Package {PacketLLM}

Add attachment to active conversation

Description

Add attachment to active conversation

Usage

add_attachment_to_active_conversation(name, content)

Arguments

Value

Logical.

Add a message to the active conversation

Description

Locks the model on the first assistant message. Sets title on the first user message.

Usage

add_message_to_active_history(role, content)

Arguments

Value

Result list (type + extra fields) or error list.

Add user message to the active conversation

Description

Add user message to the active conversation

Usage

add_user_message(text)

Arguments

Value

Invisible NULL.

Available model presets

Description

Returns the model presets shown in the PacketLLM UI.

Usage

available_model_presets()

Value

A data frame with preset labels, model IDs, and descriptions.

List of available AI models for selection in the UI

Description

List of available AI models for selection in the UI

Usage

available_openai_models

Call the configured AI model

Description

Sends messages to the Responses API and returns assistant text.

Usage

call_openai_chat(
  messages,
  model,
  reasoning_effort = "medium",
  verbosity = "low",
  max_output_tokens = NA_integer_
)

Arguments

Value

Character string with assistant reply, or NULL on unexpected response.

Examples

## Not run: 
  msgs <- list(list(role = "user", content = "What does httr do?"))
  call_openai_chat(messages = msgs, model = "gpt-5.4-mini")

## End(Not run)

Capture RStudio editor and project context

Description

Captures lightweight context from the active RStudio editor and project. The function returns a graceful empty context outside RStudio.

Usage

capture_rstudio_context(mode = "auto", max_chars = 6000)

Arguments

Value

A structured list describing the captured context.

Check API Key

Description

Verifies that OPENAI_API_KEY is set. Stops if missing.

Usage

check_api_key()

Value

Invisible TRUE if set.

Examples

## Not run: 
  check_api_key()

## End(Not run)

Create a new conversation

Description

Create a new conversation

Usage

create_new_conversation(
  activate = FALSE,
  add_initial_settings = TRUE,
  title = NULL
)

Arguments

Value

Character: conversation ID.

Delete a conversation

Description

Delete a conversation

Usage

delete_conversation(id)

Arguments

Value

TRUE if deleted, FALSE otherwise.

Get active chat history

Description

Get active chat history

Usage

get_active_chat_history()

Value

List of messages (possibly empty).

Get active conversation object

Description

Get active conversation object

Usage

get_active_conversation()

Value

List or NULL.

Get attachments for active conversation

Description

Get attachments for active conversation

Usage

get_active_conversation_attachments()

Value

List (possibly empty).

Get active conversation ID

Description

Get active conversation ID

Usage

get_active_conversation_id()

Value

Character or NULL.

Get all conversation IDs

Description

Get all conversation IDs

Usage

get_all_conversation_ids()

Value

Character vector.

Get assistant response for the active conversation

Description

Prepares messages and calls the API. Adds the reply to history.

Usage

get_assistant_response()

Value

Character with assistant reply or error message.

Get attachments by ID

Description

Get attachments by ID

Usage

get_conversation_attachments(id)

Arguments

Value

List or NULL.

Get conversation data by ID

Description

Get conversation data by ID

Usage

get_conversation_data(id)

Arguments

Value

List or NULL.

Get conversation history by ID

Description

Get conversation history by ID

Usage

get_conversation_history(id)

Arguments

Value

List or NULL.

Get model for conversation

Description

Get model for conversation

Usage

get_conversation_model(id)

Arguments

Value

Character or NULL.

Get conversation title by ID

Description

Get conversation title by ID

Usage

get_conversation_title(id)

Arguments

Value

Character or NULL.

Initialize the history manager

Description

Initializes state and creates or restores a conversation, then activates it.

Usage

initialize_history_manager(persist = FALSE)

Arguments

Value

Character: ID of the created conversation.

Has the conversation started (model locked)?

Description

Has the conversation started (model locked)?

Usage

is_conversation_started(id)

Arguments

Value

Logical.

Launch PacketLLM from the RStudio Addins menu

Description

Launch PacketLLM from the RStudio Addins menu

Usage

packetllm_addin()

Value

Value passed to run_llm_chat_app().

Parse page range

Description

This function processes a character string specifying a page range (e.g., "1-3,5") and returns a numeric vector containing the individual page numbers, sorted and unique.

Usage

parse_pages(pages_str)

Arguments

Value

A numeric vector containing the unique page numbers specified in the input string, sorted in ascending order. Returns an empty integer vector if the input string is empty or contains only whitespace. Stops with an error if the input pages_str is not a single character string or if the format within the string is invalid (e.g., non-numeric parts, invalid ranges).

