This module implements a 3D pathfinding system using the A* algorithm in a Panda3D environment. It allows AI-controlled entities to follow a target (like the player) by dynamically calculating paths through a graph loaded from a .pfs file.

• A* pathfinding between graph nodes
• Player-following AI that dynamically updates its goal
• .pfs file parser for navigation points and connections
• Smooth AI movement with frame-based interpolation
• Periodic path recalculation based on player position

• PFSParser: Parses .pfs files to extract point positions and graph structure
• Ai: Core AI controller
  • clpToobject(): Finds the closest graph node to the player or target
  • set_new_goal(pid): Sets a new goal and computes the path
  • astar(start, goal): Calculates the shortest path using A*
  • move_along_path(task): Moves the AI along the path over time
  • updateAi(task): Re-evaluates the player's position and updates the goal

1.import as an add-on "PSF_exporter.py" for blender 4.x

2.create a flat ground (plan) and subdivise into multiples faces, and select the ground

3.go to export and find export_scene.pds and select selected only and export

/!\ other objects in the scene that are not selected will be considered as obstacles.

PS: sorry if there are some element in french

<< Exported by Blender PFS Exporter >>

<Area> [AreaName] {

  <SubArea> [SubArea1] {
    <Group_of_Points> {
      <Point> 1 {
        <x> {0.0}
        <y> {0.0}
        <z> {0.0}
        <Tag> [start]
      }
      <Point> 2 {
        <x> {1.0}
        <y> {0.0}
        <z> {0.0}
      }
    }
    <Connections> {
      <Relate> {
        <Crossroad> {1}
        <Peripheral> {2}
      }
    }
  }

  <SubArea> [SubArea2] {
    <Group_of_Points> {
      <Point> 3 {
        <x> {2.0}
        <y> {0.0}
        <z> {0.0}
      }
    }
    <Connections> {
      <Relate> {
        <Crossroad> {2}
        <Peripheral> {3}  # external link with SubArea1
      }
    }
  }

}

from direct.showbase.ShowBase import ShowBase
from pathfinding import PFSParser, Ai
from panda3d.core import Vec3

class MyApp(ShowBase):
    def __init__(self):
        ShowBase.__init__(self)

        # Load the player model
        self.player = loader.loadModel("models/player.egg")
        self.player.reparentTo(render)
        self.player.setPos(0, 0, 0)

        # Load navigation graph
        parser = PFSParser("navpoint.pfs")
        areas, points, graph = parser.load()

        # Initialize the AI
        ai_model = loader.loadModel("models/enemy.egg")
        self.enemy = Ai(points=points, graph=graph, areas=areas, model=ai_model, world=None, use_bullet=False)
        self.enemy.aitarg = self.player

        # Start AI behavior
        self.taskMgr.do_method_later(0.5, self.enemy.follow, "UpdateAITask")
        self.taskMgr.add(self.enemy.move_along_path, "MoveAIPathTask")

        # Optional: Move the player slightly every frame
        self.taskMgr.add(self.move_player, "MovePlayer")

    def move_player(self, task):
        self.player.setX(self.player.getX() + 0.05)
        return task.cont

app = MyApp()
app.run()

Download and watch the demo video

Demo animation