rate limiting wip

.gitignore

@@ -1,3 +1,4 @@
 /target
 **/*.rs.bk
 Cargo.lock
+/doc

Cargo.toml

@@ -2,7 +2,7 @@
 name = "riven"
 version = "0.0.1"
 authors = ["Mingwei Samuel <mingwei.samuel@gmail.com>"]
-description = "RiotApi library (wip)"
+description = "Riot API Library (WIP)"
 license = "LGPL-3.0"
 edition = "2018"
 exclude = [
@@ -12,3 +12,6 @@ exclude = [
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
 
 [dependencies]
+async-std = "0.99"
+reqwest = { version = "0.10.0-alpha.1", features = [ "gzip", "json" ] }
+serde = "^1.0"

src/consts/mod.rs

@@ -0,0 +1 @@
+pub mod region;

src/consts/region.rs

@@ -0,0 +1,73 @@
+#[derive(Debug, PartialEq, Eq, Hash, PartialOrd, Ord)]
+pub struct Region<'a> {
+    pub key: &'a str,
+    pub platform: &'a str,
+}
+
+macro_rules! regions {
+    (
+        $(
+            $key:ident => $plat:expr ;
+        )*
+    ) => {
+        $(
+            const $key: &'static Region<'static> = &Region {
+                key: stringify!($key),
+                platform: $plat,
+            };
+        )*
+
+        #[doc="Get region by name."]
+        #[doc="# Arguments"]
+        #[doc="* `name` - Case-insensitive ASCII string to match Regions' `key` or `playform`."]
+        #[doc="# Returns"]
+        #[doc="`Some(&Region)` if match found, `None` if no match found."]
+        #[allow(unreachable_patterns)]
+        pub fn get(name: &str) -> Option<&Region> {
+            match &*name.to_ascii_uppercase() {
+                $(
+                    stringify!($key) | $plat => Some(Self::$key),
+                )*
+                _ => None
+            }
+        }
+    }
+}
+
+impl Region<'_> {
+    // Is this stupid?
+    regions! {
+        BR       => "BR1";
+        EUNE     => "EUN1";
+        EUW      => "EUW1";
+        NA       => "NA1";
+        KR       => "KR";
+        LAN      => "LA1";
+        LAS      => "LA2";
+        OCE      => "OC1";
+        RU       => "RU";
+        TR       => "TR1";
+        JP       => "JP1";
+        PBE      => "PBE1";
+        AMERICAS => "AMERICAS";
+        EUROPE   => "EUROPE";
+        ASIA     => "ASIA";
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn test_basic() {
+        assert_eq!("BR1", Region::BR.platform);
+    }
+
+    #[test]
+    fn test_get() {
+        assert_eq!(Some(Region::AMERICAS), Region::get("amEricAs"));
+        assert_eq!(Some(Region::NA), Region::get("na1"));
+        assert_eq!(None, Region::get("LA"));
+    }
+}

src/lib.rs

@@ -1,6 +1,9 @@
-mod req;
+#![allow(dead_code)] // TODO REMOVE
 
-use req::rate_limit;
+pub mod consts;
+
+mod req;
+mod riot_api_config;
 
 #[cfg(test)]
 mod tests {

src/req/mod.rs

@@ -1,2 +1,5 @@
 pub mod rate_limit;
-pub mod token_bucket;
+pub mod rate_limit_type;
+pub mod token_bucket;
+pub mod regional_requester;
+pub mod requester_manager;

src/req/rate_limit.rs

@@ -1,6 +1,83 @@
-use std::time::Duration;
+use std::cmp;
+use std::time::{
+    Duration,
+    Instant,
+};
+use std::sync::{
+    RwLock,
+};
 
-pub trait RateLimit {
+use super::token_bucket::{
-    fn get_retry_after_delay(&self) -> Duration;
+    TokenBucket,
+    VectorTokenBucket,
+};
+use super::rate_limit_type::RateLimitType;
+
+pub struct RateLimit {
+    rate_limit_type: RateLimitType,
+    // Buckets for this rate limit (synchronized).
+    // Almost always read, written only when rate limit rates are updated
+    // from API response.
+    // TODO: Question of writer starvation.
+    buckets: RwLock<Vec<VectorTokenBucket>>,
+    // Set to when we can retry if a retry-after header is received.
+    retry_after: Option<Instant>,
 }
 
