Upload, and translate a PDF file to model code via external API

Sends a file to a Dify Workflow or LLM API for translation into model code. The file should be pre-validated and optionally pre-combined using combine_uploaded_files before being passed to this function.

Usage

translate_model_code(
  ready_path,
  file_name = NULL,
  service,
  api_key,
  api_upload = NULL,
  api_chat = NULL,
  user_id = "mrgsolve_translator",
  model_gemini = "gemini-3-flash-preview",
  model_openai = "gpt-5-mini",
  model_anthropic = "claude-haiku-4-5-20251001",
  model_openrouter = "arcee-ai/trinity-large-preview:free",
  model_openai_compatible = "gpt-5-mini",
  model_deepseek = "deepseek-reasoner",
  model_apollo = "gpt-5.2",
  model_azure = "gpt-5.2",
  model_aws = "anthropic.claude-sonnet-4-6",
  display_info = TRUE,
  temperature = 0.1,
  seed = 42,
  locally_parse_file = FALSE,
  model_lang = "mrgsolve",
  deep_pdfscan = FALSE,
  force_parse = FALSE,
  mrgsolve_system_prompt,
  mrgsolve_long_user_prompt,
  mrgsolve_short_user_prompt,
  nonmem_system_prompt,
  nonmem_long_user_prompt,
  nonmem_short_user_prompt,
  rxode2_system_prompt,
  rxode2_long_user_prompt,
  rxode2_short_user_prompt,
  internal_version = TRUE,
  debug = TRUE
)

Arguments

ready_path: Path to the file to be processed. For multi-file uploads, this should be the combined output of combine_uploaded_files.
file_name: Original filename (used for display in notifications and for MIME type detection). If NULL, derived from ready_path.
service: Choice of "PROD" (BI-only), "EXP" (BI-only), "Gemini", "OpenAI", "AWS Bedrock", "Claude", "OpenRouter", "OpenAI-Compatible", "DeepSeek", "Apollo" (BI-only), "Azure OpenAI"
api_key: API Key, recommended to store it as env var called "ANTHROPIC_API_KEY" etc
api_upload: API URL for uploading of files (Dify requires a 2-step process)
api_chat: API URL for chat messages, required when using OpenAI-compatible API
user_id: user id for the request (BI-only)
model_gemini: Model to be used when calling Gemini API
model_openai: Model to be used when calling OpenAI API
model_anthropic: Model to be used when calling Anthropic API
model_openrouter: Model to be used when calling OpenRouter
model_openai_compatible: Model to be used when calling OpenAI-compatible API
model_deepseek: Model to be used when calling DeepSeek
model_apollo: Model to be used when calling Apollo (BI only)
model_azure: Model to be used when calling Azure OpenAI
model_aws: Model to be used when calling AWS Bedrock
display_info: Set to TRUE to show how much time/tokens in Shiny UI when job is finished
temperature: Goes from 0 to 1, where 0 is deterministic
seed: seed number for LLMs
locally_parse_file: if TRUE, extracts text locally and includes in prompt instead of uploading file
model_lang: Either "mrgsolve" or "nonmem" (changes the prompt)
deep_pdfscan: Use Vision to extract image data (BI only)
force_parse: Force parsing (BI only)
mrgsolve_system_prompt: String for mrgsolve system prompt
mrgsolve_long_user_prompt: String for mrgsolve long user prompt
mrgsolve_short_user_prompt: String for mrgsolve short user prompt
nonmem_system_prompt: String for nonmem system prompt
nonmem_long_user_prompt: String for nonmem long user prompt
nonmem_short_user_prompt: String for nonmem short user prompt
rxode2_system_prompt: String for rxode2 system prompt
rxode2_long_user_prompt: String for rxode2 long user prompt
rxode2_short_user_prompt: String for rxode2 short user prompt
internal_version: Logical. Only relevant for BI
debug: Displays debug messages

Value

a named list with answer, conversation_id, and chat_obj, or NULL on failure.