Recreated based on PsychProd's basketweave/quasi-vasemode concept here, as well as its remixes (e.g. #5204468, #5168851, #4976590) and the cited prior works (e.g. #4508179).

The idea is great, but perhaps it was limited by the method of generating the cyclic geometry needed for the container walls. There were complaints of weak walls for obvious reasons, so i printed a small one. I tried to flex it, and it split. That made me doubt the strength of the larger containers I had intended to print. I also wanted to modify the models for my needs, but the STL files are so full of errors that I can't import any of them into openscad, and it's beyond my meshlab skills to repair the models.

Also included is an SCAD file for rectangular boxes of the same fashion, based around a similarly inexpensive hygrometer model. I had not seen any implementations of the concept in prismatic containers at the time, so I figured it would be best to add one.

Only after I completed both files did I find this hexagonal model. While MysticalDork has tried to resolve the same limitations in PsychProd's design, it's still just a set of STL files.

I don't know of any prior published basketweave container models which are parametric, but now we have some. I've covered round canisters and rectangular boxes. If you want something else, these files contain modules to generate the weave pattern as either flat panels or cylindrical tubes. You can copy that and build what you want.

The SCAD files have lots of printing and design notes. Included is a library of premade STL files for 25, 40, 50, and 70mm dia canisters, as well as a handful of rectangular box designs with assorted lids and accessories.

BENEFITS:

• It's parametric. (gotta love openscad) This makes it configurable for other applications unconstrained by spool hole sizes.
• Square-weave design avoids splitting and sagging problems. Strength is further improved by overlapping the weave layers by a small amount.
• Improved strength allows for thinner walls and a tiny bit more interior volume.
• Result is a tiny bit lighter/cheaper. (as if that matters)
• For the typical 50mm versions, the jar mouth is wider. (nice if you also use packets)
• Bottom and lid ventilation is more open and still prints fast with comparably few retractions (27 vs 22 for layer 1 iirc) compared to 300+ for earlier lattice designs (e.g. #4508179).
• Bottom ventilation isn't prone to spreading like long angled slots do (see photos).
• Fast print time is overall comparable to the originals.
• The STL files don't have nearly as many errors. :)
• Includes escape slots to make it easier to remove the meter if necessary (similar to #4976590).
• 80 degree thread profile should be fairly forgiving to print.
• Chamfers help with deplating.

CAVEATS:

• It also renders horribly slowly. (gotta love openscad)
• There are also probably better thread libraries (maybe something with a decent blunt-start), but i want to maintain compatibility back to version 2015.03.
• The generated STL files may initially poorly optimized. See notes on how to clean that up. After cleanup, they're usually smaller than the original basketweave model files (for nfaces = 100).
• Due to the interlayer support areas, there are more retractions in the walls. Idk if that truly matters, but it's adjustable.
• I have seen some stringing in the top/bottom mesh with my print settings, but it doesn't really hurt anything.
• The very square webbing pathing might limit the max travel speed at which these can be printed, though it's probably fair to assume that the curved overhangs of the original model are at least as much of a speed handicap.
• The rectangular box designs might not be as fast as a long-bridge box design, but they're probably stiffer and stronger.
• I haven't printed a ton of these with different filaments or different printers, so thread fitment may need to be adjusted and there may be other issues i can't foresee.

TL;DR: They're customizable and a good balance of speed, printability, stiffness, strength, and minimal material usage. They're good and cool and will make you look smart. All your friends will be impressed, and you'll finally be able to sleep restfully after all these years.

EDIT: As a matter of course, now that I've posted these files, the other implementations of basketweave desiccant containers finally come out of the woodwork. It's almost like the behavior of search engines is less dependent on the query string than it is on collected user data. Imagine that.

Regardless of prior attempts to improve on PsychProd's concept, this is the only known implementation in SCAD, and it's the only one to provide both canisters and boxes. It's also going to produce the strongest containers for the material used, since none of the other square-weave models use any sort of web overlap (referred to in the files as ilh (interlayer support height)).

Regarding the influence of ilh on strength, I did 3-point bending tests on 20mm wide panels (40mm support distance). When compared to ilh=0, ilh=0.4 increased the load required to break 3mm thick samples by 50%. The strength of thinner 2mm samples was more than quadrupled. In other words, webbing overlap/intersection makes panels significantly stronger.