//! `evtclib` is a crate aiming to provide utility functions to parse and work //! with `.evtc` reports generated by arcdps. //! //! # About evtc Files //! //! evtc files are files generated by the (inofficial) arcdps addon to Guild Wars 2. They contain //! metadata about a fight in the game, such as the boss's name (if it was a raid or fractal boss), //! the participants, and a stripped-down log of the complete fight. //! //! There are other programs (such as //! [GW2-Elite-Insights-Parser](https://github.com/baaron4/GW2-Elite-Insights-Parser/)) and //! websites (such as [dps.report](https://dps.report)) which allow you to generate reports from //! evtc files. //! //! A common way to store and distribute evtc files is to zip them to either a `.evtc.zip` (old //! way) or a `.zevtc` (new way). evtclib uses [`zip`](https://crates.io/crates/zip) to read them, //! prodiving the [`raw::parse_zip`][raw::parse_zip] convenience function. //! //! # Crate Structure //! //! The crate consists of two main parts: The [`raw`][raw] parser, which is used to read structured //! data from binary input streams, and the higher-level abstrations provided in the root and //! [`event`][event] submodules. //! //! Additionally, there are some defintions (such as IDs for various game items) in the //! [`gamedata`][gamedata] module. //! //! The main structs that you should be dealing with are the [`Log`][Log] and its components, such //! as [`Event`][Event] and [`Agent`][Agent]. //! //! # Workflow //! //! `evtclib` provides two convenience functions to obtain a [`Log`][Log]: //! //! If you have a stream (that is, something that is [`Read`][std::io::Read] + //! [`Seek`][std::io::Seek]), you can use [`process_stream`][process_stream] to obtain a //! [`Log`][Log] by reading from the stream. //! //! If your evtc is saved in a file, you can use [`process_file`][process_file] to obtain a [`Log`] //! from it. This will also ensure that the buffering is set up correctly, to avoid unnecessary //! system calls. //! //! Both of those functions require the reader to be seekable, as that is what we need for zip //! archive support. If you cannot provide that, or if you need finer grained control for other //! reasons, you can use either [`raw::parse_file`][raw::parse_file] or //! [`raw::parse_zip`][raw::parse_zip] to obtain the low-level [`Evtc`][raw::Evtc] structure, and //! then turn it into a [`Log`][Log] by using [`process`][process]: //! //! ```no_run //! # fn main() -> Result<(), Box> { //! use evtclib::{Compression, Log}; //! use std::fs::File; //! // Preferred: //! let log: Log = evtclib::process_file("my_log.evtc", Compression::None)?; //! //! // If you have a stream: //! let file = File::open("my_log.evtc")?; //! let log: Log = evtclib::process_stream(file, Compression::None)?; //! //! // If you really need to do it manually: //! // Open a file for processing //! let file = File::open("my_log.evtc")?; //! // Parse the raw content of the file //! let raw_log = evtclib::raw::parse_file(file)?; //! // Process the file to do the nitty-gritty low-level stuff done //! let log: Log = evtclib::process(&raw_log)?; //! //! // In all cases, you can now do work with the log //! for player in log.players() { //! println!("Player {} participated!", player.account_name()); //! } //! # Ok(()) //! # } //! ``` //! //! Make sure to take a look at the note on "Buffering" in the [parser //! module](raw/parser/index.html#buffering) in order to increase the speed of your application. //! //! # Writing evtc Files //! //! Currently, `evtclib` does not provide a way to output or modify evtc files. This is for two //! reasons: //! //! * The only sensible source for logs is the arcdps addon itself, most applications only consume //! them. //! * The library was needed for reading support, and writing support has never been a priority. //! //! While there are legitimate use cases for writing/modification support, they are currently not //! implemented (but might be in a future version). use std::convert::TryFrom; use std::marker::PhantomData; use