Compact Universal Factory (Pre-Wires)

This is my take on the universal factory; by that I mean a factory which can create (almost) every shape of the game with minimum amount of effort to switch from the produciton of one shape to another. This is particuarly usefull for the freeplay, during which you're ask to create complex randomly generated shapes. The principle of my factory is simple: produce a full lane of every possible basic corner shape of the game (4 shapes * 8 colors * 4 orientations = 128 lanes), have a simple mechanism to pick the coners needed, and stack them, to output a full lane of the shape.
[The guide tab may not be the place for this, but I couln't find any other place to make a detailed post]
[Sorry for the approximate english]

Overview

This is what the factory looks like. The goal was to make it as compact as possible, for the beauty of it but also to limit lag (With two of these factories, the game is running at 5 FPS on my laptop..). It's divided in multiple parts showcased below.

We first want to have the 128 full lanes of each basic corner like so:

Cutters

To produce each corners, we use the quad cutters on the four basic shapes of every color variant. This part was probably the toughest to design, since it has to output a different shape on every single line.

Painters

Using the double painters divides by two the required amount of colors, and also allows an efficient use of space.

Color mixers

The color mixers are only needed for the secondary colors, and the white which takes the most space to make.

Supplies

Note that the factory takes as input 8 full lanes of each basic shape and color, which i find quite elegant. The colors are coming from the left and the shapes from the top.

Basic shapes maker

This is independent of the factory, but here is the compact design I used to produce 8 full lanes of rare basic shapes like the stars or the windmills, from 16 lanes of a shape containing a half of it:

Shape selection

Once we have the 128 lanes thanks to the combination of machines shown above, we have to select the ones we want, and put the trash the others so nothing gets stuck.
The design I found is quite big but very easy to operate: to select a shape, you just have to replace a tunnel by a belt. The ones never selected end up in a trash.
A shape can have up to 4 corners * 4 layers = 16 corners, so we need 16 of these selection lanes.
Some shapes involve the same corner on multiple layers; in this case, you can use a similar corner shape in a different orientation and rotate it once or twice in the place of the belts.

Stackers

The stackers first combine two corners in halves, which are then combined into full shapes. The last part on the right stacks the full shapes onto each other, and outputs the result at the bottom. Often, corners or a whole layer as to be left blank: in this case, there is just one belt to had and the shapes will get past the stackers without getting stuck into them (can you see the discontinued belts? this is where you have to add a belt).

Conclusion

You then end up with a full lane of the shape of your choice! Here it is, with my two factories, bringing in a shinny 3.6k/m of it, completing the level under 20 minutes.


We now have to discuss the limitations of this factory: this is not actually a completely universal factory, there is a certain category of the ~10^14 possible shapes that it can't make; for example, the shape of the level 18 (the logo of the game) since it involves more cutting during the stacking phase. There are so many edge cases related to this kind of shapes that it would be very tedious if not impossible to make the truly universal factory. Furthermore, the level 18 is the only one requiring such a shape.
Another limitation would be if you want a shape made from more than 4 of the same corner shape : you just wouldn't have enough with this factory, and splitters would be needed, significantly reducing the output.
Finally, despite this factory being compact, it lags the game a lot, to the point that I couldn't make more than 2 of it.
I would have liked to include the save or at least export the base as an image, but the game won't do it (on both Windows and Linux), probably because it's too heavy.

Anyway, I hope you found this post interesting, please give me your ideas of improvement and feedback if you have any! :D
Thanks to the dev for such a brilliant and fun game!

[Dragonflynipple]

the shapez* are random but it works



*haha

[Dragonflynipple]

and more compact too

[Dragonflynipple]

there is a waaaaaaaaaaaay easier way to do this

[Aaronstar]

2 is finally out! gonna have to rebuild what i made to fit with the new progression, but ill post more in a day or two!

[Arller2828]

Seeing this brings me so much joy :')

[Aaronstar]

managed to halve the footprint because it turns out buildings can be put on the second layer and both layers can be used to transmit items to other platforms.
currently at:
163k bpu
41 platforms
66 platform units
10.2k buildings

[Aaronstar]

got a basic thing thrown together. if people are interested, i could do a writeup about it. currently at:
240k bpu
136 platforms
193 platform units
13.8k buildings
only does shape cutting, selection, and stacking. may need to shuffle some things around to make painting fit better when it gets released.

[Aaronstar]

im gonna start trying to reproduce this in the shapez 2 demo and save it as modular blueprints. i woln't be able to do colors yet, but might be able to get a start on it

[godslayer]

I don't think having 128 different corners is entirely necessary. While it is easier you could rather have another cutter phase to reduce lag. It's also pre-wires which could also mean you could you use rotators.

[TheGamer]

need to be level 11