Giter Club home page Giter Club logo

altv-livecity's Introduction

altv-livecity

Traffic synchronization proof-of-concept alt:V Multiplayer.

What is this?

This is small proof-of-concept for spawning & syncing traffic in GTAO-like way using alt:V dev branch. Last working server version is 15.0-dev368 (dev). I made some research, how GTAO and GTA SP handles traffic. After I made some decent results, I stopped the devlopment, because of limitations from alt:V. That's why I want to share this repository for everyone. You might take some parts from it, if you want to.

This repository will be updated only to keep it working. Maybe.

If you have any questions - please use github Discussions page or reach me in alt:V Discord, @b2soft. I don't care bugs and/or, so just don't. Your PR will be never accepted. Maybe.

Data I discovered so far

There is a lot of needed data you can find in Data folder. Partially it was copied from https://github.com/KWMSources/pedSyncer

The other part is modified DurtyFree dumps https://github.com/DurtyFree/gta-v-data-dumps

CarGenerators is my file generated using CodeWalker.

Breakdown of files used

File Description Details
footpath.msgpack Contains navigation mesh info Based on DurtyFree dump, cleaned to have only needed navmesh parts, where ped can be spawned (sidewalks only)
AllowedScenarios.json Contains allowed scenarios for Scenario Ped From pedSyncer repo
CarColorsNum.json Contains Car Colors Copy from pedSyncer repo
CarGenerators.xml Contains all the Car Generators Generated from CodeWalker, used for Parked Vehicles
CarModels.json Contains all the Car Models Copy from pedSyncer repo
ColorlessCars.json Contains car models, which are not "painted" Red police car looks strange, ya?
ExtendedNodes.json Contiains all Path Nodes for cars/boats Based on DurtyFree dump, added best headings provided by GTA natives (did not help) + minified format
pedComponentVariations.json Contains variations of Drawable components for Ped models From DurtyFree dump
PedModelGroup.json Contains groups of ped models used by Scenario Peds From pedSyncer repo
PopCycle Contains Population Cycles description Copy-Paste from GTA files
PopGroups.xml Contains relations between Population Cycle group and Ped/Vehicle model Detailed description can be found below
ScenarioPoints.json Contains all Scenario points for Scenario Peds From pedSyncer repo
vehicles.json Contains Vehicle metadata From DurtyFree dump
ZoneBind.ymt Contains relations between Zone and PopZone names Detailed description can be found below
Zones Contains bounds of Zones Detailed description can be found below

Population Cycle format

Everything related to spawn in GTA follows the Population Cycles file. Each section (between POP_SCHEDULE: and END_POP_SCHEDULE) of PopCycle file represents one PopZone. SP and MP zones are set up separately. Each line of a section represents 2 hours time period, starting from 00:00. Each line contains Limits and Probabilities.

Let's break down one of lines, for instance NET_VINE_WOOD PopZone, 12:00-14:00 time period:

10 50 85 20 30 1 0 3 4 1 peds West_Vinewood_Hipsters 20 West_Vinewood_StreetGeneral 60 West_Vinewood_Vinewood 10 Club_Nighttime 00 cars VEH_RICH_MP 40 VEH_MID_MP 45 VEH_TRANSPORT_MP 05 VEH_LARGE_CITY_MP 00 VEH_BIKES_MP 10 Format of Limits is:

  • 10 - Maximum number of ambient peds
  • 50 - Maximum number of scenario peds
  • 85 - Maximum number of cars
  • 20 - Maximum number of parked cards
  • 30 - Maximum number of low priority parked cards - I do not know what does this mean. Probably GTA will destream low priority parked cars to let closer vehicle to spawn
  • 1 - Percentage of police cars
  • 0 - Percentage of peds, that are cops
  • 3 - Maximum number of ped models, to be used by scenario peds
  • 4 - Maximum number of vehicle models, to be used by scenario vehicles
  • 1 - MaxPreAssignedParked - I do not know what does this mean. Probably, how much parked vehicles GTA pre-spawns (and other are just streamed on demand)

Format of probabilities is:

  • peds - start of Peds probabilities
  • West_Vinewood_Hipsters - name of PopGroup to be used
  • 20 - Probability in percents. 20 % that spawned ped will be from West_Vinewood_Hipsters group
  • (same formatted groups and probabilities till cars section)
  • cars - start of Cars probabilities
  • VEH_RICH_MP - name of PopGroup to be used (same logic as for peds)
  • 40 - Probability in percents. 20 % that spawned vehicle will be from VEH_RICH_MP group
  • (same formatted groups and probabilities till end of line)

To match the exact ped or vehicle model PopGroups.xml is used: you can find lists of corresponding models for each PopCycle group.

