microprints-phoenix/lib/microprint.ex

defmodule Microprints.Microprint do
  @moduledoc """
  Pure functions for generating microprint visualizations from source files.

  A microprint is a compact visual representation where each line of code
  becomes a colored stripe based on syntax elements.
  """

  @type segment :: %{start: non_neg_integer(), length: non_neg_integer(), color: String.t()}
  @type line_info :: %{
          color: String.t(),
          indent: non_neg_integer(),
          length: non_neg_integer(),
          segments: [segment()] | nil
        }
  @type microprint :: %{lines: [line_info()], line_count: non_neg_integer()}

  # Color palette for different syntax elements
  @colors %{
    # purple - def, defmodule, if, case, etc.
    keyword: "#8B5CF6",
    # green - strings and charlists
    string: "#10B981",
    # gray - comments
    comment: "#6B7280",
    # red - function definitions
    function_def: "#EF4444",
    # amber - atoms
    atom: "#F59E0B",
    # blue - numbers
    number: "#3B82F6",
    # pink - module names
    module: "#EC4899",
    # purple - operators
    operator: "#8B5CF6",
    # light gray - default
    default: "#D1D5DB"
  }

  @elixir_keywords ~w(def defp defmodule defmacro defmacrop defguard defguardp
    defstruct defexception defprotocol defimpl defdelegate defoverridable
    if unless case cond with for receive try catch rescue after raise throw
    import require alias use quote unquote fn do end when and or not in)

  @doc """
  Generates a microprint for the given file path.

  Returns `{:ok, microprint}` or `{:error, reason}`.
  """
  @spec generate(String.t()) :: {:ok, microprint()} | {:error, term()}
  def generate(path) do
    case File.read(path) do
      {:ok, content} ->
        lines = String.split(content, "\n")
        line_infos = analyze_lines(path, lines)
        {:ok, %{lines: line_infos, line_count: length(line_infos)}}

      {:error, reason} ->
        {:error, reason}
    end
  end

  @doc """
  Returns the color for a given syntax type.
  """
  @spec color_for(atom()) :: String.t()
  def color_for(type), do: Map.get(@colors, type, @colors.default)

  @doc """
  Returns the color legend as a list of {label, color} tuples.
  """
  @spec color_legend() :: [{String.t(), String.t()}]
  def color_legend do
    [
      {"Function", @colors.function_def},
      {"Module", @colors.module},
      {"Keyword", @colors.keyword},
      {"String", @colors.string},
      {"Comment", @colors.comment},
      {"Atom", @colors.atom},
      {"Number", @colors.number},
      {"Other", @colors.default}
    ]
  end

  # Private functions

  defp analyze_lines(path, lines) do
    if elixir_file?(path) do
      analyze_elixir_lines(lines)
    else
      analyze_generic_lines(lines)
    end
  end

  defp analyze_elixir_lines(lines) do
    lines
    |> Enum.with_index(1)
    |> Enum.map(fn {line, line_num} ->
      segments = tokenize_elixir_line(line, line_num)

      %{
        color: colorize_elixir_line(line),
        indent: count_indent(line),
        length: content_length(line),
        segments: segments
      }
    end)
  end

  defp analyze_generic_lines(lines) do
    Enum.map(lines, fn line ->
      %{
        color: colorize_generic_line(line),
        indent: count_indent(line),
        length: content_length(line),
        segments: nil
      }
    end)
  end

  defp content_length(line) do
    # Length of line after trimming leading whitespace
    line |> String.trim_leading() |> String.length()
  end

  defp count_indent(line) do
    # Count leading spaces (treat tab as 2 spaces)
    line
    |> String.graphemes()
    |> Enum.take_while(&(&1 in [" ", "\t"]))
    |> Enum.reduce(0, fn char, acc ->
      case char do
        " " -> acc + 1
        "\t" -> acc + 2
      end
    end)
  end

  defp elixir_file?(path) do
    ext = Path.extname(path)
    ext in [".ex", ".exs"]
  end

  defp tokenize_elixir_line(line, line_num) do
    # Handle empty or whitespace-only lines
    trimmed = String.trim(line)

    if trimmed == "" do
      nil
    else
      case :elixir_tokenizer.tokenize(String.to_charlist(line), line_num, []) do
        {:ok, _end_line, _end_col, _warnings, tokens, _terminators} ->
          build_segments(line, Enum.reverse(tokens))

        {:error, _, _, _, _} ->
          # Tokenization failed, fall back to nil (use single color)
          nil
      end
    end
  end

  defp build_segments(line, tokens) do
    line_length = String.length(line)

    # Convert tokens to segments with positions
    segments =
      tokens
      |> Enum.map(&token_to_segment/1)
      |> Enum.reject(&is_nil/1)

    # Fill gaps with default color
    fill_gaps(segments, line_length)
  end

  defp token_to_segment({type, {_line, col, token_chars}, _value}) when is_list(token_chars) do
    # col is 1-indexed, convert to 0-indexed
    start = col - 1
    length = length(token_chars)
    color = token_type_to_color(type)
    %{start: start, length: length, color: color}
  end

  defp token_to_segment({type, {_line, col, _}, value}) when is_atom(value) do
    start = col - 1
    length = value |> Atom.to_string() |> String.length()
    color = token_type_to_color(type)
    %{start: start, length: length, color: color}
  end

  defp token_to_segment({type, {_line, col, nil}, _value}) do
    # Single-character tokens like (, ), etc.
    start = col - 1
    color = token_type_to_color(type)
    %{start: start, length: 1, color: color}
  end

  defp token_to_segment({type, {_line, col, length}, _value}) when is_integer(length) do
    start = col - 1
    color = token_type_to_color(type)
    %{start: start, length: max(1, length), color: color}
  end

  defp token_to_segment(_), do: nil

  defp token_type_to_color(type) do
    case type do
      # Keywords
      t when t in [:do, :end, :fn, :when, :and, :or, :not, :in] ->
        @colors.keyword