getset::{CopyGetters, Getters}; use num_traits::FromPrimitive; use thiserror::Error; pub mod raw; pub mod event; pub use event::{Event, EventKind}; mod processing; pub use processing::{process, process_file, process_stream, Compression}; pub mod gamedata; use gamedata::Boss; pub use gamedata::{EliteSpec, Encounter, Profession}; pub mod analyzers; pub use analyzers::{Analyzer, Outcome}; /// Any error that can occur during the processing of evtc files. #[derive(Error, Debug)] pub enum EvtcError { /// Error for underlying parser errors. /// /// This should never be returned from [`process`][process], only from /// [`process_stream`][process_stream] and [`process_file`][process_file]. #[error("the file could not be parsed: {0}")] ParseError(#[from] raw::ParseError), /// Generic error for invalid data in the evtc file. #[error("invalid data has been provided")] InvalidData, /// The profession id is not known. /// /// The field contains the unknown profession id. #[error("invalid profession id: {0}")] InvalidProfession(u32), /// The elite specialization id is not known. /// /// The field contains the unknown elite specialization id. #[error("invalid elite specialization id: {0}")] InvalidEliteSpec(u32), /// The file contains invalid utf-8. #[error("utf8 decoding error: {0}")] Utf8Error(#[from] std::str::Utf8Error), } /// Player-specific agent data. /// /// Player agents are characters controlled by a player and as such, they contain data about the /// account and character used (name, profession), as well as the squad composition. /// /// Note that a `Player` is only the player character itself. Any additional entities that are /// spawned by the player (clones, illusions, banners, ...) are either a [`Character`][Character] /// or a [`Gadget`][Gadget]. #[derive(Debug, Clone, Hash, PartialEq, Eq, CopyGetters)] pub struct Player { /// The player's profession. #[get_copy = "pub"] profession: Profession, /// The player's elite specialization, if any is equipped. #[get_copy = "pub"] elite: Option, character_name: String, account_name: String, /// The subgroup the player was in. #[get_copy = "pub"] subgroup: u8, } impl Player { /// The player's character name. pub fn character_name(&self) -> &str { &self.character_name } /// The player's account name. /// /// This includes the leading colon and the 4-digit denominator. pub fn account_name(&self) -> &str { &self.account_name } } /// Gadget-specific agent data. /// /// Gadgets are entities that are spawned by certain skills. They are mostly inanimate objects that /// only exist to achieve a certain skill effect. /// /// Examples of this include the [banners](https://wiki.guildwars2.com/wiki/Banner) spawned by /// Warriors, but also skill effects like the roots created by /// [Entangle](https://wiki.guildwars2.com/wiki/Entangle) or the other objects in the arena. #[derive(Debug, Clone, Hash, PartialEq, Eq, CopyGetters)] pub struct Gadget { /// The id of the gadget. /// /// Note that gadgets do not have true ids and the id is generated "through a combination of /// gadget parameters". #[get_copy = "pub"] id: u16, name: String, } impl Gadget { /// The name of the gadget. pub fn name(&self) -> &str { &self.name } } /// Character-specific agent data. /// /// Characters are NPCs such as the bosses themselves, additional mobs that they spawn, but also /// friendly characters like Mesmer's clones and illusions, Necromancer minions, and so on. #[derive(Debug, Clone, Hash, PartialEq, Eq, CopyGetters)] pub struct Character { /// The id of the character. #[get_copy = "pub"] id: u16, name: String, } impl Character { /// The name of the character. pub fn name(&self) -> &str { &self.name } } /// The type of an agent. /// /// arcdps differentiates between three types of agents: [`Player`][Player], /// [`Character`][Character] and [`Gadget`][Gadget]. This enum unifies handling between them by /// allowing you to pattern match or use one of the accessor methods. /// /// The main way to obtain a `AgentKind` is by using the [`.kind()`][Agent::kind] method on an /// [`Agent`][Agent]. In cases where you already have a [`raw::Agent`][raw::Agent] available, you /// can also use the [`TryFrom`][TryFrom]/[`TryInto`][std::convert::TryInto] traits to convert a /// `raw::Agent` or `&raw::Agent` to a `AgentKind`: /// /// ```no_run /// # use evtclib::{AgentKind, raw}; /// use std::convert::TryInto; /// // Get a raw::Agent from somewhere /// let raw_agent: raw::Agent = panic!