返回技能列表

script-blender-automation

pjt222
更新于 6 days ago
24 次查看
17
2
17
在 GitHub 上查看
apiautomationdesigndata

关于

This Claude Skill generates Blender Python scripts using the bpy API for automation and procedural generation. It's designed for automating repetitive modeling/animation tasks, creating procedural geometry, and building batch rendering pipelines. Developers should use it for advanced Blender automation, custom add-on development, or integrating Blender with external data sources.

快速安装

Claude Code

推荐
主要方式
npx skills add pjt222/agent-almanac -a claude-code
插件命令备选方式
/plugin add https://github.com/pjt222/agent-almanac
Git 克隆备选方式
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/script-blender-automation

在 Claude Code 中复制并粘贴此命令以安装该技能

技能文档

本 Blender 自動

進階 Blender Python—程序模、鍵動畫、批操、運算註、加件開發。

  • 自動重複模/動務→用
  • 自算或資生程序幾何→用
  • 建批繪管含參變→用
  • 築自運算或加件→用
  • 接 Blender 與外資管或 API→用
  • 數精動畫本→用
  • 開組工→用

自動需務述、參、限繪 100 變、動路自資
資源檔/API程序生外資CSV 坐、JSON 參、API 應
算定碼/數程序生邏分形、參曲、L 系
運算譜自工為與 UI名、屬、模互
動參鍵/資時、緩、限幀範、插曲

一:程序幾何生

用 BMesh 程生網:

import bpy
import bmesh
import math

def create_parametric_surface(name, u_res=32, v_res=32):
    """Generate parametric surface using mathematical function."""
    mesh = bpy.data.meshes.new(name)
    obj = bpy.data.objects.new(name, mesh)
    bpy.context.collection.objects.link(obj)

    bm = bmesh.new()

    verts = []
    for i in range(u_res):
        for j in range(v_res):
            u = (i / (u_res - 1)) * 2 * math.pi
            v = (j / (v_res - 1)) * math.pi

            x = math.sin(v) * math.cos(u)
            y = math.sin(v) * math.sin(u)
            z = math.cos(v)

            vert = bm.verts.new((x, y, z))
            verts.append(vert)

    bm.verts.ensure_lookup_table()
    for i in range(u_res - 1):
        for j in range(v_res - 1):
            v1 = verts[i * v_res + j]
            v2 = verts[(i + 1) * v_res + j]
            v3 = verts[(i + 1) * v_res + (j + 1)]
            v4 = verts[i * v_res + (j + 1)]
            bm.faces.new([v1, v2, v3, v4])

    bm.to_mesh(mesh)
    bm.free()

    return obj

得:自數函生複幾何。

敗:察 BMesh API、驗點索、確面流形。

二:鍵動畫自動

本鍵動與驅:

def animate_rotation(obj, start_frame=1, end_frame=250, axis='Z', rotations=2):
    """Animate object rotation over time."""
    obj.rotation_euler[2] = 0
    obj.keyframe_insert(data_path="rotation_euler", index=2, frame=start_frame)

    obj.rotation_euler[2] = rotations * 2 * math.pi
    obj.keyframe_insert(data_path="rotation_euler", index=2, frame=end_frame)

    if obj.animation_data and obj.animation_data.action:
        for fcurve in obj.animation_data.action.fcurves:
            if 'rotation_euler' in fcurve.data_path:
                for keyframe in fcurve.keyframe_points:
                    keyframe.interpolation = 'LINEAR'

def animate_material_property(mat, property_path, values, frames):
    """Animate material node values."""
    if not mat.node_tree:
        return

    nodes = mat.node_tree.nodes
    emission = nodes.get('Emission')
    if emission:
        for frame, value in zip(frames, values):
            emission.inputs['Strength'].default_value = value
            emission.inputs['Strength'].keyframe_insert(
                data_path="default_value",
                frame=frame
            )

def create_driver(obj, property_path, expression):
    """Create driver for automated animation."""
    driver = obj.driver_add(property_path)
    driver.driver.type = 'SCRIPTED'
    driver.driver.expression = expression

