DataBpy 1

import bpy
bpy.app.version_string # current blender version

'5.2.2 LTS'

import sys
from pathlib import Path

import databpy as db
import numpy as np
import polars as pl

# The kernel runs inside Blender (not in the notebook folder), so add the docs helpers to sys.path
sys.path.append(str(Path.home() / "projects/bpy-gallery/docs/utils"))
from gallery_utils import ASSETS, fresh_scene, render_result

# Set render resolution
bpy.context.scene.render.resolution_x = 500
bpy.context.scene.render.resolution_y = 200

def camera_from_above():
    camera = bpy.context.scene.camera
    camera.location = (6, 4, 10)  # Position above the XY plane
    camera.rotation_euler = (0, 0, 0)  # Rotate to look at XY plane
    # Set the camera to orthographic mode
    camera.data.type = 'ORTHO'
    # Set the orthographic scale to 12
    camera.data.ortho_scale = 9

    bpy.context.scene.render.resolution_x = 300
    bpy.context.scene.render.resolution_y = 300

Building Geometry Nodes with nodebpy

The node group that places a sphere on every point, written with nodebpy:

from nodebpy import geometry as g

with g.tree("gn_place_spheres") as tree:
    geometry = tree.inputs.geometry("Geometry")  # Socket_0
    radius = tree.inputs.float("Radius", 1.0, min_value=0.0, subtype="DISTANCE")  # Socket_1
    _ = (
        geometry
        >> g.InstanceOnPoints(instance=g.UVSphere(radius=radius, rings=10))
        >> tree.outputs.geometry("Geometry")
    )

node_group_place_spheres = tree.tree

single_vertex = np.array([[6, 4, 0]])

# Create a mesh object with the single vertex
my_point_obj = db.create_object(single_vertex, name="SinglePoint")

# Print the name of the created object
#print(my_point_obj.name)

camera_from_above()

modifier = my_point_obj.modifiers.new(name="GeometryNodes", type='NODES')
modifier.node_group = node_group_place_spheres

my_point_obj.modifiers["GeometryNodes"].properties.inputs.Socket_1.value = 4 # set radius

url = 'https://raw.githubusercontent.com/kolibril13/ipydrop/refs/heads/main/dino.csv'
df1 = pl.read_csv(url)
length_data = len(df1)

# Scale and transform the coordinates from the dataframe
x = df1['x'] / 10
y = df1['y'] / 10 * 0.8

print(f"Loaded {length_data} rows")
df1.head()

Loaded 142 rows

shape: (5, 2)

x	y
f64	f64
55.3846	97.1795
51.5385	96.0256
46.1538	94.4872
42.8205	91.4103
40.7692	88.3333

# Create 3D points with z=0
coords = np.array(list(zip(x, y, [0]*len(x))))

my_point_obj = db.create_object(coords, name="Deno")

modifier = my_point_obj.modifiers.new(name="GeometryNodes", type='NODES')
modifier.node_group = node_group_place_spheres

my_point_obj.modifiers["GeometryNodes"].properties.inputs.Socket_1.value = 0.2  # Set radius

camera_from_above()

my_point_obj.modifiers["GeometryNodes"].properties.inputs.Socket_1.value = 0.1  # Set radius

# Also possible to change the data points
url = 'https://raw.githubusercontent.com/kolibril13/ipydrop/refs/heads/main/star.csv'
df = pl.read_csv(url)

x = df['x'] / 10 
y = df['y'] / 10 * 0.8

df.head()

shape: (5, 2)

x	y
f64	f64
58.213608	91.881892
58.196054	92.214989
58.718231	90.310532
57.278373	89.907607
58.08202	92.008145

coords = np.array(list(zip(x, y, [0]*len(x))))

db.store_named_attribute(my_point_obj, coords, "position")
my_point_obj.data.update()
bpy.context.view_layer.update()