+impl RateLimit {
+    /// Header specifying which RateLimitType caused a 429.
+    const HEADER_XRATELIMITTYPE: &'static str = "X-Rate-Limit-Type";
+    /// Header specifying retry after time in seconds after a 429.
+    const HEADER_RETRYAFTER: &'static str = "Retry-After";
+
+    pub fn new(rate_limit_type: RateLimitType) -> Self {
+        let initial_bucket = VectorTokenBucket::new(Duration::from_secs(1), 1);
+        RateLimit {
+            rate_limit_type: rate_limit_type,
+            // Rate limit before getting from response: 1/s.
+            buckets: RwLock::new(vec![initial_bucket]),
+            retry_after: None,
+        }
+    }
+
+    pub fn get_both_or_delay(app_rate_limit: &Self, method_rate_limit: &Self) -> Option<Duration> {
+        // Check retry after.
+        let retry_after_delay = app_rate_limit.get_retry_after_delay()
+            .and_then(|a| method_rate_limit.get_retry_after_delay().map(|m| cmp::max(a, m)));
+        if retry_after_delay.is_some() {
+            return retry_after_delay
+        }
+        // Check buckets.
+        let app_buckets = app_rate_limit.buckets.read().unwrap();
+        let method_buckets = app_rate_limit.buckets.read().unwrap();
+        for bucket in app_buckets.iter().chain(method_buckets.iter()) {
+            let delay = bucket.get_delay();
+            if delay.is_some() {
+                return delay;
+            }
+        }
+        // Success.
+        for bucket in app_buckets.iter().chain(method_buckets.iter()) {
+            bucket.get_tokens(1);
+        }
+        None
+    }
+
+    pub fn get_retry_after_delay(&self) -> Option<Duration> {
+        self.retry_after.and_then(|i| Instant::now().checked_duration_since(i))
+    }
+
+    pub fn on_response(&self, _response: &reqwest::Response) {
+        unimplemented!();
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    fn send_sync() {
+        fn is_send_sync<T: Send + Sync>() {}
+        is_send_sync::<RateLimit>();
+    }
+}

src/req/rate_limit_type.rs

@@ -0,0 +1,27 @@
+pub enum RateLimitType {
+    Application,
+    Method,
+}
+
+impl RateLimitType {
+    pub fn type_name(self) -> &'static str {
+        match self {
+            Self::Application => "application",
+            Self::Method => "method",
+        }
+    }
+
+    pub fn limit_header(self) -> &'static str {
+        match self {
+            Self::Application => "X-App-Rate-Limit",
+            Self::Method => "X-Method-Rate-Limit",
+        }
+    }
+
+    pub fn count_header(self) -> &'static str {
+        match self {
+            Self::Application => "X-App-Rate-Limit-Count",
+            Self::Method => "X-Method-Rate-Limit-Count",
+        }
+    }
+}