得:鍵插、動正回放。

敗:察屬路、驗 data_path 法、確物可鍵。

三:批處操

批處諸物或檔:

import os
from pathlib import Path

def batch_import_and_render(input_dir, output_dir, file_pattern="*.obj"):
    """Import multiple files and render each."""
    input_path = Path(input_dir)
    output_path = Path(output_dir)
    output_path.mkdir(exist_ok=True)

    scene = bpy.context.scene

    for obj_file in input_path.glob(file_pattern):
        bpy.ops.object.select_all(action='SELECT')
        bpy.ops.object.delete()

        bpy.ops.import_scene.obj(filepath=str(obj_file))

        setup_camera()
        setup_lighting()

        output_file = output_path / f"{obj_file.stem}.png"
        scene.render.filepath = str(output_file)
        bpy.ops.render.render(write_still=True)

        print(f"Rendered: {output_file}")

def batch_material_variation(base_object, colors, output_prefix):
    """Render object with multiple material colors."""
    mat = base_object.data.materials[0]
    bsdf = mat.node_tree.nodes.get('Principled BSDF')

    if not bsdf:
        return

    for i, color in enumerate(colors):
        bsdf.inputs['Base Color'].default_value = color + (1.0,)

        bpy.context.scene.render.filepath = f"{output_prefix}_{i:03d}.png"
        bpy.ops.render.render(write_still=True)

得:諸檔處、各變生繪。

敗:察徑存、驗入運算、理缺材。

四:自運算開發

築自運算為複用工:

import bpy
from bpy.props import FloatProperty, IntProperty

class OBJECT_OT_generate_spiral(bpy.types.Operator):
    """Generate a spiral curve"""
    bl_idname = "object.generate_spiral"
    bl_label = "Generate Spiral"
    bl_options = {'REGISTER', 'UNDO'}

    radius: FloatProperty(
        name="Radius",
        description="Spiral radius",
        default=2.0,
        min=0.1,
        max=10.0
    )

    turns: IntProperty(
        name="Turns",
        description="Number of spiral turns",
        default=5,
        min=1,
        max=20
    )

    resolution: IntProperty(
        name="Resolution",
        description="Points per turn",
        default=32,
        min=8,
        max=128
    )

    def execute(self, context):
        curve = bpy.data.curves.new('Spiral', 'CURVE')
        curve.dimensions = '3D'

        spline = curve.splines.new('NURBS')
        num_points = self.turns * self.resolution

        spline.points.add(num_points - 1)

        for i in range(num_points):
            t = i / self.resolution
            angle = t * 2 * math.pi

            x = self.radius * math.cos(angle)
            y = self.radius * math.sin(angle)
            z = t * 0.5

            spline.points[i].co = (x, y, z, 1.0)

        obj = bpy.data.objects.new('Spiral', curve)
        context.collection.objects.link(obj)
        obj.select_set(True)
        context.view_layer.objects.active = obj

        self.report({'INFO'}, f"Generated spiral with {num_points} points")
        return {'FINISHED'}

def register():
    bpy.utils.register_class(OBJECT_OT_generate_spiral)

def unregister():
    bpy.utils.unregister_class(OBJECT_OT_generate_spiral)

if __name__ == "__main__":
    register()

得:運算現於搜、行支撤。

敗:察 bl_idname 式(小寫底線)、驗屬型。

五:模運算為互動工

築互動模運算:

class OBJECT_OT_modal_scale(bpy.types.Operator):
    """Interactive scaling with mouse"""
    bl_idname = "object.modal_scale"
    bl_label = "Modal Scale"
    bl_options = {'REGISTER', 'UNDO'}

    def __init__(self):
        self.initial_mouse_x = 0
        self.initial_scale = 1.0

    def modal(self, context, event):
        if event.type == 'MOUSEMOVE':
            delta = event.mouse_x - self.initial_mouse_x
            scale = self.initial_scale + (delta / 100.0)
            scale = max(0.1, scale)

            context.active_object.scale = (scale, scale, scale)

        elif event.type == 'LEFTMOUSE':
            return {'FINISHED'}

        elif event.type in {'RIGHTMOUSE', 'ESC'}:
            context.active_object.scale = (
                self.initial_scale,
                self.initial_scale,
                self.initial_scale
            )
            return {'CANCELLED'}

        return {'RUNNING_MODAL'}

    def invoke(self, context, event):
        if context.active_object:
            self.initial_mouse_x = event.mouse_x
            self.initial_scale = context.active_object.scale[0]

            context.window_manager.modal_handler_add(self)
            return {'RUNNING_MODAL'}
        else:
            self.report({'WARNING'}, "No active object")
            return {'CANCELLED'}

