@neset

refactored a little bit of Runner to make it more consistent.
changed the actions to a parsed json rather than hardcoded to make it easier for others to use rp, and so that its easier down the line

took a TON of time and learning but i made a new file and some new methods in order to accurately display status!
it updates independently every second, rather than every loop. and, everything there is actual information, nothing pre-written!

todo:

• bug fix
• make the ui more interactive and user-friendly
• add more commands
• idk

wrote a TON of code in order to get to this point. drawing the status lines is now halfway there (there’s front end but no back end). i also implemented it in such a way that (hopefully) making this look pretty wont mean rewriting a lot of code.
onto writing the functionality!

added a stop command which allows me to stop a command that i ran from the tui!

commands now run in their own thread! this makes sure that i cant run multiple commands on the pico at the same time.

finally have commands running with their output!

moved everything into their own files, to keep stuff more organized.
learned what .hpp is.
now i can move on to actually making it functional !!!

did a lot of backend changes! i added a lot of error handling, changed the terminal and each section to classes, and added colors. i also added titles to each section.

i made a vector with actions and keybinds to each of those actions! i can also go up and down with arrow keys and use enter or spacebar to select that action. i also made one of my windows into logs, and made sure that it only shows the latest lines!

learning c++ asw as how to make a tui!
i made something that splits my windows into 3, outputs a changing variable (time), and takes my keyboard inputs and does stuff with them!

im hoping to learn more UI and then start on the actual functionality.

bug fixes

ready to ship!

• added an intro explaining what this is
• added cookies so u can track ur own data
• made my data global for everyone to see
• minor bug fixes

todo:

• add more drinks

i designed the board, including connections for a raspberry pi zero 2 w for computing, a dac, a small screen, a rotary encoder w click and 3 buttons.
next:

• print pcb
• acquire components
• solder/assemble
• program
• iterate

OKAY so i kinda forgot to make updates…
but here’s what ive done so far:

• ran hours of simulations to test draw rates of other existing bots
• made the framework so i can start training
  next:
• train the neural network!

LOTS of hours put in. in summary:

• made uci, which allows me to spin up instances of existing chess engines for verification & simulation
• made fastboard, a fork of pychess’s board class, which is much faster at checking for draws
• made a baseline “dumb” bot that looks at legal moves and sees which one would increase stockfish’s evaluation of a draw the most, then makes that move.
• made a game class that allows bots to play against each other. it takes ~70s for a game between the baseline “dumb” bot & stockfish
• made a pgn class that allows me to save what happened in a game and the result, which will help for data analysis down the line

image is of a real chess match that my bots played, saved and then opened in lichess. you can see in the bottom right that it ended in a draw, which is a good thing

had a lot of brainstorming and ended up designing these 6 letters. i’m unhappy with a few but i will go back and fix later

spent lots of time and a lot of headaches improving email speeds. its now a bit slower than gmail but a lot faster than previous.

read receipts on emails that carry over to other email apps. also, i can now write and send emails :)

YAYYYYY!!!
i have REAL emails from my google account!
still a lot to do, but this is great progress :)