(); /// // Convert it /// let agent: AgentKind = raw_agent.try_into().unwrap(); /// ``` #[derive(Debug, Clone, Hash, PartialEq, Eq)] pub enum AgentKind { /// The agent is a player. /// /// The player-specific data is in the included [`Player`][Player] struct. Player(Player), /// The agent is a gadget. /// /// The gadget-specific data is in the included [`Gadget`][Gadget] struct. Gadget(Gadget), /// The agent is a character. /// /// The character-specific data is in the included [`Character`][Character] struct. Character(Character), } impl AgentKind { fn from_raw_character(raw_agent: &raw::Agent) -> Result { assert!(raw_agent.is_character()); let name = raw::cstr_up_to_nul(&raw_agent.name).ok_or(EvtcError::InvalidData)?; Ok(Character { id: raw_agent.prof as u16, name: name.to_str()?.to_owned(), }) } fn from_raw_gadget(raw_agent: &raw::Agent) -> Result { assert!(raw_agent.is_gadget()); let name = raw::cstr_up_to_nul(&raw_agent.name).ok_or(EvtcError::InvalidData)?; Ok(Gadget { id: raw_agent.prof as u16, name: name.to_str()?.to_owned(), }) } fn from_raw_player(raw_agent: &raw::Agent) -> Result { assert!(raw_agent.is_player()); let character_name = raw::cstr_up_to_nul(&raw_agent.name) .ok_or(EvtcError::InvalidData)? .to_str()?; let account_name = raw::cstr_up_to_nul(&raw_agent.name[character_name.len() + 1..]) .ok_or(EvtcError::InvalidData)? .to_str()?; let subgroup = raw_agent.name[character_name.len() + account_name.len() + 2] - b'0'; let elite = if raw_agent.is_elite == 0 { None } else { Some( EliteSpec::from_u32(raw_agent.is_elite) .ok_or(EvtcError::InvalidEliteSpec(raw_agent.is_elite))?, ) }; Ok(Player { profession: Profession::from_u32(raw_agent.prof) .ok_or(EvtcError::InvalidProfession(raw_agent.prof))?, elite, character_name: character_name.to_owned(), account_name: account_name.to_owned(), subgroup, }) } /// Accesses the inner [`Player`][Player] struct, if available. pub fn as_player(&self) -> Option<&Player> { if let AgentKind::Player(ref player) = *self { Some(player) } else { None } } /// Determines whether this `AgentKind` contains a player. pub fn is_player(&self) -> bool { self.as_player().is_some() } /// Accesses the inner [`Gadget`][Gadget] struct, if available. pub fn as_gadget(&self) -> Option<&Gadget> { if let AgentKind::Gadget(ref gadget) = *self { Some(gadget) } else { None } } /// Determines whether this `AgentKind` contains a gadget. pub fn is_gadget(&self) -> bool { self.as_gadget().is_some() } /// Accesses the inner [`Character`][Character] struct, if available. pub fn as_character(&self) -> Option<&Character> { if let AgentKind::Character(ref character) = *self { Some(character) } else { None } } /// Determines whether this `AgentKind` contains a character. pub fn is_character(&self) -> bool { self.as_character().is_some() } } impl TryFrom for AgentKind { type Error = EvtcError; /// Convenience method to avoid manual borrowing. /// /// Note that this conversion will consume the agent, so if you plan on re-using it, use the /// `TryFrom<&raw::Agent>` implementation that works with a reference. fn try_from(raw_agent: raw::Agent) -> Result { Self::try_from(&raw_agent) } } impl TryFrom<&raw::Agent> for AgentKind { type Error = EvtcError; /// Extract the correct `AgentKind` from the given [raw agent][raw::Agent]. /// /// This automatically discerns between player, gadget and characters. /// /// Note that in most cases, you probably want to use `Agent::try_from` or even /// [`process`][process] instead of this function. fn try_from(raw_agent: &raw::Agent) -> Result { if raw_agent.is_character() { Ok(AgentKind::Character(AgentKind::from_raw_character( raw_agent, )?)) } else