得:互動運算應鼠、左確、ESC 撤。

敗:察事型、確模處加、理無活物。

六:加件包

組碼為可裝加件:

bl_info = {
    "name": "Custom Tools",
    "author": "Your Name",
    "version": (1, 0, 0),
    "blender": (3, 0, 0),
    "location": "View3D > Add > Mesh",
    "description": "Collection of custom modeling tools",
    "category": "Add Mesh",
}

import bpy

from .operators import OBJECT_OT_generate_spiral

classes = (
    OBJECT_OT_generate_spiral,
)

def menu_func(self, context):
    """Add to menu."""
    self.layout.operator(OBJECT_OT_generate_spiral.bl_idname)

def register():
    for cls in classes:
        bpy.utils.register_class(cls)

    bpy.types.VIEW3D_MT_mesh_add.append(menu_func)

def unregister():
    bpy.types.VIEW3D_MT_mesh_add.remove(menu_func)

    for cls in reversed(classes):
        bpy.utils.unregister_class(cls)

if __name__ == "__main__":
    register()

得:加件經 Preferences 裝、運算現於菜。

敗:察 bl_info 式、驗 Blender 本需、確諸類列。

七:資導程序生

自外資生幾何:

import csv
import json

def create_from_csv(filepath):
    """Generate objects from CSV data."""
    with open(filepath, 'r') as f:
        reader = csv.DictReader(f)

        for row in reader:
            name = row['name']
            x, y, z = float(row['x']), float(row['y']), float(row['z'])
            scale = float(row.get('scale', 1.0))

            bpy.ops.mesh.primitive_uv_sphere_add(location=(x, y, z))
            obj = bpy.context.active_object
            obj.name = name
            obj.scale = (scale, scale, scale)

def create_from_json(filepath):
    """Generate scene from JSON configuration."""
    with open(filepath, 'r') as f:
        config = json.load(f)

    for obj_config in config.get('objects', []):
        obj_type = obj_config['type']
        location = obj_config['location']

        if obj_type == 'cube':
            bpy.ops.mesh.primitive_cube_add(location=location)
        elif obj_type == 'sphere':
            bpy.ops.mesh.primitive_uv_sphere_add(location=location)

        obj = bpy.context.active_object
        obj.name = obj_config.get('name', 'Object')

        if 'material' in obj_config:
            mat_name = obj_config['material']
            mat = bpy.data.materials.get(mat_name)
            if mat:
                obj.data.materials.append(mat)

得:物自外資檔生。

敗:驗檔式、理缺欄、予默值。

  • 本於 Blender Python 行無誤
  • 程序幾何如期生
  • 動鍵插於正幀
  • 批操處諸檔
  • 自運算現於搜並正行
  • 模運算應鼠/鍵
  • 加件裝/卸潔
  • 外資檔正解
  • 誤理覆邊例
  • 碼循 PEP 8

  1. 加件循入:用相對入、慎組模
  2. 運算名:bl_idname 必小寫單底線(類.名)
  3. 屬型:用正 bpy.props 型(FloatProperty、IntProperty 等)
  4. 境訪:非諸運算於諸境(視口 vs 繪)
  5. BMesh 清bm.to_mesh() 後恆 bm.free() 防漏
  6. 動鍵時:幀始 1 非 0
  7. 驅式誤:驗式、用安名空
  8. 模運算阻:勿阻於 modal()、用非阻操
  9. 加件裝徑:置於 Blender scripts/addons
  10. 本相容:API 跨本變、文錄需

GitHub 仓库

pjt222/agent-almanac
路径: i18n/wenyan-ultra/skills/script-blender-automation
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

相关推荐技能

content-collections

Content Collections 是一个 TypeScript 优先的构建工具,可将本地 Markdown/MDX 文件转换为类型安全的数据集合。它专为构建博客、文档站和内容密集型 Vite+React 应用而设计,提供基于 Zod 的自动模式验证。该工具涵盖从 Vite 插件配置、MDX 编译到生产环境部署的完整工作流。

查看技能

polymarket

这个Claude Skill为开发者提供完整的Polymarket预测市场开发支持,涵盖API调用、交易执行和市场数据分析。关键特性包括实时WebSocket数据流,可监控实时交易、订单和市场动态。开发者可用它构建预测市场应用、实施交易策略并集成实时市场预测功能。

查看技能

creating-opencode-plugins

该Skill帮助开发者创建OpenCode插件,用于接入命令、文件、LSP等25+种事件。它提供了插件结构、事件API规范和JavaScript/TypeScript实现模式,适合需要拦截操作、扩展功能或自定义事件处理的场景。开发者可通过它快速构建响应式模块来增强OpenCode AI助手的能力。

查看技能

sglang

SGLang是一个专为LLM设计的高性能推理框架,特别适用于需要结构化输出的场景。它通过RadixAttention前缀缓存技术,在处理JSON、正则表达式、工具调用等具有重复前缀的复杂工作流时,能实现极速生成。如果你正在构建智能体或多轮对话系统,并追求远超vLLM的推理性能,SGLang是理想选择。

查看技能