A few 3DCodeData factory scripts fail to execute in Blender 5.0.1

[llada60]

Thanks for the great benchmark!

I ran a full executability sweep over all Blender 5.0.1 scripts with names ending in factory; nearly all ran cleanly, but three failed.

Environment

• Blender 5.0.1

Sweep result: of all factory .py + _geo.py scripts, only the 3 below fail to execute.

---

1. CanFactory/CanFactory_geo.py

Symptom

File "CanFactory/CanFactory_geo.py", line 473, in <module>
    (make_text_label, 1.0),
NameError: name 'make_text_label' is not defined

Fix
Added the missing """ at line 362 and line 388 to properly close/open the docstrings, so make_text_label (and the neighboring material constructor) parse as real functions again. Root cause was a docstring-quoting issue, not the call site.

---

2. LeafMapleFactory/LeafMapleFactory.py

Symptom

File "LeafMapleFactory/LeafMapleFactory.py", line 340, in make_nodegroup_sub_vein
    mix.inputs['Fac'].default_value = 0.9
KeyError: 'bpy_prop_collection[key]: key "Fac" not found'

Debug — shared root cause with #3
Both scripts mis-address the ShaderNodeMix node (Blender 4.0+ replacement for the legacy ShaderNodeMixRGB). On the new node the socket names changed and are duplicated per data-type, so name-based lookups like 'Fac' / 'Color1' either don't exist or hit the wrong-typed variant. Probing ShaderNodeMix with data_type='RGBA' in Blender 5.0.1 gives:

INPUTS:  [0] Factor(Float)  [1] Factor(Vector)  [2] A(Float)  [3] B(Float)
         [4] A(Vector) [5] B(Vector)  [6] A(RGBA)  [7] B(RGBA)  ...
OUTPUTS: [0] Result(Float)  [1] Result(Vector)  [2] Result(RGBA)  ...

So for RGBA mixing the correct sockets are Factor=inputs[0], A=inputs[6], B=inputs[7], Result=outputs[2] — and these must be addressed by index, not by name (names are ambiguous/duplicated).

Here make_nodegroup_sub_vein used mix.inputs['Fac'], mix.inputs['Color1'], mix.inputs['Color2'], and mix.outputs[0] — all legacy MixRGB socket names/indices.

Fix (make_nodegroup_sub_vein, ~L339-346)

mix = ng.nodes.new('ShaderNodeMix'); mix.data_type = 'RGBA'
mix.inputs[0].default_value = 0.9                     # Factor
ng.links.new(noise.outputs['Color'], mix.inputs[6])  # A (RGBA)
ng.links.new(comb.outputs[0],        mix.inputs[7])  # B (RGBA)
...
ng.links.new(mix.outputs[2], vor.inputs['Vector'])   # Result (RGBA)

3. BoulderPileFactory/BoulderPileFactory.py

Symptom

File "BoulderPileFactory/BoulderPileFactory.py", line 235, in make_mountain_material
    ao_mix.inputs[1].default_value = (0.0, 0.0, 0.0, 1.0)
ValueError: bpy_struct: item.attr = val: sequences of dimension 0 should contain 3 items, not 4

Debug — same ShaderNodeMix socket-layout issue as #2
The code assigned an RGBA 4-tuple to ao_mix.inputs[1], but on ShaderNodeMix inputs[1] is the Vector Factor (3 components) — hence "should contain 3 items, not 4". The RGBA color inputs are inputs[6]/inputs[7]. Three ShaderNodeMix (RGBA) blocks were affected — make_mountain_material (~L233) and two in make_stone_material (~L406, ~L434) — each using inputs[1]/inputs[2] for colors and outputs[0] for the result.

Fix — addressed all three blocks by index (Factor=inputs[0], A=inputs[6], B=inputs[7], Result=outputs[2]):

Block 1 — make_mountain_material (AO mix):

ao_mix = nt.nodes.new("ShaderNodeMix"); ao_mix.data_type = "RGBA"
ln(_fac_output(amb_occl), ao_mix.inputs[0])             # Factor
ao_mix.inputs[6].default_value = (0.0, 0.0, 0.0, 1.0)   # A (RGBA)   was inputs[1]
ln(ramp.outputs[0], ao_mix.inputs[7])                   # B (RGBA)   was inputs[2]
...
ln(layered_color.outputs[2], color_mul.inputs[0])       # Result(RGBA)  was outputs[0]

Block 2 — make_stone_material (cr2/cr1 mix):

mix = nt.nodes.new("ShaderNodeMix"); mix.data_type = "RGBA"
ln(cr2.outputs["Color"], mix.inputs[0])                 # Factor
ln(cr1.outputs["Color"], mix.inputs[6])                 # A (RGBA)   was inputs[1]

Block 3 — make_stone_material (stone_base_color mix):

stone_base_color = nt.nodes.new("ShaderNodeMix"); stone_base_color.data_type = "RGBA"
ln(mix.outputs[2], stone_base_color.inputs[0])                   # Result(RGBA)->Factor  was outputs[0]
stone_base_color.inputs[6].default_value = (0.0, 0.0, 0.0, 1.0)  # A (RGBA)   was inputs[1]
ln(cr_base.outputs["Color"], stone_base_color.inputs[7])         # B (RGBA)   was inputs[2]
...
ln(stone_base_color.outputs[2], bsdf.inputs["Base Color"])       # Result(RGBA)  was outputs[0]

---

Thanks for reviewing this. The affected scripts now execute successfully in Blender 5.0.1.

[gaoypeng]

Thanks for verifying the data! I'm gonna fix them and update the scripts in Hugging Face.