TensorReactions is a fully-featured and extremely optimized reactions system, that incorporates both timelines for savage and ultimate raids, as well as more traditional event-based reactions.

Notable featuers:

• Works with every ACR and even Skill Manager
• No FPS drops/lag
• Importable and Exportable reaction profiles
• Automatically updates trigger profiles when a timeline update gets pushed
• Can execute reactions at certain points in the timeline
• Can execute reactions based on event triggers, such as on entity cast
• Extremely powerful GUI that covers most options for reactions, but also has lua for either full lua reaction or combining gui conditions and actions with lua conditions and actions
• Unlimited possibilities with lua code, almost every reaction is possible
• Advanced ACR override system can override ACR and Skill Manager to weave whatever ability you want, or choose not to follow weaving rules to execute in custom order
• Much more but this list would be too long

This addon is huge in scope, and covering everything in this wiki is impossible. I have provided basic usage information and examples, however you should go and experiment with the various options

General theme with the UI is to right click when in doubt
Timeline - This is the fight for which you add reactions for.

Trigger Profile - This the trigger profile for the selected fight. Contains all your reactions. To swap between profiles you just select a different one in this dropdown.

Next, going left to right, top to bottom with the ui elements:

Timeline List - Has all the mechanics in the selected fight. You can right click on any element in this list to copy/paste the entire set of reactions, or remove all the reactions entirely.

Mechanic Info Panel - Has some useful information for the selected mechanic

Conditions - Has a list of conditions. These cannot be reordered, and have indexes so you know which one is which.

Actions - List of actions that will be executed in that reaction. The actions are executed in that exact priority, and ogcds/gdcs are weaved correctly depending on priority. Note: Priority =/= order! If you want to execute a sequence of actions, you need a last skill check on each action. For more info, please see the examples and/or ask in the discord.

Name: Can be changed to whatever you want, duplicates are allowed (internal uuid system is used)

Timer Offset: Relative to the selected mechanic, when the reaction will be executed. For example, if the selected mechanic time is 247.6, a timer offset of -4 will use that reaction 4s before that mechanic, at 243.6s. The timer offset has a 2.5s window to execute the reaction before it will be dequeued as stale. I recommend using Timer Offset for instant actions such as variables and misc reactions.

Use Time Range: Select the window of time for when the reaction will be attempted to be used. For example, a Timer Start Offset of -5 and a Timer End Offset of 5 will attempt to execute the reaction starting from 5s before the mechanic and until 5s after the mechanic. If the time of the mechanic is at 247.6, the reaction will be attempted to be executed starting at 242.6 and until 252.6. I recommend using Time Range for Skill overrides that are manually used by TensorReactions, so that you have enough time for the action to go off.

Loop Reaction: Every reaction will run only once for that given pull, however loop reaction will run it continuously in the window it's set to be executed in. By default, the timer offset window is 2.5s, and the Time Range window is whatever you set it to be. I highly reocmmend using this in conjunction with the Time Range option for reactions to continuously execute before a given mechanic. For example, a reaction would continuously heal low hp players starting from -30s until -5s, using an action priority that will end on the highest priority action.

Lua: Each lua reaction will have 3 variables that it can use (and one that it must):

self: the reaction object. You must set self.used = true if the reaction successfully went off and can be dequeued!

data: Global table passed to every lua reaction, condition, and action, gets reset on wipe! This table is useful for checking variables set by other lua reactions as well as setting them. For example: a lua reaction that would check

if data.lastUse == nil or TimeSince(data.lastUse) > 500 then
    data.lastUse = Now()
    self.used = true
end

would add a 500ms delay to this reaction.

You can also invoke the ACR override by returning the following from your lua code:

table action, number targetID, bool ignoreWeaveRules, bool allowInterrupt

Argument action must be a valid action from the ActionList. If you choose to return an action from lua, make sure the “Lua Returns Action” box is checked!

The way this list works is, you create your conditions here, and then link them to specific actions below. That way you can have multiple actions using the same condition without having to make a new condition every time. Every action will also have it's own independent list of linked conditions.

Category: General category that exposes more options. Self contains most of the important ones, such as Gauge, Map ID, Skill Cooldown checks, etc, however I encourage you to explore the options and ask in the discord if you don't understand any of them.

