index.qmd

---
title: ""
format: 
  revealjs:
    theme: [default, custom.scss]
    footer: '<span style="color:#aa5b31;">github.com/simonpcouch/openscapes-25</span>'
editor: source
knitr:
  opts_chunk: 
    echo: true
    collapse: true
    comment: "#>"
---

## Writing R code with the help of LLMs

<a href="github.com/simonpcouch/openscapes-25"><img src="figures/wall.png" align="right" height="350" alt="Hex wall" /></a>


<br><br><br><br><br><br><br>
_Simon Couch_ - Posit, PBC

Open Source Group, R / LLMs


```{r}
#| label: load-pkgs
#| message: false
#| warning: false
#| eval: false
#| include: false
remotes::install_github("luisdva/hexsession")
library(hexsession)
library(acquaint)
library(btw)
library(chores)
library(ellmer)
library(gander)
library(vitals)
library(hexsession)
```

## 

```{r}
#| label: productivity-plot
#| echo: false
#| fig.width: 10
#| fig.height: 6
library(ggplot2)
library(dplyr)

set.seed(123)
data <- tibble(
  gain = c(rnorm(100, mean = 1.4, sd = 0.3), rnorm(100, mean = 10, sd = 1.5)),
  type = rep(
    c("Complex, production software", "Fun little throwaway programs"), 
    each = 100
  )
)

ggplot(data, aes(x = gain, fill = type)) +
  geom_density() +
  scale_fill_manual(values = c("#aa5b31", "#3b7c98")) +
  labs(
    title = "Productivity gains from LLMs are bimodal",
    x = "Productivity gain",
    y = "Density",
    fill = "Type"
  ) +
  theme_minimal() +
  theme(
    legend.position = "bottom",
    text = element_text(size = 16)
  ) +
  scale_x_continuous(breaks = c(1, 5, 10), labels = c("1x", "5x", "10x"))
```

. . .

Today, we're talking about the LHS⬅️

# Part 1: Have a chat

## Part 1: Have a chat <a href="https://ellmer.tidyverse.org"><img src="figures/hexes/ellmer.png" align="right" height="400" alt="ellmer website" /></a>

Meet ellmer!🐘 

<br>

. . .

::: columns
::: {.column width="50%"}
```{r}
#| label: install-ellmer
#| eval: false
install.packages("ellmer")
```
:::
::: {.column width="50%"}
:::
:::

## Part 1: Have a chat

```{r}
#| label: load-ellmer
#| include: false
library(ellmer)
```

These are the same:

::: columns
::: {.column width="50%"}
![](figures/chats/hello.png)
:::

::: {.column width="50%"}
<br>
```{r}
#| label: chat-example
#| eval: false
library(ellmer)

ch <- chat_github(
  model = "gpt-4o"
)

ch$chat("hey!")
#> "Hey there!😊 What can I 
#>  help you with today?"
```

:::
:::

## Part 1: Have a chat

Your turn! Create a chat object and say "hey!"

* `chat_github()` might "just work"
* If not, set up `chat_anthropic()` using instructions linked below

```{r 1-have-a-chat}
#| echo: false
countdown::countdown(minutes = 3, id = "have-a-chat")
```

<!--
Interactively:
* Chat a couple times
* Extract the turns information
-->

# Part 2: The system prompt

## Part 2: The system prompt

* An "invisible message" at the start of your chat
* Use it to influence behavior, give knowledge, define output format, etc

. . .

```{r}
#| label: system-prompt-example
#| eval: false
ch <- chat_anthropic(
  system_prompt = 
    "Try and tie any response back to the Openscapes organization."
)

ch$chat("What's 2+2?")
```

## Part 2: The system prompt

Your turn: adjust the system prompt to the model so that, when supplied a question like "What's 2+2?", the model returns _only_ the answer as a _word_ (rather than a digit), no punctuation or exposition.