src/req/regional_requester.rs

@@ -0,0 +1,127 @@
+use std::collections::HashMap;
+use std::future::Future;
+use std::sync::RwLock;
+
+use async_std::task;
+use reqwest::{
+    Client,
+    StatusCode,
+};
+use serde::de::DeserializeOwned;
+
+use super::rate_limit::RateLimit;
+use super::rate_limit_type::RateLimitType;
+use crate::riot_api_config::RiotApiConfig;
+use crate::consts::region::Region;
+
+pub struct RegionalRequester<'a> {
+    /// Configuration settings.
+    riot_api_config: &'a RiotApiConfig<'a>,
+    /// Client for making requests.
+    client: &'a Client,
+
+    /// Represents the app rate limit.
+    app_rate_limit: RateLimit,
+    /// Represents method rate limits.
+    method_rate_limits: HashMap<&'a str, RateLimit>,
+}
+
+impl <'a> RegionalRequester<'a> {
+    /// Request header name for the Riot API key.
+    const RIOT_KEY_HEADER: &'static str = "X-Riot-Token";
+
+    /// HttpStatus codes that are considered a success, but will return None.
+    const NONE_STATUS_CODES: [u16; 3] = [ 204, 404, 422 ];
+
+
+    pub fn new(riot_api_config: &'a RiotApiConfig<'a>, client: &'a Client) -> RegionalRequester<'a> {
+        RegionalRequester {
+            riot_api_config: riot_api_config,
+            client: client,
+            app_rate_limit: RateLimit::new(RateLimitType::Application),
+            method_rate_limits: HashMap::new(),
+        }
+    }
+
+    pub async fn get<T: DeserializeOwned>(
+        &mut self, method_id: &'a str, relative_url: &'_ str,
+        region: &'_ Region<'_>, query: &[(&'_ str, &'_ str)]) -> Result<Option<T>, reqwest::Error> {
+
+        let mut attempts: u8 = 0;
+        for _ in 0..=self.riot_api_config.retries {
+            attempts += 1;
+
+            // Rate limiting.
+            let app_rate_limit = &self.app_rate_limit;
+            let method_rate_limit = self.method_rate_limits.entry(method_id)
+                .or_insert_with(|| RateLimit::new(RateLimitType::Method));
+
+            while let Some(delay) = RateLimit::get_both_or_delay(app_rate_limit, method_rate_limit) {
+                task::sleep(delay).await;
+            }
+
+            // Send request.
+            let url = &*format!("https://{}.api.riotgames.com{}", region.platform, relative_url);
+            let result = self.client.get(url)
+                .header(Self::RIOT_KEY_HEADER, self.riot_api_config.api_key)
+                .query(query)
+                .send()
+                .await;
+
+            let response = match result {
+                Err(e) => return Err(e),
+                Ok(r) => r,
+            };
+
+            // Update rate limits (if needed).
+            app_rate_limit.on_response(&response);
+            method_rate_limit.on_response(&response);
+
+            // Handle response.
+            let status = response.status();
+            // Success, return None.
+            if Self::is_none_status_code(&status) {
+                return Ok(None);
+            }
+            // Success, return a value.
+            if status.is_success() {
+                let value = response.json::<T>().await;
+                return match value {
+                    Err(e) => Err(e),
+                    Ok(v) => Ok(Some(v)),
+                }
+            }
+            // Retryable.
+            if StatusCode::TOO_MANY_REQUESTS == status || status.is_server_error() {
+                continue;
+            }
+            // Failure (non-retryable).
+            if status.is_client_error() {
+                break;
+            }
+            panic!("NOT HANDLED: {}!", status);
+        }
+        // TODO: return error.
+        panic!("FAILED AFTER {} ATTEMPTS!", attempts);
+    }
+
+    pub fn get2<T: 'a + DeserializeOwned>(&'a mut self, method_id: &'a str, relative_url: &'a str,
+        region: &'a Region<'_>, query: &'a [(&'a str, &'a str)]) -> impl Future<Output = Result<Option<T>, reqwest::Error>> + 'a {
+
+        self.get(method_id, relative_url, region, query)
+    }
+
+    fn is_none_status_code(status: &StatusCode) -> bool {
+        Self::NONE_STATUS_CODES.contains(&status.as_u16())
+    }
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    fn send_sync() {
+        fn is_send_sync<T: Send + Sync>() {}
+        is_send_sync::<RegionalRequester>();
+    }
+}

src/req/requester_manager.rs

@@ -0,0 +1,21 @@
+use std::collections::HashMap;
+
+use reqwest::{
+    Client,
+};
+
+use super::regional_requester::RegionalRequester;
+use crate::riot_api_config::RiotApiConfig;
+use crate::consts::region::Region;
+
+// pub struct RequesterManager<'a> {
+//     /// Configuration settings.
+//     riot_api_config: &'a RiotApiConfig<'a>,
+//     /// Client for making requests.
+//     client: &'a Client,
+
+//     /// Represents the app rate limit.
+//     app_rate_limit: RateLimit,
+//     /// Represents method rate limits.
+//     method_rate_limits: HashMap<&'a str, RateLimit>,
+// }

src/req/token_bucket.rs

@@ -1,18 +1,19 @@
 use std::collections::VecDeque;
 use std::time::{Duration, Instant};
+use std::sync::{Mutex, MutexGuard};
 
 pub trait TokenBucket {
     /// Get the duration til the next available token, or 0 duration if a token is available.
     /// # Returns
     /// Duration or 0 duration.
-    fn get_delay(&mut self) -> Duration;
+    fn get_delay(&self) -> Option<Duration>;
 