Examples

# Example 1: Simple range and single page
page_string1 <- "1-3, 5"
parsed_pages1 <- parse_pages(page_string1)
print(parsed_pages1) # Output: [1] 1 2 3 5

# Example 2: Multiple ranges and single pages, with spaces and duplicates
page_string2 <- " 2, 4-6, 9 , 11-12, 5 "
parsed_pages2 <- parse_pages(page_string2)
print(parsed_pages2) # Output: [1] 2 4 5 6 9 11 12 (sorted, unique)

# Example 3: Single number
page_string3 <- "10"
parsed_pages3 <- parse_pages(page_string3)
print(parsed_pages3) # Output: [1] 10

# Example 4: Empty string input
page_string_empty <- ""
parsed_pages_empty <- parse_pages(page_string_empty)
print(parsed_pages_empty) # Output: integer(0)

# Example 5: Invalid input (non-numeric) - demonstrates error handling
page_string_invalid <- "1-3, five"
## Not run: 
# This will stop with an error message about "five"
tryCatch(parse_pages(page_string_invalid), error = function(e) print(e$message))

## End(Not run)

# Example 6: Invalid range format (missing end) - demonstrates error handling
page_string_invalid_range <- "1-"
## Not run: 
# This will stop with an error message about invalid range format
tryCatch(parse_pages(page_string_invalid_range), error = function(e) print(e$message))

## End(Not run)

# Example 7: Invalid range format (start > end) - demonstrates error handling
page_string_invalid_order <- "5-3"
## Not run: 
# This will stop with an error message about invalid range values
tryCatch(parse_pages(page_string_invalid_order), error = function(e) print(e$message))

## End(Not run)

Read file content

Description

This function reads the content of a file with the extension .R, .pdf, or .docx and returns it as a single character string. TXT files are also supported. For PDF files, if the pages parameter is provided, only the selected pages will be read.

Usage

read_file_content(file_path, pages = NULL)

Arguments

Value

A character string containing the file content, with pages separated by double newlines for PDF files. Stops with an error if the file does not exist, the format is unsupported, or required packages (pdftools for PDF, readtext for DOCX) are not installed or if pages is not numeric when provided.

Examples

# --- Example for reading an R file ---
# Create a temporary R file
temp_r_file <- tempfile(fileext = ".R")
writeLines(c("x <- 1", "print(x + 1)"), temp_r_file)

# Read the content
r_content <- tryCatch(read_file_content(temp_r_file), error = function(e) e$message)
print(r_content)

# Clean up the temporary file
unlink(temp_r_file)

# --- Example for reading a TXT file ---
temp_txt_file <- tempfile(fileext = ".txt")
writeLines(c("Line one.", "Second line."), temp_txt_file)
txt_content <- tryCatch(read_file_content(temp_txt_file), error = function(e) e$message)
print(txt_content)
unlink(temp_txt_file)

# --- Example for PDF (requires pdftools, only run if installed) ---
## Not run: 
# This part requires the 'pdftools' package and a valid PDF file.
# Provide a path to an actual PDF file to test this functionality.
# Replace "path/to/your/sample.pdf" with a real path.

pdf_file_path <- "path/to/your/sample.pdf"

# Check if pdftools is installed and the file exists
if (requireNamespace("pdftools", quietly = TRUE) && file.exists(pdf_file_path)) {

  # Example: Read all pages
  pdf_content_all <- tryCatch(
    read_file_content(pdf_file_path),
    error = function(e) paste("Error reading all pages:", e$message)
  )
  # print(substr(pdf_content_all, 1, 100)) # Print first 100 chars

  # Example: Read only page 1
  pdf_content_page1 <- tryCatch(
    read_file_content(pdf_file_path, pages = 1),
    error = function(e) paste("Error reading page 1:", e$message)
  )
  # print(pdf_content_page1)

} else if (!requireNamespace("pdftools", quietly = TRUE)) {
  message("Skipping PDF example: 'pdftools' package not installed.")
} else {
  message("Skipping PDF example: File not found at '", pdf_file_path, "'")
}

## End(Not run)
# Note: Reading DOCX files is also supported if the 'readtext' package
# is installed, but a simple runnable example is difficult to create
# without including a sample file or complex setup.

Reset the history manager

Description

Reset the history manager

Usage

reset_history_manager(clear_persistent = FALSE)

Arguments

Value

Invisible NULL.

Run the PacketLLM gadget

Description

Launches PacketLLM as an RStudio/Shiny gadget. The gadget can use RStudio editor context when available, while still running outside RStudio with reduced functionality.

Usage

run_llm_chat_app()

Value

Value passed to shiny::stopApp() (typically NULL).

Examples

## Not run: 
run_llm_chat_app()

## End(Not run)

Set the active conversation

Description

Set the active conversation

Usage

set_active_conversation(id)

Arguments

Value

Invisible NULL.

Set model for conversation (if not started)

Description

Set model for conversation (if not started)

Usage

set_conversation_model(id, model_name)

Arguments

Value

Logical.

Set system message for conversation

Description

Set system message for conversation

Usage

set_conversation_system_message(id, message)

Arguments

Value

Logical.