nathania

LLM Wiki

A pattern for building personal knowledge bases using LLMs.

This is an idea file, it is designed to be copy pasted to your own LLM Agent (e.g. OpenAI Codex, Claude Code, OpenCode / Pi, or etc.). Its goal is to communicate the high level idea, but your agent will build out the specifics in collaboration with you.

The core idea

Most people's experience with LLMs and documents looks like RAG: you upload a collection of files, the LLM retrieves relevant chunks at query time, and generates an answer. This works, but the LLM is rediscovering knowledge from scratch on every question. There's no accumulation. Ask a subtle question that requires synthesizing five documents, and the LLM has to find and piece together the relevant fragments every time. Nothing is built up. NotebookLM, ChatGPT file uploads, and most RAG systems work this way.

nathania

Install native ARM R: https://cloud.r-project.org/

Install RStudio (probably preview) which is not native as of today: https://www.rstudio.com/products/rstudio/download/preview/

This blog post is helpful for getting Tensorflow working, but it is older, before the ARM version of R and the RStudio version that supports it was available.

This other blog post has specific steps outlined, but some of it has also been superseded.

nathania

# install.packages(c("here", "fs", "stringr", "purrr", "git2r"))

# to add invisibly in your R profile, open with usethis::edit_r_profile()
# then define it in an environment, e.g.
# .env <- new.env()
# .env$move_slides_to_web <- {function definition}

move_slides_to_web <- function(folder = NULL, index = NULL) {
  
  if (is.null(folder)) {

nathania

# https://community.rstudio.com/t/how-to-apply-a-plotting-function-to-a-list-of-dataframes-and-a-character-vector/26033

library(tidyverse)
library(scales)
library(naniar)


# original viz ------------------------------------------------------------

cols <- c('Good' = '#1a9641', 

nathania

ruler <- function(width = getOption("width")) {
  x <- seq_len(width)
  y <- dplyr::case_when(
    x %% 10 == 0 ~ as.character((x %/% 10) %% 10),
    x %% 5 == 0  ~ "+",
    TRUE         ~ "-"
  )
  cat(y, "\n", sep = "")
  cat(x %% 10, "\n", sep = "")
}

nathania

# `expand(df, nesting(patient, PCP), event_type)`
#   generates all unique combinations in `(patient, PCP) x event_type`.
#   That is; there will be `n_distinct(df$event_type)` rows per `patient, PCP`
#   tuple in the data
#  
#  df <- tibble(x = 1:4, y = letters[1:4]) 
# testthat::expect_equivalent(expand(df, nesting(x, y)), distinct(df, x, y)) 
# testthat::expect_equivalent(expand(df, crossing(x, y)), expand(df, x, y))

# generate all combinations of elements in a set S

nathania

plot(1:10)

nathania

# Adapted from https://stackoverflow.com/a/7267364/1036500 by Andrie de Vries

library(ggplot2)

td <- expand.grid(
  hjust=c(0, 0.5, 1),
  vjust=c(0, 0.5, 1),
  angle=c(0, 45, 90),
  text="text"
)

nathania

#!/usr/bin/env python
# numpy and biopython are required -- pip install numpy biopython

from Bio import Entrez
from Bio import Medline

MAX_COUNT = 10
TERM = 'Tuberculosis'

print('Getting {0} publications containing {1}...'.format(MAX_COUNT, TERM))

nathania

<link rel="stylesheet" href="https://maxcdn.bootstrapcdn.com/font-awesome/4.4.0/css/font-awesome.min.css">
<div class="rating">
  <span><i class="fa fa-male"></i></span>
  <span><i class="fa fa-male"></i></span>
  <span><i class="fa fa-male"></i></span>
  <span><i class="fa fa-male"></i></span>
  <span><i class="fa fa-male"></i></span>
</div>