     /// Gets n tokens, regardless of whether they are available.
     /// # Parameters
     /// * `n` - Number of tokens to take.
     /// # Returns
     /// True if the tokens were obtained without violating limits, false otherwise.
-    fn get_tokens(&mut self, n: usize) -> bool;
+    fn get_tokens(&self, n: usize) -> bool;
 
     /// Get the duration of this bucket.
     /// # Returns
@@ -25,54 +26,61 @@ pub trait TokenBucket {
     fn get_total_limit(&self) -> usize;
 }
 
-struct VectorTokenBucket {
+pub struct VectorTokenBucket {
     /// Duration of this TokenBucket.
     duration: Duration,
     // Total tokens available from this TokenBucket.
     total_limit: usize,
-    // Record of timestamps.
+    // Record of timestamps (synchronized).
-    timestamps: VecDeque<Instant>,
+    timestamps: Mutex<VecDeque<Instant>>,
 }
 
 impl VectorTokenBucket {
-    fn create(duration: Duration, total_limit: usize) -> Self {
+    pub fn new(duration: Duration, total_limit: usize) -> Self {
         VectorTokenBucket {
             duration: duration,
             total_limit: total_limit,
-            timestamps: VecDeque::new(),
+            timestamps: Mutex::new(VecDeque::new()),
         }
     }
 
-    fn update_state(&mut self) {
+    fn update_get_timestamps(&self) -> MutexGuard<VecDeque<Instant>> {
+        let mut timestamps = self.timestamps.lock().unwrap();
         let cutoff = Instant::now() - self.duration;
-        while self.timestamps.back().map_or(false, |ts| ts < &cutoff) {
+        while timestamps.back().map_or(false, |ts| ts < &cutoff) {
-            self.timestamps.pop_back();
+            timestamps.pop_back();
         }
+        return timestamps;
     }
 }
 
 impl TokenBucket for VectorTokenBucket {
 
-    fn get_delay(&mut self) -> Duration {
+    fn get_delay(&self) -> Option<Duration> {
-        self.update_state();
+        let timestamps = self.update_get_timestamps();
-        if self.timestamps.len() < self.total_limit {
+
-            Duration::new(0, 0)
+        if timestamps.len() < self.total_limit {
+            None
         }
         else {
-            let ts = *self.timestamps.get(self.total_limit - 1).unwrap();
+            // Timestamp that needs to be popped before
-            Instant::now().saturating_duration_since(ts)
+            // we can enter another timestamp.
+            let ts = *timestamps.get(self.total_limit - 1).unwrap();
+            Instant::now().checked_duration_since(ts)
+                .and_then(|passed_dur| self.duration.checked_sub(passed_dur))
         }
     }
 
-    fn get_tokens(&mut self, n: usize) -> bool {
+    fn get_tokens(&self, n: usize) -> bool {
-        self.update_state();
+        let mut timestamps = self.update_get_timestamps();
+
         let now = Instant::now();
 
-        self.timestamps.reserve(n);
+        timestamps.reserve(n);
         for _ in 0..n {
-            self.timestamps.push_front(now);
+            timestamps.push_front(now);
         }
-        self.timestamps.len() <= self.total_limit
+        timestamps.len() <= self.total_limit
     }
 
     fn get_bucket_duration(&self) -> Duration {
@@ -86,6 +94,13 @@ impl TokenBucket for VectorTokenBucket {
 
 #[cfg(test)]
 mod tests {
+    use super::*;
+
+    fn send_sync() {
+        fn is_send_sync<T: Send + Sync>() {}
+        is_send_sync::<VectorTokenBucket>();
+    }
+
     #[test]
     fn it_works() {
         assert_eq!(2 + 2, 4);

src/riot_api_config.rs

@@ -0,0 +1,4 @@
+pub struct RiotApiConfig<'a> {
+    pub api_key: &'a str,
+    pub retries: u8,
+}