notex.nvim/lua/notex/parser/yaml.lua

-- YAML header parsing module
local M = {}

local lyaml = require('lyaml')
local utils = require('notex.utils')

-- Extract YAML header from markdown content
function M.extract_yaml_header(content)
  if not content or content == "" then
    return nil, "Empty content provided"
  end

  -- Check for YAML header delimiters
  local start_pos = content:find("^%s*%-%-%-%s*\n")
  if not start_pos then
    return nil, "No YAML header found"
  end

  local end_pos = content:find("\n%s*%-%-%-%s*\n", start_pos + 4)
  if not end_pos then
    return nil, "Unclosed YAML header"
  end

  -- Extract YAML content
  local yaml_content = content:sub(start_pos + 4, end_pos - 1)

  return yaml_content, nil
end

-- Parse YAML header content
function M.parse_yaml(yaml_content)
  if not yaml_content or yaml_content == "" then
    return {}, nil
  end

  local ok, data = pcall(lyaml.load, yaml_content)

  if not ok then
    return nil, "YAML parsing failed: " .. tostring(data)
  end

  if type(data) ~= "table" then
    return {}, nil
  end

  return data, nil
end

-- Parse markdown file and extract YAML header
function M.parse_markdown_file(file_path)
  -- Validate file exists
  if not utils.file_exists(file_path) then
    return nil, "File not found: " .. file_path
  end

  -- Validate UTF-8 encoding
  if not utils.is_utf8(file_path) then
    return nil, "File is not valid UTF-8: " .. file_path
  end

  -- Read file content
  local content, err = utils.read_file(file_path)
  if not content then
    return nil, err
  end

  -- Extract YAML header
  local yaml_content, extract_err = M.extract_yaml_header(content)
  if not yaml_content then
    return nil, extract_err
  end

  -- Parse YAML
  local yaml_data, parse_err = M.parse_yaml(yaml_content)
  if not yaml_data then
    return nil, parse_err
  end

  return yaml_data, nil
end

-- Flatten YAML data into key-value pairs
function M.flatten_yaml(data, prefix)
  local flattened = {}
  prefix = prefix or ""

  for key, value in pairs(data) do
    local full_key = prefix .. (prefix ~= "" and "." or "") .. key

    if type(value) == "table" then
      -- Recursively flatten nested tables
      local nested = M.flatten_yaml(value, full_key)
      for nested_key, nested_value in pairs(nested) do
        flattened[nested_key] = nested_value
      end
    else
      flattened[full_key] = value
    end
  end

  return flattened
end

-- Validate YAML structure
function M.validate_yaml(yaml_data)
  local errors = {}

  if type(yaml_data) ~= "table" then
    table.insert(errors, "YAML data must be a table")
    return errors
  end

  -- Check for required fields (if any)
  local required_fields = {} -- Add required fields as needed

  for _, field in ipairs(required_fields) do
    if yaml_data[field] == nil then
      table.insert(errors, string.format("Required field '%s' is missing", field))
    end
  end

  -- Validate field types
  local field_types = {
    -- Define expected types for specific fields
  }

  for field, expected_type in pairs(field_types) do
    if yaml_data[field] ~= nil and type(yaml_data[field]) ~= expected_type then
      table.insert(errors, string.format("Field '%s' should be %s, got %s",
                                         field, expected_type, type(yaml_data[field])))
    end
  end

  return errors
end

-- Detect and convert property types
function M.detect_property_type(value)
  local value_type = type(value)

  if value_type == "boolean" then
    return "boolean", value
  elseif value_type == "number" then
    return "number", value
  elseif value_type == "string" then
    -- Check for ISO 8601 date format
    if value:match("^%d%d%d%d%-%d%d%-%d%d$") or
       value:match("^%d%d%d%d%-%d%d%-%d%dT%d%d:%d%d:%d%dZ?$") then
      return "date", value
    end

    -- Check for numeric strings
    local num = tonumber(value)
    if num and value:match("^%-?%d+%.?%d*$") then
      return "number", num
    end

    -- Check for boolean strings
    local lower = value:lower()
    if lower == "true" then
      return "boolean", true
    elseif lower == "false" then
      return "boolean", false
    end

    return "string", value
  elseif value_type == "table" then
    return "array", vim.json.encode(value)
  else
    return "string", tostring(value)
  end
end

-- Process YAML data into property format
function M.process_properties(yaml_data)
  local flattened = M.flatten_yaml(yaml_data)
  local properties = {}

  for key, value in pairs(flattened) do
    local prop_type, processed_value = M.detect_property_type(value)

    table.insert(properties, {
      key = key,
      value = processed_value,
      value_type = prop_type
    })
  end

  return properties
end

return M