Action CD: This list will only show actions that are usable by your current job and level. You can, of course, switch jobs and it will display the actions for that job. Once the action is set, it is stored regardless of job, so you don't need to worry about switching jobs and losing the set skill

Emnity: This value is target.aggropercentage on your current target.

Filter: Filter condition will loop over your target list and set detection target, then run your conditions. In the linked conditions to filter, you must set detection target to be the target you are checking on (party > target: detection target), and then once it's done it will look at the list of all targets that match, pick a subtype, and then set that as detection target. Then you link your skill action and use detection target as your target, with the filter condition linked. Tip: You can set “min number of entities” to 0, and enable “match all”, and effectively this will become an “OR” condition. Setting min entities to 0 just means reactions will ignore your target list and simply evaluate the conditions as is.

Lua: If you make a lua reaction, you must return a boolean, else your condition will be ignored. Important variables that you can use:

data: Global table passed to every lua reaction, condition, and action, gets reset on wipe! This table is useful for checking variables set by other lua reactions. For example: a lua reaction that would check

return data.lastUse ~= nil and TimeSince(data.lastUse) > 500

would add a 500ms delay to any action that used this condition.

Skill: This will call the action override system that works with any acr / skillmgr profile and will follow gcd/ogcd weaving rules.

Interrupt Channels: This will interrupt channels to use your ability. If Interrupt Channels is checked, OGCDs will automatically ignore ogcd rules regardless of the “Ignore Weave Rules” setting!

Ignore Weave Rules: OGCD Will ignore weave rules and will be allowed usage during a gcd window. Example usage: Emergency Tactics into adlo on sch, using reassemble/queen before boss is targettable on mch.

Untarget: Untargets enemy. Will not show if Set Target is checked.

Set Target: Targets an enemy. Will not show if Untarget is checked.

Condition List: List of conditions that the action will check. Each condition must be satisfied for the action to execute. This list can be dragged and dropped, you can also right click on the condition to view options.

Lua: There are 3 useful variables that you can (and one that you must) use here.

self: the action object. You must set self.used = true if the action successfully went off and can be dequeued!

data: Global table passed to every lua reaction, condition, and action, gets reset on wipe! This table is useful for checking variables set by other lua reactions as well as setting them. For example: a lua action that executes the following code:

local target = TensorCore.getEntityByGroup("Main Tank")
local minne = ActionList:Get(1, 7408)
if minne.cdmax - minne.cd < 1.0 and target~= nil then
    ACR_TensorRequiem_Hotbar_NaturesMinne_Custom = true
    ACR_TensorRequiem_Hotbar_NaturesMinne_Custom_TargetID = target.id
    self.used = true
end

would invoke the hotbar method inside of TensorRequiem to use natures minne on your main tank.

You can also invoke the ACR override by returning the following from your lua code:

table action, number targetID, bool ignoreWeaveRules, bool allowInterrupt

Argument action must be a valid action from the ActionList. If you choose to return an action from lua, make sure the “Lua Returns Action” box is checked!

Profile: This is the profile that will be loaded. General reactions are loaded regardless of timeline and map, hence the name “general”.

Each reaction in this list will be executed in the order of the listing and depending on event. The UI is the same as Timeline Reactions but with a few minor tweaks.

Event: List of events to trigger your event with. OnUpdate should be used as a last resort, as it's a bit resource heavy depending on the reaction.

Lua: every lua reaction and action inside general reactions will have the same variables as the timeline ones, however they will also have an additional variable, eventArgs:

self: the same as timelines, which means for lua reactions and lua actions, self.used must be set if the (re)action goes off successfully

data: same as timelines, for you to use as you wish. This data variable may prove very useful in general reactions as you can set it based on certain events happening. For example: on limit cut in TEA, you get marker number 5. You can use the OnEntityMarkerAdd event and see which marker is added on you, and store that in a data.myLimitCutNumber variable. You can use that variable to use certain skills in the timeline for tea (remember that data object is shared between everything, general timeline reactions included

eventArgs: this is the new arg for general lua reactions/actions/conditions. It is a table, and each key is different for each event (listed below). Example usage: d(eventArgs.entityID)

You can also invoke the ACR override by returning the following from your lua code:

table action, number targetID, bool ignoreWeaveRules, bool allowInterrupt

Argument action must be a valid action from the ActionList. If you choose to return an action from lua, make sure the “Lua Returns Action” box is checked!

