Package index • dragonfarm

Setup

Drop in documents, ask them questions, write and review training questions, build the training set.

dragon_documents(): Read documents into a document set
dragon_documents_load(): Reopen a saved document set
dragon_documents_list(): List saved document sets
dragon_example_documents(): Example documents
dragon_chunks(): The chunks of a document set
dragon_index(): Index a document set for retrieval
dragon_retrieve(): Find the chunks most relevant to a question
dragon_ask(): Ask a question of your documents
dragon_ellmer_tools(): dragonfarm's tools for an ellmer chat or an MCP client
dragon_mcp_server(): Serve dragonfarm's tools to MCP clients
dragon_questions() dragon_questions_progress() dragon_questions_wait(): Write questions and answers from a document set
dragon_questions_load() dragon_questions_save() dragon_questions_list(): Load, save, and list question sets
dragon_questions_dataset(): Training data from reviewed questions

dragon_dataset(): Create a dataset for fine-tuning
dragon_map(): Map dataset columns to prompt, response, and system text
dragon_map_pairs(): Map columns for preference optimization
dragon_map_prompts(): Map columns for reinforcement learning
dragon_conversations(): Multi-turn conversations as training data
dragon_synthesize(): Write fine-tuning data with a teacher model
dragon_synthesize_pairs(): Build preference pairs from a model's own samples
dragon_prompts(): Prompts from a run's data files
dragon_preview(): Preview mapped rows as chat turns
dragon_split(): Hold out rows for evaluation
dragon_example_data(): Path to the bundled example dataset

dragon_run(): Reopen an existing run
dragon_runs(): List runs
dragon_archived_runs(): Archived runs
dragon_archive_run() dragon_unarchive_run() dragon_delete_run(): Archive, restore, or delete a run
dragon_runs_dir(): Directory where runs are stored
dragon_status() dragon_progress() dragon_logs(): Inspect a run
dragon_code(): R code that reproduces a run

dragon_evaluate(): Evaluate a finished run
dragon_metrics() dragon_metric_regex(): Task metrics for generated replies
dragon_judge(): Judge a run's replies with a language model
dragon_llm_anthropic() dragon_judge_anthropic() dragon_llm_ellmer() dragon_judge_ellmer(): Language models as functions: the Claude API and ellmer
dragon_compare(): Compare runs side by side

dragon_backend_local() dragon_backend_server() dragon_backend_ollama() dragon_backend(): Inference backends
dragon_serve_ollama(): Serve a run's model with Ollama
dragon_worker_stop(): Stop the local inference worker
dragon_chat() dragon_chat_load(): Talk to a model, with memory of the conversation
dragon_feedback(): Training data from chat feedback

dragon_pipeline(): Run several post-training stages as one pipeline
dragon_step_train() dragon_step_prefer() dragon_step_synthesize_pairs() dragon_step_reinforce() dragon_step_judge() dragon_step_evaluate() dragon_step_merge() dragon_step_publish(): Steps of a post-training pipeline
dragon_pipeline_status(): Progress of a pipeline
dragon_pipeline_cancel(): Cancel a pipeline