Boxes.py

As people surely have noticed the web instance was pretty slow at times. The issue was not really too much load but some weird latency or time out problem. As my hackerspace has setup a new server I asked if they would host Boxes.py. So now a new instance is available at https://boxes.hackerspace-bamberg.de. I added a redirect to the old location so you end up there no matter what. For now this is still in testing. Please report any issues you have here or on the GitHub project. If everything works out fine I will change all the links to the new instance.

Update: Looks like everything is working fine. I switched over the URL in all places I could think of. So the web instance is now officially move to its new home. Please enjoy the much quicker response times!

Being able to put more time into the project have I looked into my local repository and it is full of half finished projects - most of them new "box" generators. I have been working on getting some of them in shape for inclusion. Additionally there where a few needs that popped up that also needed instant addressing. Funny enough a lot of these new generators come in pairs.

So I did a roll holder for the kitchen after the old, 3D printed one broke:

Well, it is mounted to the wall... so a "Wall" variant was needed:

After finishing both I found another 20% finished holder in my working dir. It is more ambitious with 3 layers on each side that allow locking in the axle by rotating the middle layer. No idea why that didn't get finished. May be one day the two above get a fancy sibling...

For another project I need to 3D print a small part (yes, I know, heresy) but the filament didn't come with it's own spool. So, I did a spool that can be separated into two parts using the BayonetBox as a basis: 

And as I was at it it was just too easy not the strip it down to a one piece spool (and then add some fancy reinforcements):

One generator I started a long time ago was one for generating Tetris blocks. At the time this wasn't that easy and quite a bit of work. Nowadays this is much easier as we have the polygonWall method that gets a list of lengths and angles and handles the rest. There also is a companion method to create all the side walls from the same list. So the only tricky part was getting the lengths right as we do want the walls to be within the perimeter of the block structure. I am optimistic they fit, but I have not yet cut them out. So if you feel adventurous and "need some Tetris blocks in your life" please donate a few good pictures.

Another "thing" that I started long ago and started over with more than once is my Seven Segments feature. The idea is to use a LED stripe on its side and wind it through the seven segments. So you can create a seven segment display without much soldering or special circuit boards and from semi small to quite big. The obvious use case is a seven segment clock. Both of them are still flagged "unstable". But I am not sure if I get to build one any time soon. So if one brave soul is looking an LED stripe fueled clock project let me know.

So far my backlog is all but empty and I still have more ideas. I just started with a wire straightener holding two rows of v-groove ball bearings. I have a cryptic font that is mostly finished with each letter is just a few shapes that are only legible when surrounded by a border. I wanted to create a calendar with them, preferably perpetual or even some clever mechanical one. I saw a cube box at the Fablab of the University of Oulu that is cut directly between pieces without creating gaps. And I have a half working one for arbitrary sizes.  I have toyed a bit with flexures and thought about doing a laser cut microscope like the OpenFlexure Microscope although with very different kinematic. And I have a dozen more that are not much more than the first few lines of code and I am not confident enough to publicly admit here.

It's been a while since the last proper log entry...

The last few years were not that great for many things - including Boxes.py. But even with things going slow or not moving anywhere for months a lot of changes have piled up over this long time. Also it looks like this year is a lot better and things have actually already picked up speed in December of last year. Not only have I been able to do more but several old and new contributors have shown up and gave the project a very notable boost. To deal with the long backlog of new features and box generators while there is so much new stuff coming in I decide to mix in the old with the new.

Wall Mounted Boxes and Tool Holders

After a brief discussion here in the comments in early 2022 I factored and renamed the SlatWall group to WallMounted. While slat walls are fancy not many people have them. The whole group is only there because a few panels fell into the hands of our hacker space. Now the generators can have different styles of hooks and also work with DIN rails or French Cleats. It is relatively easy to add new ways of hooking the boxes and tool holders to different things. So it you have ideas open a ticket on GitHub - or start a discussion in the comments below.

There are also few new generators in this group:

• SlatWallPinRow allows you to hang stuff on the wall.
• WallStairs allows you to design your own tool holders.

and I recently added a few more sample images.

On the search for the right box generator

Caleb Crome is one of our new and very prolific contributors. One theme of his work is making finding the right box generator easier. He wrote a Gallery page that shows all sample images side by side for a quick overview.

The second, complementary new feature is a search field on the menu page. It shows all matching box generators and matches the English name but also the translated name and short description. This is all done with some - not even too complicated - Java Script magic directly in the browser.

A third feature is not merged yet. It will allow getting or adding QR codes with a link and settings of a generated box. That way it is easier to pass the settings around with a mobile phone or to figure out how to get the drawing from an already laser cut piece.

These new features are not all Caleb has done so far. Prepare to hear more from him as soon as we get into new generators.

The last few months complaints about poor performance and an increasing number of 502 errors have been piling up. The later were especially puzzling as it means there is a problem in the communication between the web server and uwsgi instance in the container. While there was some load on the machine everything looked fine at the first glance and there was nothing in the logs that hinted at a bug.

At the same time there were similar complaints for the web2py application that runs on the same machine. Well, actually this thing is a VM that runs on a server of a friend of a friend that does a few other things and is only maintained to the bare necessity. When I put Boxes.py there the question I asked myself was not "How do I host the greatest box generator on the internet?" but "Now that I hacked argparse to produce web pages how can I show this to a few people?"

Over the holidays I had time to look a bit deeper. Turns out the web2py app offers the data with different sorting and filtering and a horde of web crawlers eating it all up. Adding a robots.txt got rid of them and dropped the load from 0.7 to below 0.1. I may loosening this up a bit in the future but for now I just banned bots completely - including from Boxes.py.