      # Identifiers that are keywords
      :identifier ->
        @colors.default

      :kw_identifier ->
        @colors.keyword

      :paren_identifier ->
        @colors.function_def

      :do_identifier ->
        @colors.function_def

      # Atoms
      :atom ->
        @colors.atom

      :kw_identifier_unsafe ->
        @colors.atom

      # Strings
      :bin_string ->
        @colors.string

      :list_string ->
        @colors.string

      :bin_heredoc ->
        @colors.string

      :list_heredoc ->
        @colors.string

      :interpolation_start ->
        @colors.string

      :interpolation_end ->
        @colors.string

      # Numbers
      :int ->
        @colors.number

      :flt ->
        @colors.number

      # Operators
      t
      when t in [
             :dual_op,
             :mult_op,
             :power_op,
             :concat_op,
             :range_op,
             :xor_op,
             :pipe_op,
             :match_op,
             :assoc_op,
             :arrow_op,
             :rel_op,
             :comp_op,
             :and_op,
             :or_op,
             :not_op,
             :capture_op,
             :ternary_op,
             :at_op,
             :unary_op,
             :type_op,
             :stab_op
           ] ->
        @colors.operator

      # Module aliases
      :alias ->
        @colors.module

      # Comments are handled separately (not tokenized)
      # Punctuation - use default
      _ ->
        @colors.default
    end
  end

  defp fill_gaps(segments, line_length) do
    # Sort by start position
    sorted = Enum.sort_by(segments, & &1.start)

    # Build list with gaps filled
    {filled, last_end} =
      Enum.reduce(sorted, {[], 0}, fn segment, {acc, current_pos} ->
        # Add gap segment if needed
        acc =
          if segment.start > current_pos do
            gap = %{
              start: current_pos,
              length: segment.start - current_pos,
              color: @colors.default
            }

            [gap | acc]
          else
            acc
          end

        segment_end = segment.start + segment.length
        {[segment | acc], max(current_pos, segment_end)}
      end)

    # Add trailing gap if needed
    filled =
      if last_end < line_length do
        gap = %{start: last_end, length: line_length - last_end, color: @colors.default}
        [gap | filled]
      else
        filled
      end

    Enum.reverse(filled)
  end

  defp colorize_elixir_line(line) do
    trimmed = String.trim(line)

    cond do
      # Empty line
      trimmed == "" ->
        @colors.default

      # Comment line
      String.starts_with?(trimmed, "#") ->
        @colors.comment

      # Function definition
      Regex.match?(~r/^\s*(def|defp|defmacro|defmacrop|defguard|defguardp)\s+\w+/, line) ->
        @colors.function_def

      # Module definition
      Regex.match?(~r/^\s*defmodule\s+/, line) ->
        @colors.module

      # String line (starts with quote or contains prominent string)
      Regex.match?(~r/^\s*["']/, line) || Regex.match?(~r/~[a-zA-Z]\[|~[a-zA-Z]{/, line) ->
        @colors.string

      # Doc attribute
      Regex.match?(~r/^\s*@(moduledoc|doc|spec|type|typep|typedoc)/, line) ->
        @colors.comment

      # Keyword-heavy line
      keyword_line?(trimmed) ->
        @colors.keyword

      # Atom-heavy line
      Regex.match?(~r/^\s*:[\w_]+/, line) || Regex.match?(~r/@\w+/, line) ->
        @colors.atom

      # Default
      true ->
        dominant_color(line)
    end
  end

  defp keyword_line?(trimmed) do
    first_word = trimmed |> String.split(~r/[\s(]/, parts: 2) |> List.first() || ""
    first_word in @elixir_keywords
  end

  defp dominant_color(line) do
    cond do
      # Has numbers
      Regex.match?(~r/\b\d+\.?\d*\b/, line) && !Regex.match?(~r/[a-zA-Z]/, line) ->
        @colors.number

      # Has operators
      Regex.match?(~r/[|><+\-*\/=]+/, line) ->
        @colors.operator

      # Has function calls (word followed by parenthesis)
      Regex.match?(~r/\w+\(/, line) ->
        @colors.default

      true ->
        @colors.default
    end
  end

  defp colorize_generic_line(line) do
    trimmed = String.trim(line)

    cond do
      trimmed == "" ->
        @colors.default

      # Common comment patterns
      String.starts_with?(trimmed, "//") ||
        String.starts_with?(trimmed, "#") ||
        String.starts_with?(trimmed, "/*") ||
          String.starts_with?(trimmed, "*") ->
        @colors.comment

      # String-heavy line
      Regex.match?(~r/["'`]/, line) ->
        @colors.string

      # Keywords for various languages
      Regex.match?(
        ~r/\b(function|const|let|var|class|import|export|return|if|else|for|while)\b/,
        line
      ) ->
        @colors.keyword

      true ->
        @colors.default
    end
  end
end