<br>

```{r}
#| label: system-prompt-exercise
#| eval: false
ch$chat("What's 2+2?")
#> four
```

```{r 2-system-prompt}
#| echo: false
countdown::countdown(minutes = 3, id = "system-prompt")
```

## Part 2: The system prompt

You can get a [lot of mileage](https://github.com/simonpcouch/chores/blob/main/inst/prompts/roxygen-prefix.md) out of the system prompt:

```{r}
#| label: complex-system-prompt
#| eval: false
ch <- chat_anthropic(
  system_prompt = paste0(readLines(
    "https://raw.githubusercontent.com/simonpcouch/chores/refs/heads/main/inst/prompts/roxygen-prefix.md"
  ), collapse = "\n")
)
```

## Part 2: The system prompt
What if that wasn't super difficult to interface with? 

{{< video figures/chore-roxygen.mov >}}

## Part 2: The system prompt <a href="https://simonpcouch.github.io/chores/"><img src="figures/hexes/chores.png" align="right" height="200" alt="chores website" /></a>

<br>
```r
install.packages("chores")
```

* Attach system prompts to a key command in your IDE
* Select some code, press the command, and watch code stream in
* Supports common R package development actions by default

# Intermission: tokens

## Intermission: tokens

OpenAI and Anthropic have two main ways they make money from their models:

1) Subscription plans (like chatgpt.com)
2) API usage (like from ellmer)

## Intermission: tokens

API usage is "pay-as-you-go" by _token_:

- Words, parts of words, or individual characters
  - "hello" → 1 token
  - "unconventional" → 3 tokens: `un|con|ventional`

## Intermission: tokens

Here's the pricing _per million tokens_ for some common models:

<br>

```{r}
#| label: pricing-table
#| echo: false
pricing <- 
  tibble::tribble(
    ~Name, ~Input, ~Output,
    "GPT 4o", "$3.75", "$15.00",
    "GPT 4o-mini", "$0.15", "$0.60",
    "Claude 4 Sonnet", "$3.00", "$15.00"
  )

knitr::kable(pricing)
```

## Intermission: tokens

To put that into context, the source code for these slides so far is 650 tokens.

If I input them to GPT 4o:

$$
650 \text{ tokens} \times \frac{\$3.75 }{1,000,000~\text{tokens}} = \$0.00244
$$

# Part 3: Hallucination and code assistance

## Part 3: Hallucination and code assistance

```{r}
#| label: stackoverflow-data
#| eval: false
library(ggplot2)
library(modeldata)
stackoverflow
#> # A tibble: 5,594 × 21
#>    Country        Salary YearsCodedJob OpenSource Hobby CompanySizeNumber Remote
#>    <fct>           <dbl>         <int>      <dbl> <dbl>             <dbl> <fct> 
#>  1 United Kingdom 1   e5            20          0     1              5000 Remote
#>  2 United States  1.3 e5            20          1     1              1000 Remote
#>  3 United States  1.75e5            16          0     1             10000 Not r…
#> # ℹ 5,590 more rows
#> # ℹ 14 more variables: CareerSatisfaction <int>, Data_scientist <dbl>, …
```

If I type “plot salary vs experience”, what information does the model need access to complete that request?

## Part 3: Hallucination and code assistance

If I type “plot salary vs experience”, what information does the model need access to?

:::incremental
* The name of the relevant data frame
* My preferred plotting library
* The names of the relevant columns
:::

. . . 

The first two can be inferred from the source code, but the third requires access to your R session.

## Part 3: Hallucination and code assistance  <a href="https://simonpcouch.github.io/gander/"><img src="figures/hexes/gander.png" align="right" height="200" alt="gander website" /></a>

<br>
```r
install.packages("gander")
```

* An R coding assistant written in R
* Automatically incorporates relevant context from your R sessions


# Appendix A: Providers

## Appendix A: Providers

A "provider" is a service that hosts models on an API.

In ellmer, each provider has its own `chat_*()` function, like `chat_github()` or `chat_anthropic()`

. . .

* `chat_github()` serves some popular models, like GPT-4o, for "free"
    - "free" in the sense of "we're going to use all of the data you send us"
    - heavily rate-limited; you'll need to pay for even modest usage

## Appendix A: Providers

* `chat_openai()` 
    - traditionally, more consumer-focused
    - weaker privacy guarantees 

## Appendix A: Providers

* `chat_anthropic()` serves Claude Sonnet
    - traditionally more developer/enterprised-focused
    - stronger privacy guarantees
    - subsidizes credits via Claude for Education

## Appendix A: Providers

You can be your own "provider", too:

* `chat_ollama()` uses a model that runs on your laptop
    - much less powerful than the Big Ones
    - "free" in the usual sense

## Appendix A: Providers

Many organizations have private deployments of models set up for internal, secure use. ellmer supports the common ones.

Ask around to see if this is the case at NOAA/NASA!

## Learn more

<br>

<span style="font-size:130%">

<center>[github.com/simonpcouch/openscapes-25]{style="color:#c46938;"}</center>

</span>

<br>

<a href="github.com/simonpcouch/openscapes-25"><img src="figures/wall.png" align="right" height="350" alt="Hex wall" /></a>