if raw_agent.is_gadget() { Ok(AgentKind::Gadget(AgentKind::from_raw_gadget(raw_agent)?)) } else if raw_agent.is_player() { Ok(AgentKind::Player(AgentKind::from_raw_player(raw_agent)?)) } else { Err(EvtcError::InvalidData) } } } /// An agent. /// /// Agents in arcdps are very versatile, as a lot of things end up being an "agent". This includes: /// * Players /// * Bosses /// * Any additional mobs that spawn /// * Mesmer illusions /// * Ranger spirits, pets /// * Guardian spirit weapons /// * ... /// /// Generally, you can divide them into three kinds ([`AgentKind`][AgentKind]): /// * [`Player`][Player]: All players themselves. /// * [`Character`][Character]: Non-player mobs, including most bosses, "adds" and player-generated /// characters. /// * [`Gadget`][Gadget]: Some additional gadgets, such as ley rifts, continuum split, ... /// /// All of these agents share some common fields, which are the ones accessible in `Agent`. /// The kind can be retrieved using [`.kind()`][Agent::kind], which can be matched on. /// /// # Obtaining an agent /// /// The normal way to obtain the agents is to use the [`.agents()`](Log::agents) method on a /// [`Log`][Log], or one of the other accessor methods (like [`.players()`][Log::players] or /// [`.agent_by_addr()`][Log::agent_by_addr]). /// /// In the cases where you already have a [`raw::Agent`][raw::Agent] available, you can also /// convert it to an [`Agent`][Agent] by using the standard /// [`TryFrom`][TryFrom]/[`TryInto`][std::convert::TryInto] traits: /// /// ```no_run /// # use evtclib::{Agent, raw}; /// use std::convert::TryInto; /// let raw_agent: raw::Agent = panic!(); /// let agent: Agent = raw_agent.try_into().unwrap(); /// ``` /// /// Note that you can convert references as well, so if you plan on re-using the raw agent /// afterwards, you should opt for `Agent::try_from(&raw_agent)` instead. /// /// # The `Kind` parameter /// /// The type parameter is not actually used and only exists at the type level. It can be used to /// tag `Agent`s containing a known kind. For example, `Agent` implements /// [`.player()`][Agent::player], which returns a `&Player` directly (instead of a /// `Option<&Player>`). This works because such tagged `Agent`s can only be constructed (safely) /// using [`.as_player()`][Agent::as_player], [`.as_gadget()`][Agent::as_gadget] or /// [`.as_character()`][Agent::as_character]. This is useful since functions like /// [`Log::players`][Log::players], which already filter only players, don't require the consumer /// to do another check/pattern match for the right agent kind. /// /// The unit type `()` is used to tag `Agent`s which contain an undetermined type, and it is the /// default if you write `Agent` without any parameters. /// /// The downside is that methods which work on `Agent`s theoretically should be generic over /// `Kind`. An escape hatch is the method [`.erase()`][Agent::erase], which erases the kind /// information and produces the default `Agent<()>`. Functions/methods that only take `Agent<()>` /// can therefore be used by any other agent as well. #[derive(Debug, Clone, Hash, PartialEq, Eq, Getters, CopyGetters)] // For the reasoning of #[repr(C)] see Agent::transmute. #[repr(C)] pub struct Agent { /// The address of this agent. /// /// This is not actually the address of the in-memory Rust object, but rather a serialization /// detail of arcdps. You should consider this as an opaque number and only compare it to other /// agent addresses. #[get_copy = "pub"] addr: u64, /// The kind of this agent. #[get = "pub"] kind: AgentKind, /// The toughness of this agent. /// /// This is not an absolute number, but a relative indicator that indicates this agent's /// toughness relative to the other people in the squad. /// /// 0 means lowest toughness, 10 means highest toughness. #[get_copy = "pub"] toughness: i16, /// The concentration of this agent. /// /// This is not an absolute number, but a relative indicator that indicates this agent's /// concentration relative to the other people in the squad. /// /// 0 means