Stemlight v1.1.10-beta

[ncolyer11]

Features

This double digits update brings some final features and quality of life changes to the Playerless Core Tool, along with a heap of bug fixes across the whole program:

• The simulated annealing algorithm, used for optimising dispenser placements, now takes into account: cycles, blocked blocks, and cleared dispensers:
  

  • If you optimise a dispenser layout with at least 1 cleared dispenser, then the program will find the most optimal layout including cleared dispensers, leading to even faster layouts in some cases (more research and science :) is required here to deduce when cleared dispensers are more effective than not).

• Added a new config menu with the option to calibrate the run times of the optimisation algorithm to match the speed of your computer more closely:
  

  • If you run the program from the terminal, you can see it logging how much the calibration has refined the run time:
    
  • Here it goes from being 129% off from the desired run time of 15 seconds, to 10%, and then just 0.7% off.

• Added (and then removed lmao) the option to optimise for max wart block output instead of just for max fungus output:
  

  • It turns out a max fungus layout is also a max wart block layout in every case I tried, so this feature was removed from the front end but all the logic is still in the code for any possible future addition.

• A biproduct of the above experimental addition is that the program is now able to display the wart blocks generated if you were to bone meal the expected distribution of fungus after the elapsed amount of cycles have taken place:
  

  • Meaning you can now also go full circle and find the net bone meal return by calculating bone meal from composting the wart blocks, given some empirically set blast chamber efficiency (mhmm, application of feedback network theory to calculate tnt interference on dispenser layouts soontm??).

Bug Fixes

• See comment below.

Known Issues

• The program can become seemingly unresponsive when handling intensive tasks such as optimising or calibrating dispenser layouts. This is due to everything being done on just the 1 thread, I may pursue using dual or multi-threads in the future and adding a loading bar but this is a non-trivial task. For now just be patient! (with both the program and me lol).

---

This discussion was created from the release Stemlight v1.1.10-beta.

[ncolyer11]

Been working on this since May so hope some research and development can be done with it now that it's all quite refined. This should be the last v1.1.x-beta patch fingers crossed!

[ncolyer11]

Bug Fixes

• Blocked blocks don't get removed after placing a dispenser over them.
• Blocked blocks aren't displayed after switching nylium types.
• The calculated wart block figure is lower than it should be in some configurations.
• Complimentary probabilities for wart block distribution don't respect y-distribution.
• Exported heatmap probabilities don't respect blocks generated earlier than them.
  • e.g. Wart block probabilities aren't lowered due to a trunk from another fungus generating in the way of the hat and blocking potential wart block generation spots.
• Cumulative block total grid used to calculate total block distribution for external heatmaps double counts some values leading to skewed, and inaccurate, distributions and results.
• Column calculations when writing to spreadsheet on excel are calculated incorrectly due to swapped x and z coordinates.
• Some alternate dispenser placements are left out/cutoff in the outputted alternate_dispensers.txt file due to x and y (z) coordinates being switched (if I had a penny for every time there was a bug relating to width and length being mixed up...).
• Blast chamber efficiency only works if set to 100% due to faulty string to int conversion.
• Fixed laggy/clunky behaviour when resizing the grid due to frivolous wart block calculations with 0 dispensers placed.
• Net bone meal doesn't take into account composted foliage.
• Tick/check mark symbol indicating selected amount of decimal places is cut off in numerous menus throughout the program.
• Stemlight 'Creation' date refers to the first commit rather than the first official alpha release.
• viable.txt name for alternate dispenser placements is unclear.