• TensorReactions_CurrentTimer
  • Contains a copy of the current internal timer that reactions uses

• OnUpdate: no args, simply an event that runs every frame

• OnEntityCast: Triggered the moment an entity casts an instant cast action, or successfully finishes a channel action
  • entityID
    • id of entity that is casting action
  • spellID
    • id of spell that entity is casting

• OnEntityChannel: Triggered the moment an entity starts channeling an action
  • entityID
    • id of entity that's channeling the spell
  • spellID
    • id of the action being channeled
  • targetID
    • id of target the action is being casted on
  • channelTimeMax
    • how long the entity needs to channel for before the cast is finished

• OnEntityMarkerAdd: Triggered whenever a marker is added above an entity (stack marker etc). I recommend making a debug trigger on this one so you can find out the marker types, or ask in the discord.
  • entityID
    • id of entity the marker is added to
  • markerID
    • marker type that is being added (this value is consistent; it will always be the same for the same mechanic)

• OnEntityAdd: Triggered whenever a new entity is added/loaded in. Useful for detecting when add spawns happen. These will also trigger on reset so you should have a combat check on any entity add/remove reactions.
  • entityID
    • id of entity that was added
  • entityName
    • name of entity that was added

• OnEntityRemove: Triggered whenever an entity is removed/unloaded. This is not the same as when an entity dies, often an entity will die but still exist for a 5-10 seconds after These will also trigger on reset so you should have a combat check on any entity add/remove reactions.
  • entityID
    • id of entity that was removed
  • entityName
    • name of entity that was removed

• OnNewChatLine: Triggered whenever a new chat message is sent, including combat log. Useful for detecting certain phase transitions and various misc reactions
  • line
    • line object that was sent. Use line.line to get the actual text.

• OnNewBuffEntry: Triggered whenever a buff is added to an entity
  • entityID
    • id of entity that buff was added to
  • buffID
    • id of buff that was added
  • buffDuration
    • duration of buff that was added
  • ownerID
    • owner of the buff that was added to entity

• OnWipe: no args, simply gets triggered when you go out of combat, and only gets triggered if a valid timeline is loaded to prevent situations like dungeons constantly triggering onwipe()

italics = type, underline = optional

• TensorCore.isAnyEntityCasting(number castid, string elistFilters)
  • Returns bool if any entity is channeling action castid, and table entity that is channeling castid. Optionally, elistFilters can be specified to check a more narrow list of entities.
• TensorCore.isEntityAlive(number contentid, string elistFilters)
  • Returns bool if entity with contentid is alive, and table entity that is matches contentid and is alive. Optionally, elistFilters can be specified to check a more narrow list of entities.
• TensorCore.getEntityByGroup(string strGroup, string or table subGroup, bool inParty, number range, number szCluster)
  • Returns table entity of specified group and subgroup. If strGroup is “Named Target” or “ContentID” then subgroup must be a table that contains keys “name”, “subgroup”, and “contentid”. Example: {name = “Striking Dummy”, subgroup = “Nearest”, contentid = 541}. Important note: you don't need contentid if strGroup is “Named Target”, and you don't need name if strGroup is “ContentID”. Subgroup will default to nearest in all cases.
  • Full list of strGroups: {“All”, “Self”, “Named Target”, “ContentID”, “Current Target”, “Enemy”, “Target of Current Target”, “Target Spell of Current Target”, “Most Clustered Ally”, “Most Clustered Enemy”, “Party”, “Main Tank”, “Off Tank”, “Tank”, “Melee DPS”, “Ranged DPS”, “Ranged Physical DPS”, “Ranged Caster DPS”, “Healer”}
  • Full list of subGroups: {“Nearest”, “Lowest HP”, “Highest HP”, “Number”} Important Note: Number will simply return the list of strGroup targets back to you.
  • Full list of strGroups that require a subgroup: {“Named Target”, “Enemy”, “Party”, “Tank”, “ContentID”, “Melee DPS”, “Ranged DPS”, “Ranged Physical DPS”, “Ranged Caster DPS”, “Healer”, “Target Spell of Current Target”}
• TensorCore.resetTTKTargets(number …)
  • This function accepts a variable number of arguments. Given an id, it will reset all time to kill data on that entity. Useful for Living Liquid / Living hand split in TEA p1 that throws off TTK. Example Usage: TensorCore.resetTTKTargets(123, 456). This will reset ttk info on entities with id 123 and 456.