lowest concentration, 10 means highest concentration. #[get_copy = "pub"] concentration: i16, /// The healing power of this agent. /// /// This is not an absolute number, but a relative indicator that indicates this agent's /// healing power relative to the other people in the squad. /// /// 0 means lowest healing power, 10 means highest healing power. #[get_copy = "pub"] healing: i16, /// The condition damage of this agent. /// /// This is not an absolute number, but a relative indicator that indicates this agent's /// condition damage relative to the other people in the squad. /// /// 0 means lowest condition damage, 10 means highest condition damage. #[get_copy = "pub"] condition: i16, /// The instance ID of this agent. #[get_copy = "pub"] instance_id: u16, /// The timestamp of the first event entry with this agent. #[get_copy = "pub"] first_aware: u64, /// The timestamp of the last event entry with this agent. #[get_copy = "pub"] last_aware: u64, /// The master agent's address. #[get_copy = "pub"] master_agent: Option, phantom_data: PhantomData, } impl TryFrom<&raw::Agent> for Agent { type Error = EvtcError; /// Parse a raw agent. fn try_from(raw_agent: &raw::Agent) -> Result { let kind = AgentKind::try_from(raw_agent)?; Ok(Agent { addr: raw_agent.addr, kind, toughness: raw_agent.toughness, concentration: raw_agent.concentration, healing: raw_agent.healing, condition: raw_agent.condition, instance_id: 0, first_aware: 0, last_aware: u64::max_value(), master_agent: None, phantom_data: PhantomData, }) } } impl TryFrom for Agent { type Error = EvtcError; /// Convenience method to avoid manual borrowing. /// /// Note that this conversion will consume the agent, so if you plan on re-using it, use the /// `TryFrom<&raw::Agent>` implementation that works with a reference. fn try_from(raw_agent: raw::Agent) -> Result { Agent::try_from(&raw_agent) } } impl Agent { /// Unconditionally change the tagged type. #[inline] fn transmute(&self) -> &Agent { // Beware, unsafe code ahead! // // What are we doing here? // In Agent, T is a marker type that only exists at the type level. There is no actual // value of type T being held, instead, we use PhantomData under the hood. This is so we // can implement special methods on Agent, Agent and Agent, // which allows us in some cases to avoid the "second check" (e.g. Log::players() can // return Agent, as the function already makes sure all returned agents are // players). This makes the interface more ergonomical, as we can prove to the type checker // at compile time that a given Agent has a certain AgentKind. // // Why is this safe? // PhantomData (which is what Agent boils down to) is a zero-sized type, which means // it does not actually change the layout of the struct. There is some discussion in [1], // which suggests that this is true for #[repr(C)] structs (which Agent is). We can // therefore safely transmute from Agent to Agent, for any U and T. // // Can this lead to unsafety? // No, the actual data access is still done through safe rust and a if-let. In the worst // case it can lead to an unexpected panic, but the "guarantee" made by T is rather weak in // that regard. // // What are the alternatives? // None, as far as I'm aware. Going from Agent to Agent is possible in safe Rust by // destructuring the struct, or alternatively by [2] (if it would be implemented). However, // when dealing with references, there seems to be no way to safely go from Agent to // Agent, even if they share the same layout. // // [1]: https://www.reddit.com/r/rust/comments/avrbvc/is_it_safe_to_transmute_foox_to_fooy_if_the/ // [2]: https://github.com/rust-lang/rfcs/pull/2528 unsafe { &*(self as *const Agent as *const Agent) } } /// Erase any extra information about the contained agent kind. #[inline] pub fn erase(&self) -> &Agent { self.transmute() } /// Try to convert this `Agent` to an `Agent` representing a `Player`. #[inline] pub fn as_player(&self) -> Option<&Agent> { if self.kind.is_player() { Some(self.transmute()) } else { None } } /// Try to convert this `Agent` to an `Agent` representing a `Gadget`. #[inline] pub fn as_gadget(&self) -> Option<&Agent> { if self.kind.is_gadget() { Some(self.transmute()) } else { None } } /// Try to convert this `Agent` to an `Agent` representing a `Character`. #[inline] pub fn as_character(&self) -> Option<&Agent> { if self.kind.is_character() { Some(self.transmute()) } else { None } } } impl Agent { /// Directly access the underlying player data. #[inline] pub fn player(&self) -> &Player { self.kind.as_player().expect("Agent had no player!") } /// Shorthand to get the player's account name. #[inline] pub fn account_name(&self) -> &str { self.player().account_name() } /// Shorthand to get the player's character name. #[inline] pub fn character_name(&self) -> &str { self.player().character_name() } /// Shorthand to get the player's elite specialization. #[inline] pub fn elite(&self) -> Option { self.player().elite() } /// Shorthand to get the player's profession. #[inline] pub fn profession(&self) -> Profession { self.player().profession() } /// Shorthand to get the player's subgroup. #[inline] pub fn subgroup(&self) -> u8 { self.player().subgroup() } } impl Agent { /// Directly access the underlying gadget data. #[inline] pub fn gadget(&self) -> &Gadget { self.kind.as_gadget().expect("Agent had no gadget!") } /// Shorthand to get the gadget's id. #[inline] pub fn id(&self) -> u16 { self.gadget().id() } /// Shorthand to get the gadget's name. #[inline] pub fn name(&self) -> &str { self.gadget().name() } } impl Agent { /// Directly access the underlying character data. #[inline] pub fn character(&self) -> &Character { self.kind .as_character() .expect("Agent had no character!") } /// Shorthand to get the character's id. #[inline] pub fn id(&self) -> u16 { self.character().id() } /// Shorthand to get the character's name. #[inline] pub fn name(&self) -> &str { self.character().name() } } /// A fully processed log file. #[derive(Debug, Clone)] pub struct Log { agents: Vec, events: Vec, boss_id: u16, } impl Log { /// Return all agents present in this log. #[inline] pub fn agents(&self) -> &[Agent] { &self.agents } /// Return an agent based on its address. pub fn agent_by_addr(&self, addr: u64) -> Option<&Agent> { self.agents.iter().find(|a| a.addr == addr) } /// Return an agent based on the instance ID. pub fn agent_by_instance_id(&self, instance_id: u16) -> Option<&Agent> { self.agents.iter().find(|a| a.instance_id == instance_id) } /// Return the master agent of the given agent. /// /// * `addr` - The address of the agent which to get the master for. pub fn master_agent(&self, addr: u64) -> Option<&Agent> { self.agent_by_addr(addr) .and_then(|a| a.master_agent) .and_then(|a| self.agent_by_addr(a)) } /// Return an iterator over all agents that represent player characters. pub fn players(&self) -> impl Iterator> { self.agents.iter().filter_map(|a| a.as_player()) } /// Return an iterator over all agents that are NPCs. pub fn npcs(&self) -> impl Iterator> { self.agents.iter().filter_map(|a| a.as_character()) } /// Return an iterator over all agents that are gadgets. pub fn gadgets(&self) -> impl Iterator> { self.agents.iter().filter_map(|a| a.as_gadget()) } /// Return the boss agent. /// /// Be careful with encounters that have multiple boss agents, such as Trio /// and Xera. pub fn boss(&self) -> &Agent { self.npcs() .find(|c| c.character().id == self.boss_id) .map(Agent::erase) .expect("Boss has no agent!") } /// Return all boss agents. /// /// This correctly returns multiple agents on encounters where multiple /// agents are needed. pub fn boss_agents(&self) -> Vec<&Agent> { let bosses = self .encounter() .map(Encounter::bosses) .unwrap_or(&[] as &[_]); self.npcs() .filter(|c| bosses.iter().any(|boss| *boss as u16 == c.character().id)) .map(Agent::erase) .collect() } /// Check whether the given address is a boss agent. pub fn is_boss(&self, addr: u64) -> bool { self.boss_agents().into_iter().any(|a| a.addr() == addr) } /// Returns the encounter id. #[inline] pub fn encounter_id(&self) -> u16 { self.boss_id } /// Returns the encounter, if present. /// /// Some logs don't have an encounter set or have an ID that is unknown to us (for example, if /// people set up arcdps with custom IDs). Therefore, this method can only return the encounter /// if we know about it in [`Encounter`]. #[inline] pub fn encounter(&self) -> Option { Boss::from_u16(self.boss_id).map(Boss::encounter) } /// Return an analyzer suitable to analyze the given log. pub fn analyzer<'s>(&'s self) -> Option> { analyzers::for_log(&self) } /// Return all events present in this log. #[inline] pub fn events(&self) -> &[Event] { &self.events } /// Returns the timespan of the log in milliseconds. /// /// The timespan is the time between the first registered event and the last registered event, /// measured in milliseconds. /// /// Note that this does not necessarily equate to the fight/encounter duration, as arcdps /// starts logging as soon as you enter combat, but some bosses are still invulnerable (e.g. /// Ensolyss). It does however give a good idea and is cheap to compute. /// /// In the rare occassions that a log does not have any events, this function will return 0. pub fn span(&self) -> u64 { let first = self.events().first().map(Event::time).unwrap_or(0); let last = self.events().last().map(Event::time).unwrap_or(0); last - first } } /// Convenience data accessing funtions for [`Log`][Log]s. /// /// The information that is gathered by those functions is "expensive" to compute, as we have to /// loop through every event. They are not saved in the header, and instead are implemented using /// special [`EventKind`][EventKind]s. This is not a limitation of `evtclib`, but rather a result /// of how arcdps stores the data. /// /// This also means that those functions are fallible because we cannot guarantee that the special /// events that we're looking for is actually present in every log file. /// /// Use those functions only if necessary, and prefer to cache the result if it will be reused! impl Log { /// Check whether the fight was done with challenge mote activated. /// /// This function always returns `false` if /// * The fight was done without CM /// * The fight does not have a CM /// * We cannot determine whether the CM was active /// * The boss is not known pub fn is_cm(&self) -> bool { self.analyzer().map(|a| a.is_cm()).unwrap_or(false) } /// Get the timestamp of when the log was started. /// /// The returned value is a unix timestamp in the local time zone. /// /// If the [`LogStart`][EventKind::LogStart] event cannot be found, this function returns /// `None`. pub fn local_start_timestamp(&self) -> Option { self.events().iter().find_map(|e| { if let EventKind::LogStart { local_timestamp, .. } = e.kind() { Some(*local_timestamp) } else { None } }) } /// Get the timestamp of when the log was ended. /// /// The returned value is a unix timestamp in the local time zone. /// /// If the [`LogEnd`][EventKind::LogEnd] event cannot be found, this function returns `None`. pub fn local_end_timestamp(&self) -> Option { self.events().iter().find_map(|e| { if let EventKind::LogEnd { local_timestamp, .. } = e.kind() { Some(*local_timestamp) } else { None } }) } /// Check if rewards for this fight have been given out. /// /// This can be used as an indication whether the fight was successful (`true`) or not /// (`false`). /// /// If you want to properly determine whether a fight was successful, check the /// [`Analyzer::outcome`][Analyzer::outcome] method, which does more sophisticated checks /// (dependent on the boss). pub fn was_rewarded(&self) -> bool { self.events() .iter() .any(|e| matches!(e.kind(), EventKind::Reward { .. })) } /// Returns all error strings that were captured. /// /// If no errors were encountered, an empty vec is returned. /// /// Note that those are errors reported verbatim by arcdps, nothing that evtclib /// produces/interprets. pub fn errors(&self) -> Vec<&str> { self.events() .iter() .filter_map(|e| { if let EventKind::Error { ref text } = e.kind() { Some(text as &str) } else { None } }) .collect() } }