encoding 1.0.1 (registry+https://github.com/rust-lang/crates.io-index)",
+ "regex 1.1.0 (registry+https://github.com/rust-lang/crates.io-index)",
  "reqwest 0.9.10 (registry+https://github.com/rust-lang/crates.io-index)",
  "serde 1.0.88 (registry+https://github.com/rust-lang/crates.io-index)",
  "serde_derive 1.0.88 (registry+https://github.com/rust-lang/crates.io-index)",

Cargo.toml

@@ -12,6 +12,7 @@ serde = "1.0.88"
 serde_derive = "1.0.88"
 serde_json = "1.0"
 failure = "0.1.5"
-lazy_static = "1.2.0"
+#lazy_static = "1.2.0"
 log = "0.4"
 env_logger = "0.6"
+regex = "1.1.0"

README.md

@@ -0,0 +1,18 @@
+RapBlock
+========
+
+This is a daemon that auto-skips rap and hip-hop songs on Spotify via MPRIS.
+
+Music genre is determined using the MusicBrainz JSON API, as that is not reported via MPRIS.
+
+Compatibility
+-------------
+
+This daemon should work with any MPRIS compatible player, so long as it reports artist names 
+in the TrackChanged event metadata.
+
+Customization
+-------------
+
+You can easily modify this to skip e.g. folk, country or any other genre. Just change the
+`BAD_GENRES` list in `src/brainz.rs`.

rapblock.example.json

@@ -0,0 +1,17 @@
+{
+  "logging": "info",
+  "blacklist": {
+    "tag": [],
+    "tag_partial": ["hip-hop", "hip hop", "rap"],
+    "artist": ["Higher Brothers"]
+  },
+  "whitelist": {
+    "tag": [],
+    "artist": []
+  },
+  "artist_min_score": 95,
+  "max_artists_per_track": 3,
+  "player_find_interval_ms": 2500,
+  "cooldown_ms": 500,
+  "api_timeout_ms": 2000
+}

src/brainz.rs

@@ -2,6 +2,7 @@ use percent_encoding::{utf8_percent_encode, DEFAULT_ENCODE_SET};
 use failure::Error;
 use std::io::Read;
 use std::time::Duration;
+use regex::Regex;
 
 #[derive(Serialize, Deserialize, Debug)]
 struct MBArtistQueryResult {
@@ -27,15 +28,9 @@ struct MBArtist {
     tags: Option<Vec<Tag>>
 }
 
-lazy_static! {
-    static ref BAD_GENRES : Vec<&'static str> = vec![
-        "hip hop", "hip-hop", "hiphop", "rnb", "rap", "rapper",
-        "hardcore hip hop", "new york hip hop", "east coast hip hop",
-        "pop rap"
-    ];
-}
+const VERSION: &'static str = env!("CARGO_PKG_VERSION");
 
-pub fn artist_sucks(artist : &str) -> Result<bool, Error> {
+pub fn check_artist(config : &crate::Config, artist : &str) -> Result<bool, Error> {
     let query = utf8_percent_encode(&format!("\"{}\"", artist), DEFAULT_ENCODE_SET).to_string();
     let url = format!("https://musicbrainz.org/ws/2/artist?query={}&fmt=json&inc=tags", query);
 
@@ -43,11 +38,15 @@ pub fn artist_sucks(artist : &str) -> Result<bool, Error> {
 
     let mut resp = String::new();
 