Zones drama and how spawn works

You may be noticed, that name of PopZone from PopCycle file does not equal any known zone names from natives etc. Imagine we want to spawn a ped at given coordinate. To match the zone, following steps are required:

  1. Based on a given position, find the Zone using Zones.txt file, for instance our position belongs to Z_Alta3 Zone
  2. Find PopZone using ZoneBind.ymt, for Z_Alta3 you will have 2 names alta and net_vine_wood. The former is used for GTA SP, the latter - for GTAO
  3. Taking net_vine_wood find the corresponding PopZone in PopCycle file.
  4. Based on current time, select line (same as we took for Population Cycle format): 10 50 85 20 30 1 0 3 4 1 peds West_Vinewood_Hipsters 20 West_Vinewood_StreetGeneral 60 West_Vinewood_Vinewood 10 Club_Nighttime 00 cars VEH_RICH_MP 40 VEH_MID_MP 45 VEH_TRANSPORT_MP 05 VEH_LARGE_CITY_MP 00 VEH_BIKES_MP 10
  5. Select weightned random distributed PopGroup, for instance, West_Vinewood_Hipsters
  6. Select random ped model using PopGroups.xml file, for instance a_f_y_hipster_01
  7. Spawn

Code

Copy the Data folder content to [altv-server-root]/data. Project uses C# for server and client. Open LiveCity.sln with Visual Studio. If you have ALTV_SERVER_ROOT environment var pointing to the altv server root, you can choose DebugLocal configuration.

Data is loaded in Server/Navigation/NavigationMeshProvider.cs and Server/LiveCity/LiveCityDataService.cs.

Main logic can be found in Server/LiveCity/LiveCityService.cs and Client/Resource/LiveCity/LiveCityService.cs.

How does it work in LiveCity

What did I find in GTAO?

I found that GTAO spawns traffic within 600 meters around the player and extensively stream it in and out based on camera angle + cache "hidden" entities. After 600m distance GTA uses billboards, probably native.

Cars have no peds driving, unless car is within 80m radius of player + when you zoom in using scope, peds are shown on demand. Helicopters have drivers at all times.

How is it done in this repo?

Main Algorithm

There are 4 types of NPCs spawned:

  • Wander Peds - spawned randomly on a sidewlk, just wandering around
  • Scenario Peds - spawned at Scenario Points taking into account allowed scenarios (like smoking, drinking coffee etc)
  • Wander Vehicles - spawned randomly at eligible Street Nodes. Vehicle's door are locked, drivers follow the safe driving style, obey the rules
  • Parked Vehicles - spawned at Car Generators positions taking into account rules from GTA

When player connects to server, it's being added to list of players to track. After player spawn is complete, each 100 ms LiveCityService tries to fill the NPCs within two ranges, considering streaming range is default (400 m):

  • Close range - circle zone around the player, default radius is 200 m
  • Far range - a sector of ring zone around the player, min radius 200m, max radius 400 m, sector angle is 90 degrees default, main direction is player orientation (tried several times with camera forward vector instead, without caching looks worse, than entity rotation, imo)

Entities are never spawned closer than 100 m default. First it counts how much entities are around the player and if it's possible to add new entities. Budgets are set to not overshoot hard limits of GTA (like 128 peds maximum). Budgets and ranges are specified at GlobalConfig.cs.

How model is chosen

Based on different data dumps, these rules are applied.

  1. Wander Ped models depend on Population Cycle groups and position of spawn
  2. Scenario Ped models depend on allowed models for specific scenarios + based on a Population Cycle as a backup, if scenario is not limited to specific models
  3. Wander Vehicles models depend on Population Cycle groups and position of spawn. Drivers for Wander Vehicles are chosen the same way as Wander Peds
  4. Parked Vehicles models depend on Car Generators' limitations + based on a Population Cycle as a backup, if car generator does not specify exact models

Known Issues and Limits

  1. No caching. Entities are destroyed instead of cache them.
  2. No driverless vehicles. While it can give some room to have more vehicles, it implies for fine control over creating Wander Vehicles.
  3. GetAll* used, becuase GetAll*InRange don't work
  4. Wander Vehicles are not reacting to Emergency vehicle sirens, like police. Probably, becuase alt:V creates peds as Mission type, and not ambient. I tried different natives to fix this, but it's not possible to change Ped type to Ambient
  5. No Billboards/Fake vehicles. Native is blocked by alt:V, no workaround provided for now.
  6. 128 entities is a hardcoded limit of alt:V. LiveCity is adjusted to have 100 Entities around (+ 1 player + 27 other players room). Wander vehicle costs 2 Entities (Vehicle + Driver)
  7. No 2+ ped scenarios. Right now only one ped is playing scenario. GTA has 2 people talking mext to each other, for example.
  8. No Vehicle scenarios. GTA has scenarios to vehicles parking, or going to drive-by Clickin' Bell. Not supporte in LiveCity.

altv-livecity's People

Contributors

radeon590 avatar b2soft avatar

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.