(*---------------------------------------------------------------------------
  Copyright (c) 2025 Anil Madhavapeddy <anil@recoil.org>
  SPDX-License-Identifier: MIT
 ---------------------------------------------------------------------------*)

let detect_language input_text =
  let detector = Langdetect.create_default () in
  let results = Langdetect.detect detector input_text in
  List.iter
    (fun (r : Langdetect.result) -> Printf.printf "%s %.4f\n" r.lang r.prob)
    results

let read_all_stdin () =
  let buf = Buffer.create 4096 in
  try
    while true do
      Buffer.add_channel buf stdin 4096
    done;
    Buffer.contents buf
  with End_of_file -> Buffer.contents buf

let read_file path =
  let ic = open_in path in
  let n = in_channel_length ic in
  let s = really_input_string ic n in
  close_in ic;
  s

let run file_opt =
  let text =
    match file_opt with
    | Some path -> read_file path
    | None -> read_all_stdin ()
  in
  if String.length (String.trim text) = 0 then
    `Error (false, "No input text provided")
  else begin
    detect_language text;
    `Ok ()
  end

open Cmdliner

let file_arg =
  let doc = "Input file to detect language from. If not provided, reads from stdin." in
  Arg.(value & pos 0 (some file) None & info [] ~docv:"FILE" ~doc)

let cmd =
  let doc = "Detect the language of text" in
  let man =
    [
      `S Manpage.s_description;
      `P "Detects the natural language of input text using n-gram frequency analysis.";
      `P "Outputs detected language codes and their probabilities as space-separated values, one per line, sorted by probability (highest first).";
      `S Manpage.s_examples;
      `P "Detect language from a file:";
      `Pre "  langdetect document.txt";
      `P "Detect language from stdin:";
      `Pre "  echo 'Hello world' | langdetect";
    ]
  in
  let info = Cmd.info "langdetect" ~version:"%%VERSION%%" ~doc ~man in
  Cmd.v info Term.(ret (const run $ file_arg))

let () = exit (Cmd.eval cmd)