+    let ua = format!("Bad Song Skipper / {v} ( https://git.ondrovo.com/MightyPork/rapblock )", v=VERSION);
+
+    debug!("Using UA: {}", ua);
+
     reqwest::Client::builder()
-        .timeout(Duration::from_millis(2000))
+        .timeout(Duration::from_millis(config.api_timeout_ms))
         .build()?
         .get(&url)
-        .header(reqwest::header::USER_AGENT, "Bad Spotify Song Skipper/0.0.1 ( ondra@ondrovo.com )")
+        .header(reqwest::header::USER_AGENT, ua.as_str())
         .header(reqwest::header::ACCEPT, "text/json")
         .send()?
         .read_to_string(&mut resp)?;
@@ -64,22 +63,64 @@ pub fn artist_sucks(artist : &str) -> Result<bool, Error> {
         warn!("No results!");
         return Err(failure::err_msg("Artist not found"));
     } else {
-        info!("Got {} results", result.count);
+        info!("Got {} results, checking where score >= {}", result.count, config.artist_min_score);
         let artists = result.artists.as_ref().unwrap();
-        if artists[0].tags.is_some() {
-            let tags = artists[0].tags.as_ref().unwrap();
-            let as_vec : Vec<&String> = tags.iter().map(|t| &t.name).collect();
-            debug!("First artist has tags: {:?}", tags);
-            for tag in as_vec {
-                if BAD_GENRES.contains(&tag.as_str()) {
-                    info!("Found a bad tag: {}", tag);
-                    return Ok(true);
+
+        let mut confidence = false;
+        let mut passed = true;
+
+        'artists: for (an, a) in artists.iter().enumerate() {
+            if a.score >= config.artist_min_score {
+                if a.tags.is_some() {
+                    let tags = a.tags.as_ref().unwrap();
+                    let as_vec: Vec<&String> = tags.iter().map(|t| &t.name).collect();
+
+                    info!("Artist #{} - \"{}\" has tags: {:?}", an+1, a.name, as_vec);
+                    'tags: for tag in as_vec {
+                        if config.whitelist.tag.contains(tag) {
+                            info!("+ Whitelisted tag \"{}\"", tag);
+                            continue 'tags;
+                        }
+
+                        let mut blacklisted = false;
+                        if config.blacklist.tag.contains(&tag) {
+                            info!("- Blacklisted tag \"{}\"", tag);
+                            blacklisted = true;
+                        } else {
+                            'blacklist:
+                            for t in config.blacklist.tag_partial.iter() {
+                                let re = Regex::new(&format!(r"\b{}\b", regex::escape(t)));
+
+                                let matches = match re {
+                                    Ok(re) => re.is_match(tag),
+                                    Err(_) => tag.contains(t)
+                                };
+
+                                if matches {
+                                    info!("- Blacklisted tag \"{}\" - due to substring \"{}\"", tag, t);
+                                    blacklisted = true;
+                                    break 'blacklist;
+                                }
+                            }
+                        }
+
+                        if blacklisted {
+                            confidence = true;
+                            passed = false;
+                            break 'artists;
                         }
                     }
                     info!("All tags OK");
-            return Ok(false);
+                    confidence = true;
+                } else {
+                    warn!("Artist #{} - \"{}\" has no tags, can't determine genre", an+1, a.name);
+                }
+            }
+        }
+
+        if confidence {
+            return Ok(passed);
         } else {
-            warn!("No tags, can't determine genre");
             return Err(failure::err_msg("Artist found, but has no tags"));
         }
     }

src/main.rs

@@ -1,19 +1,140 @@
 #[macro_use] extern crate serde_derive;
-#[macro_use] extern crate lazy_static;
+//#[macro_use] extern crate lazy_static;
 #[macro_use] extern crate log;
+#[macro_use] extern crate failure;
 
 mod brainz;
 
 use mpris::{PlayerFinder,Event};
 use failure::Error;
 use std::time::Duration;
+use std::collections::HashSet;
+use std::env;
+use std::path::Path;
+use std::fs::File;
+use std::io::Read;
+use std::collections::HashMap;
 
