bitbutter

## Explaining My Work (writeups, status updates, summaries of what I did)

When I write to the user ABOUT work on a codebase — what I changed, what I found, what a test proves, why something failed — the reader was NOT in the work with me and does not share the vocabulary I built up doing it. This is a hard default:

- **Meaning before name.** Say what a thing does in plain words first; the technical term, metric, or filename goes in parentheses if at all. "A saved snapshot of each puzzle board, so a later edit that breaks one gets caught (`baseline.json`)" — NOT "21 goldens in baseline.json".
- **Plainify.** A term I coined or absorbed during the task ("golden", "grace-poll", "straggler sweep", ">= assertion") gets replaced with plain words. No undefined insider terms.
- **A number is evidence, not the point.** Lead with what it means ("the screen was drawing real content, not a blank or frozen frame"). Never a bare "0.207 → 0.06" with no labels.
- **Say why it matters to the user.** Tie each item to somethi

bitbutter

# Data Model

Defines every data structure in the project. A rebuild AI reads this to recreate the exact project file format and in-memory model.

## Project File Format

Serialized as **MessagePack** (via `rmp-serde`). The schema below is defined in Rust-struct notation but represents the canonical format.

---

bitbutter

@tool
extends EditorScript

## Auto-generate collision polygons from tile pixel alpha.
## Run from Script Editor: File > Run (Ctrl+Shift+X)

## Path to the TileSet resource to process.
const TILESET_PATH := "res://game_specific/resources/tilesets/greentileset.tres"

## Physics layer index to assign collision polygons to.

bitbutter

'''
Based on https://shaggydev.com/2025/06/12/godot-awaiting-signals/

# usage
var sigs=[
	get_tree().create_timer(0.1).timeout,
	get_tree().create_timer(0.1).timeout,
	get_tree().create_timer(0.1).timeout,
]

bitbutter

class_name SignalWaiter

signal AllFinished

# Internal variables
var _active_nodes = {}
var _signal_connections = {}
var _callables = []

## External method to add a node, its callable, and the signal to listen for.

bitbutter

import json
import spacy
import zipfile
import os
import pickle

# File paths
zip_path = r"pathtotwitterarchive.zip" # Path to zipped twitter archive
extract_path = r"somepath\twitter_data" # Path to extract twitter data
embeddings_path = r"somepath\tweet_embeddings.pkl" # Path to save embeddings

bitbutter

import chromadb
import openai
import os
from tqdm import tqdm
import tiktoken
from chromadb.errors import NotEnoughElementsException
import re
from colorama import Fore, Style

# Instructions (assumes Windows OS)

bitbutter

import json
import sys
sys.stdout.reconfigure(encoding='utf-8')
import io
import os
import traceback
from datetime import datetime
import re
booktitle = os.environ['ESPANSO_BOOKTITLE']
clippings_file = "E:/documents/My Clippings.txt"

bitbutter

import sys
sys.stdout.reconfigure(encoding='utf-8')
from datetime import datetime

# Open the clippings file and read its contents
with open("E:/documents/My Clippings.txt", "r", encoding="utf-8") as f:
    contents = f.read()

# Split the contents into separate notes
notes = contents.split("==========\n")

bitbutter

# append this to [username]/AppData/Roaming/espanso/match
  
  - trigger: kindle_import_book_from_clippings
    replace: "{{output}}"
    vars:
      - name: booktitles
        type: script
        params:
          args:
            - python