The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.
check_joint() — LLM-based joint validation was
unreliable (circular LLM self-validation) and its parsing logic was
fragile.magrittr dependency. The %>%
re-export is no longer provided. Use the native R pipe
|> (available since R 4.1.0).cat() to message() for user-facing
output in set_llm() and chat_llm(), complying
with CRAN best practices.cat() preview from
parse_joint().joint_prompt() @returns tag
(incorrectly stated data.frame; now correctly states
character string).@returns tag for
set_llm().thinking mode support, the author does not have
the energy to set up individual thinking adaptation interfaces for the
countless LLMs.build_joint() now takes a raw LLM response string
instead of data.frames. Signature changed from
build_joint(x, y, key1, key2, ...) to
build_joint(llm_response, key1, key2). The LLM call and
prompt construction are now orchestrated by
llm_join().llm_join() explicitly orchestrates the full pipeline:
joint_prompt() → chat_llm() →
build_joint().~/.LLMJOIN.yml to
tools::R_user_dir("llmjoin", "config"), complying with CRAN
policy on writing to the user home directory.@examples to set_llm() and
check_joint().LazyData: false in DESCRIPTION to avoid CRAN NOTE
about missing data directory.%>% re-export,
validate_llm_config(), and
test_llm_service_minimal().chat_llm() enables max reasoning intensity by default.
Claude: thinking: {type: enabled, budget_tokens: 16000}.
OpenAI / Gemini: reasoning_effort: "high".R/providers.R) with pluggable
.providers registry supporting OpenAI, Claude (Anthropic),
and Gemini. Four internal helpers handle per-provider auth, request
body, response parsing, and URL construction.validate_llm_config() writes
VERIFIED: true into the YAML after a successful probe,
skipping redundant network checks.test_llm_service_minimal() for one-token connectivity
probing.%||% null-coalescing operator.magrittr from Imports.chat_llm() input validation: missing
.message now properly errors.set_llm().provider_headers()).R/network-utils.R →
R/connection.R, R/main.R →
R/llmjoin.R.R/utils.R with pipe re-export, %||%,
httr/jsonlite imports.testthat
(>= 3.0.0) with local_mocked_bindings mock patterns for
offline testing.check_joint() for LLM-based joint validation. Asks the
LLM to identify problematic mappings and filters them from the
result.joint_prompt() for generating structured matching
prompts from two key columns.tbl2md() for converting data.frames to markdown
tables.build_joint() for constructing fuzzy-join mapping
tables via LLM.llm_join() for end-to-end fuzzy join: build_joint +
optional check_joint + merge.set_llm() for configuring LLM service credentials
(provider, URL, key, model) stored in ~/.LLMJOIN.yml.chat_llm() as the single entry point for LLM API calls,
supporting OpenAI, Claude, and Gemini providers.These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.