-const DELAY_FIND_PLAYER : Duration = Duration::from_millis(1000);
+#[derive(Serialize, Deserialize, Debug)]
+#[serde(default)]
+pub struct BlacklistConf {
+    pub tag: Vec<String>,
+    pub tag_partial: Vec<String>,
+    pub artist: Vec<String>,
+}
+
+impl Default for BlacklistConf {
+    fn default() -> Self {
+        Self {
+            tag: vec![],
+            artist: vec![],
+            tag_partial: vec![
+                "hip-hop".to_owned(),
+                "hip hop".to_owned(),
+                "hiphop".to_owned(),
+                "rap".to_owned()
+            ],
+        }
+    }
+}
+
+#[derive(Serialize, Deserialize, Debug)]
+#[serde(default)]
+pub struct WhiteList {
+    pub tag: Vec<String>,
+    pub artist: Vec<String>,
+}
+
+impl Default for WhiteList {
+    fn default() -> Self {
+        Self {
+            tag: vec![],
+            artist: vec![]
+        }
+    }
+}
+
+#[derive(Serialize, Deserialize, Debug)]
+#[serde(default)]
+pub struct Config {
+    /// Logging - trace, debug, (info), warning, error
+    pub logging: String,
+    /// Blacklists
+    pub blacklist: BlacklistConf,
+    /// Whitelist (overrides blacklist)
+    pub whitelist: WhiteList,
+    /// Min MusicBrainz search score for artist look-up
+    pub artist_min_score : i32,
+    /// Max nr of artists to check per track
+    pub max_artists_per_track : u64,
+    /// Interval in which the daemon probes DBUS for open MPRIS channels
+    pub player_find_interval_ms: u64,
+    /// Delay after a skip or allow, e.g. to prevent infinite skip chain when someone starts a rap playlist
+    pub cooldown_ms: u64,
+    /// Music Brainz API access timeout
+    pub api_timeout_ms: u64,
+    /// Allow playing songs from artists we couldn't verify
+    pub allow_by_default: bool,
+}
+
+impl Default for Config {
+    fn default() -> Self {
+        Self {
+            logging: "info".to_owned(),
+            blacklist: BlacklistConf::default(),
+            whitelist: WhiteList::default(),
+            artist_min_score: 95,
+            max_artists_per_track: 3,
+            player_find_interval_ms: 2500,
+            cooldown_ms: 1000,
+            api_timeout_ms: 2000,
+            allow_by_default: true
+        }
+    }
+}
+
+const CONFIG_FILE: &'static str = "rapblock.json";
+
+fn load_config() -> Result<Config, Error> {
+    let config_file_path = env::current_dir()?.join(CONFIG_FILE);
+    let buf = read_file(config_file_path)?;
+
+    let config : Config = serde_json::from_str(&buf)?;
+
+    // Validations
+    match config.logging.as_ref() {
+        "info" | "debug" | "trace" | "warning" | "error" => (),
+        _ => bail!("Invalid value for \"logging\""),
+    };
+
+    Ok(config)
+}
+
+pub fn read_file<P: AsRef<Path>>(path: P) -> Result<String, Error> {
+    let path = path.as_ref();
+    let mut file = File::open(path)?;
+
+    let mut buf = String::new();
+    file.read_to_string(&mut buf)?;
+    Ok(buf)
+}
 
 fn main() -> Result<(), Error> {
-    let env = env_logger::Env::default().default_filter_or("info");
+    let config = match load_config() {
+        Ok(c) => c,
+        Err(e) => {
+            eprintln!("Could not load config from \"{}\": {}", CONFIG_FILE, e);
+            Config::default()
+        }
+    };
+
+    let env = env_logger::Env::default().default_filter_or(&config.logging);
     env_logger::Builder::from_env(env).init();
 
+    debug!("Loaded config:\n{}", serde_json::to_string_pretty(&config)?);
+
+    let mut artist_cache = HashMap::<String, bool>::new();
+
     'main_loop:
     loop {
         let player = PlayerFinder::new()
@@ -27,7 +148,7 @@ fn main() -> Result<(), Error> {
             if events.is_err() {
                 error!("Could not start event stream!");
                 // add a delay so we don't run too hot here
-                ::std::thread::sleep(DELAY_FIND_PLAYER);
+                ::std::thread::sleep(Duration::from_millis(config.player_find_interval_ms));
                 continue 'main_loop;
             }
 
