Batch Process LLM Text Completions Using a Data Frame

Description

Batch process large language model (LLM) text completions by looping across the rows of a data frame column. The package currently supports OpenAI's GPT, Anthropic's Claude, and Google's Gemini models, with built-in delays for API rate limiting. The package provides advanced text processing features, including automatic logging of batches and metadata to local files, side-by-side comparison of outputs from different LLMs, and integration of a user-friendly Shiny App Addin. Use cases include natural language processing tasks such as sentiment analysis, thematic analysis, classification, labeling or tagging, and language translation.

Usage

batchLLM(
  df,
  df_name = NULL,
  col,
  prompt,
  LLM = "openai",
  model = "gpt-4o-mini",
  temperature = 0.5,
  max_tokens = 500,
  batch_delay = "random",
  batch_size = 10,
  case_convert = NULL,
  sanitize = FALSE,
  attempts = 1,
  log_name = "batchLLM-log",
  hash_algo = "crc32c",
  ...
)

Arguments

Value

Returns the input data frame with an additional column containing the text completion output. The function also writes the output and metadata to the log file after each batch in a nested list format.

Examples

## Not run: 
library(batchLLM)

# Set API keys
Sys.setenv(OPENAI_API_KEY = "your_openai_api_key")
Sys.setenv(ANTHROPIC_API_KEY = "your_anthropic_api_key")
Sys.setenv(GEMINI_API_KEY = "your_gemini_api_key")

# Define LLM configurations
llm_configs <- list(
  list(LLM = "openai", model = "gpt-4o-mini"),
  list(LLM = "anthropic", model = "claude-3-haiku-20240307"),
  list(LLM = "google", model = "1.5-flash")
)

# Apply batchLLM function to each configuration
beliefs <- lapply(llm_configs, function(config) {
  batchLLM(
    df = beliefs,
    col = statement,
    prompt = "classify as a fact or misinformation in one word",
    LLM = config$LLM,
    model = config$model,
    batch_size = 10,
    batch_delay = "1min",
    case_convert = "lower"
  )
})[[length(llm_configs)]]

# Print the updated data frame
print(beliefs)

## End(Not run)

Interact with batchLLM via a Shiny Gadget

Description

This function provides a user interface using Shiny to interact with the batchLLM package. It allows users to configure and execute batch processing through an interactive dashboard.

Usage

batchLLM_shiny()

Value

No return value. Launches a Shiny Gadget that allows users to interact with the batchLLM package.

Beliefs Dataset

Description

The beliefs dataset consists of 20 statements representing opposing views on various scientific, environmental, and societal topics.

Usage

beliefs

Format

A data frame with 20 rows and 1 variable:

statement

A character string with a statement representing a belief.

Examples

head(beliefs)

Interact with Anthropic's Claude API

Description

This function provides an interface to interact with Claude AI models via Anthropic's API, allowing for flexible text generation based on user inputs. This function was adapted from the claudeR repository by yrvelez on GitHub (MIT License).

Usage

claudeR(
  prompt,
  model = "claude-3-5-sonnet-20240620",
  max_tokens = 500,
  stop_sequences = NULL,
  temperature = 0.7,
  top_k = -1,
  top_p = -1,
  api_key = NULL,
  system_prompt = NULL
)

Arguments

Value

The resulting completion up to and excluding the stop sequences.

Examples

## Not run: 
library(batchLLM)

# Set API in the env or use api_key parameter in the claudeR call
Sys.setenv(ANTHROPIC_API_KEY = "your_anthropic_api_key")

# Using Claude-2
response <- claudeR(
  prompt = "What is the capital of France?",
  model = "claude-2.1",
  max_tokens = 50
)
cat(response)

# Using Claude-3
response <- claudeR(
  prompt = list(
    list(role = "user", content = "What is the capital of France?")
  ),
  model = "claude-3-5-sonnet-20240620",
  max_tokens = 50,
  temperature = 0.8
)
cat(response)

# Using a system prompt
response <- claudeR(
  prompt = list(
    list(role = "user", content = "Summarize the history of France in one paragraph.")
  ),
  system_prompt = "You are a concise summarization assistant.",
  max_tokens = 500
)
cat(response)

## End(Not run)

Get Batches

Description

Get batches of generated output in a single data frame from the .rds log file.

Usage

get_batches(df_name = NULL, log_name = "batchLLM-log")

Arguments

Value

A data frame containing the generated output.

Examples

## Not run: 
library(batchLLM)

# Assuming you have a log file with data for "beliefs_40a3012b" (see batchLLM example)
batches <- get_batches("beliefs_40a3012b")
head(batches)

# Using a custom log file name
custom_batches <- get_batches("beliefs_40a3012b", log_name = "custom-log.rds")
head(custom_batches)

## End(Not run)

Scrape Metadata

Description

Scrape metadata from the .rds log file.

Usage

scrape_metadata(df_name = NULL, log_name = "batchLLM-log")

Arguments

Value

A data frame containing metadata.

Examples

library(batchLLM)

# Scrape metadata for all data frames in the default log file
all_metadata <- scrape_metadata()
head(all_metadata)

# Scrape metadata for a specific data frame
specific_metadata <- scrape_metadata("beliefs_40a3012b")
head(specific_metadata)

# Use a custom log file name
custom_metadata <- scrape_metadata(log_name = "custom-log")
head(custom_metadata)