Skip to contents

Process text chunks through Anthropic's Messages API with batch file output

Source: R/anthropic_messages.R
ant_complete_chunks.Rd

Processes large volumes of text through Anthropic's Messages API in configurable chunks, writing results progressively to parquet files. Handles concurrent requests, automatic retries, and structured outputs.

Usage

ant_complete_chunks(
  texts,
  ids,
  chunk_size = 5000L,
  model = .ANT_DEFAULT_MODEL,
  system_prompt = NULL,
  output_dir = "auto",
  overwrite = FALSE,
  schema = NULL,
  concurrent_requests = 5L,
  temperature = 0,
  max_tokens = 4096L,
  max_retries = 5L,
  timeout = 120L,
  key_name = "ANTHROPIC_API_KEY",
  endpoint_url = .ANT_MESSAGES_ENDPOINT,
  id_col_name = "id",
  effort = NULL,
  thinking = NULL
)

Arguments

texts

Character vector of texts to process

ids

Vector of unique identifiers (same length as texts)

chunk_size

Number of texts per chunk before writing to disk

model

Anthropic model to use (default: "claude-haiku-5-5")

system_prompt

Optional system prompt (applied to all requests). Prompt caching is enabled automatically, reducing costs when the same system prompt is shared across many requests.

output_dir

Directory for parquet chunks ("auto" generates timestamped dir)

overwrite

If FALSE (default), errors when output_dir already contains chunk (.parquet) or metadata.json files. Set to TRUE to delete them and write fresh outputs; other files are left untouched.

schema

Optional JSON schema for structured output

concurrent_requests

Number of concurrent requests

temperature

Sampling temperature (0-1), included in the request only when non-NULL. Dropped with a warning on models that reject sampling parameters (Claude Opus 4.7+, Haiku 5.5, Sonnet 5, Fable 5), and when thinking is "adaptive" or a budget.

max_tokens

Maximum tokens in the response, including thinking tokens

max_retries

Maximum retry attempts per request

timeout

Request timeout in seconds

key_name

Environment variable name for API key

endpoint_url

Anthropic API endpoint URL

id_col_name

Name for ID column in output

effort

Optional effort level, one of "low", "medium", "high", "xhigh", "max". Sent as output_config$effort. Not accepted by Haiku 4.5 or Sonnet 4.5. See the "Effort and thinking" section.

thinking

Optional thinking setting. NULL (the default) leaves it out of the request, so the model uses its own default. One of "adaptive", "disabled" or "between_tools", or a whole number for a fixed thinking budget in tokens. See the "Effort and thinking" section.

Value

A tibble with columns: the id column, content, .error, .error_msg, stop_reason, input_tokens, output_tokens, thinking_tokens, .status and .chunk

Details

This function is designed for processing large text datasets. It divides input into chunks, processes each chunk with concurrent API requests, and writes results to disk to minimise memory usage and possibility of data loss.

Results are written as parquet files in the specified output directory, along with a metadata.json file containing processing parameters.

When using a custom output_dir, existing chunk files are protected by default. Set overwrite = TRUE to replace them.

The results have the same stop_reason, input_tokens, output_tokens and thinking_tokens columns as ant_batch_results(). thinking_tokens is NA when the API doesn't report it.

Effort and thinking

Most users only need effort. It controls how much the model thinks and how many tokens it uses overall. Lower effort is cheaper and faster. The default is "high" on most models, and "medium" on Claude Opus 5.5 and Claude Haiku 5.5. Haiku 4.5 and Sonnet 4.5 don't accept effort.

Use thinking for two things: to turn thinking on for older models, and to turn it off on the models that allow it.

  • Haiku 4.5, Sonnet 4.5 and Opus 4.5 only think with a fixed budget, e.g. thinking = 2048. The budget must be at least 1024 and less than max_tokens.

  • Models from 4.6 onwards use thinking = "adaptive", and models from 4.7 onwards reject a fixed budget.

  • thinking = "disabled" turns thinking off on Opus 4.7, Opus 4.8, Sonnet 5, Opus 5 and Haiku 5.5. Opus 5 and Haiku 5.5 only accept it at effort "high" or lower.

  • Sonnet 5.5 uses thinking = "between_tools" instead, at effort "high" or lower.

  • Opus 5.5, Fable and Mythos can't turn thinking off. Lower effort instead.

EndpointR checks these rules before sending, and stops with a message that suggests a setting that works. Models it doesn't recognise are sent unchecked.

Thinking tokens count towards max_tokens. If a model spends all of max_tokens thinking, the response has no text, content is NA, and stop_reason is "max_tokens". Raise max_tokens, or lower effort or the thinking budget, if that happens. A thinking budget is a target rather than a hard limit. In our tests, Haiku 4.5 with thinking = 2048 sometimes thought for over 4000 tokens, so leave plenty of room above the budget.

Anthropic reject a changed temperature while thinking is on, so when thinking is "adaptive" or a budget, temperature is dropped with a warning.

Before you submit a large batch, send one request with ant_complete_text() using the same settings. The API only reports a bad setting in a batch after the whole batch has run.