@@ -52,39 +173,77 @@ fn main() -> Result<(), Error> {
                                     continue 'event_loop;
                                 }
 
-                                'artists_loop:
-                                for ar in metadata.artists().into_iter().chain(metadata.artists()) {
-                                    for a in ar {
-                                        info!("Checking artist: {}", a);
+                                let mut artists = HashSet::new();
 
-                                        let verdict = brainz::artist_sucks(&a);
-                                        match verdict {
-                                            Ok(verdict) => {
-                                                if verdict {
-                                                    if player.can_go_next().unwrap_or(false) {
-                                                        info!(">>>>>> SKIP >>>>>>");
-                                                        if player.next().is_err() {
-                                                            break 'artists_loop;
+                                if let Some(aa) = metadata.artists() {
+                                    for a in aa { artists.insert(a); }
                                 }
-                                                    } else {
-                                                        info!("<><><> STOP <><><>");
-                                                        if player.pause().is_err() {
+                                if let Some(aa) = metadata.album_artists() {
+                                    for a in aa { artists.insert(a); }
+                                }
+
+                                let mut skip = false;
+                                let mut confidence = false;
+                                'artists_loop:
+                                for (an, a) in artists.iter().take(config.max_artists_per_track as usize).enumerate() {
+                                    info!("Checking artist #{}: {}", an+1, a);
+                                    if let Some(resolution) = artist_cache.get(a.as_str()) {
+                                        confidence = true;
+                                        if !resolution {
+                                            info!("~ result cached: BAD");
+                                            skip = true;
                                             break 'artists_loop;
                                         }
+                                        info!("~ result cached: GOOD");
+                                        continue 'artists_loop;
+                                    }
+
+                                    if config.whitelist.artist.contains(a) {
+                                        info!("+ Whitelisted artist!");
+                                        // there may be other co-artists that spoil the song -> don't break yet
+                                        artist_cache.insert(a.to_string(), true);
+                                        confidence = true;
+                                        continue 'artists_loop;
+                                    }
+
+                                    if config.blacklist.artist.contains(a) {
+                                        info!("- Blacklisted artist!");
+                                        skip = true;
+                                        artist_cache.insert(a.to_string(), false);
+                                        confidence = true;
+                                        break 'artists_loop;
                                     }
-                                                    // we add a delay here to prevent going insane on rap playlists
-                                                    ::std::thread::sleep(Duration::from_millis(1000));
+
+                                    let verdict = brainz::check_artist(&config, &a);
+                                    match verdict {
+                                        Ok(allow) => {
+                                            confidence = true;
+                                            artist_cache.insert(a.to_string(), allow);
+                                            if allow {
+                                                info!("Artist passed");
                                             } else {
-                                                    info!("Good artist, let it play");
+                                                skip = true;
+                                                break 'artists_loop;
                                             }
                                         },
                                         Err(e) => {
-                                                error!("Something went wrong: {}", e);
-                                                info!("Letting to play");
+                                            warn!("Something went wrong: {}", e);
+                                            // probably no tags, or not found - use the default action
+                                            artist_cache.insert(a.to_string(), config.allow_by_default);
                                         }
                                     }
                                 }
+
+                                if skip || (!confidence && !config.allow_by_default) {
+                                    info!(">>>>>> SKIP : {} >>>>>>\n", title);
+                                    if player.next().is_err() {
+                                        break 'event_loop;
                                     }
+                                } else {
+                                    info!("Let it play...\n");
+                                }
+
+                                ::std::thread::sleep(Duration::from_millis(config.cooldown_ms));
                             },
                             _ => {
                                 debug!("Event not handled.");
@@ -102,7 +261,7 @@ fn main() -> Result<(), Error> {
             debug!("No player found, waiting...");
         }
 
-        ::std::thread::sleep(DELAY_FIND_PLAYER);
+        ::std::thread::sleep(Duration::from_millis(config.player_find_interval_ms));
     }
 }