Looking at the logs from Boxes.py it was very apparent that there are even more connections coming in. Most of them of course are these nice pop-up images in the menu. While I still didn't have access the the main machine it dawned on me that the Boxes.py instance probably gets throttled down due to too many request. This would create timeouts and though the 502 errors. To fix that I moved all the pictures to the GH (documentation) pages and adjusted the links to point there. As this would be very annoying for people that run Boxes.py locally (me) this change affects the wsgi mode only. Yes, the boxesserver should grow a config file for things like that but I just couldn't be bothered.

So performance has notably improved and I hope you all can now create your boxes without seeing 502 error messages. Looking into the access log it turns out there are quite  a lot of you. There are tens of thousands of request per day - even with the pictures moved elsewhere. This is kinda cool but also kinda scary...

...until someone makes an insert for Agricola - then it is even more fun ... and even more games:

Kudos to Guillaume Collic who wrote this master piece. I especially like how he made use of already existing code and pieces and combines them with new edges types and new ideas. It re-uses the Divider Tray and regular boxes but combines them with new solutions like cutouts to make gripping the tiles easier. A very neat idea (but a bit difficult to see in the picture) are the sliding dividers in the trays for the wooden pieces. They are prevented from tumbling around by small triangular feet but are otherwise free to move left and right depending on how much space is needed in the compartments.

The two box types holding the tiles clearly deserve to get a generator on their own. Guillaume already has ticket opened to make it easier to re-use complete generators by others. Splitting these boxes out might be a good exercise to create the infrastructure still missing.

If you are on your way to your laser cutter you should probably stick to the default material thickness of 3mm or something close. While not stated in the generator description I have the suspicion that adding too much width with thicker material might end up with an insert that cannot be - well - inserted into the box.

Inserts and organizers for board games is a very welcome new topic in the Boxes.py. I hope at some point we can make this a category in the UI and rescue this gem from the "Misc" section. I have the hope that Guillaume may contribute some more. But may his works also inspires others. The more pieces and features there already are the easier such task becomes. 

Often just a generic box is just not enough, but you need something that can house the electronics of your project. The ElectronicsBox was the first variant with this in mind. But in the end it is just a box meant to be bolted down and screwed shut:

QuitButton needed something else: A housing with a slanted front to mount a LCDisplay and a few buttons.

As a first stop gap solution (other than just turning a pentagon box on it's side) I implemented Console using the relatively new polygonWall() code that turns a list of lengths and angles into a side wall. The actual win here is that there is a .polygonWalls() method that is rendering all the walls between the sides. The result is still a bit simple, but  at least has the right shape:

But as with a lot of other boxes there is a severe lack of lockable doors or hatches. And while non rectangular finger joints are cool they are rather distracting around the front panel.

So here is an improved Console2:

There are a few variants. You can choose with or without removable back wall, with or without removable front panel and the front panel - as seen here -  clean and glued or attached with finger joints on the sides.

The panel is held in place by spring loaded tabs:

The two tabs per side and both springs being one single, laser cut piece.

The back wall is hooked in at the bottom and has two flush latches at the top:

The latches are made small enough that they cannot be opened or closed by hand but need a small tool like a screwdriver. This is on purpose to avoid people accidentally sticking their fingers into the electric parts inside. To prevent the latches from opening they have an integral leaf spring and a very small recess that locks them in place when closed. This way they can only be re-opened when pressed in a little bit against the spring.

Heise.de is the German speaking IT news site. The Heise publishing house also issues the Make magazine. So via the Make news feed this article in German (translated by Google) made it to get global feed at heise.de.

There is nothing in the article regular readers of this project log don't already know about. But getting this level of exposure is quite nice and made my day.

Herzlich Willkommen an alle Heise-Leser!

So the new back end is finally merged into the master branch. If you update your own installation you need to install the affine Python module which is available both via pip but also in most distributions. Cairo is no longer needed. This should make installation on Windows much easier.

The SVG produce by the new back end is also much better than the old one. Artifacts like repeated moves are filtered out and the outline of parts is much more continuous. The weird inverted arcs in inner corners are - for the most part - replaced with smooth Bezier loops. This should reduce acceleration in the corners and allow for faster cutting.

I am still pondering whether it is worth adding a native DXF back end. While there is only very few commands needed by Boxes.py the format does not look like fun. I am also scared that it won't support the text the way we might need. For now the output still uses pstoedit for converting to DXF.

Boxes.py is using the Cairo graphics library to generate SVG and PostScript (which is then converted to more formats with pstoedit). This has always been a pain. There is little control over the actual output. Getting the library to run on Windows is a real pain and creates a continuous stream of tickets, questions and complaints.

So there has always been the thought of replacing Cairo with something else. But thoka finally put in the work to put together a native Python backend. After me not finding the time to take care of the PR for much too long it is now ready to be tested. Please check out https://www.festi.info/boxes.nocairo/

Beside not needing Cairo it has the following improvements:

• Continues outlines for most parts
• Better grouping within the SVG file
• Improved burn correction with inner corners now being a loop instead of changing direction twice

Other changes

• Text rendered as text not as glyphs
• <deleted>No use of hlines and vlines (yet)</deleted>

Please test with you tool chain and report issue back. Either here as comments or in the GitHub PR or a new ticket.

If you want to test locally there is a nocairo branch in the main repository.

Thank you!

Update:

Turns out I had to redo the text "rendering". The text was never that well thought out as Boxes.py is not a word processor. I now do know much more about PostScript and it's font system than I ever wanted...

To give the code more exposure I now moved the main instance (https://www.festi.info/boxes.py/) to the nocairo branch. The Cairo back end is still available at https://www.festi.info/boxes.cairo/ . But I expect to call the new back end final and ditch Cairo completely soon - if there are no major issues popping up.