notex.nvim/lua/notex/utils/errors.lua

-- Centralized error handling and recovery system
local M = {}

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

-- Error types with specific handling strategies
local ERROR_TYPES = {
  DATABASE_CONNECTION = {
    category = "database",
    recoverable = true,
    retry_strategy = "exponential_backoff",
    max_retries = 3,
    user_message = "Database connection error. Retrying..."
  },
  DATABASE_QUERY = {
    category = "database",
    recoverable = false,
    retry_strategy = "none",
    max_retries = 0,
    user_message = "Query execution failed. Please check your query syntax."
  },
  FILE_NOT_FOUND = {
    category = "filesystem",
    recoverable = true,
    retry_strategy = "immediate",
    max_retries = 1,
    user_message = "File not found. It may have been moved or deleted."
  },
  FILE_PARSE_ERROR = {
    category = "parsing",
    recoverable = false,
    retry_strategy = "none",
    max_retries = 0,
    user_message = "Failed to parse file. Please check the file format."
  },
  QUERY_SYNTAX_ERROR = {
    category = "query",
    recoverable = false,
    retry_strategy = "none",
    max_retries = 0,
    user_message = "Query syntax error. Please check your query syntax."
  },
  VALIDATION_ERROR = {
    category = "validation",
    recoverable = false,
    retry_strategy = "none",
    max_retries = 0,
    user_message = "Validation error. Please check your input."
  },
  UI_ERROR = {
    category = "ui",
    recoverable = true,
    retry_strategy = "immediate",
    max_retries = 1,
    user_message = "UI error. Attempting to recover..."
  },
  PERMISSION_ERROR = {
    category = "filesystem",
    recoverable = false,
    retry_strategy = "none",
    max_retries = 0,
    user_message = "Permission denied. Please check file permissions."
  },
  NETWORK_ERROR = {
    category = "network",
    recoverable = true,
    retry_strategy = "exponential_backoff",
    max_retries = 3,
    user_message = "Network error. Retrying..."
  },
  PERFORMANCE_TIMEOUT = {
    category = "performance",
    recoverable = true,
    retry_strategy = "immediate",
    max_retries = 1,
    user_message = "Operation timed out. Retrying with simpler approach..."
  }
}

-- Error state tracking
local error_state = {
  recent_errors = {},
  error_counts = {},
  last_recovery_attempt = {},
  recovery_in_progress = {}
}

-- Create standardized error object
function M.create_error(error_type, message, context, original_error)
  local error_def = ERROR_TYPES[error_type] or ERROR_TYPES.UI_ERROR
  local error_obj = {
    type = error_type,
    message = message,
    context = context or {},
    original_error = original_error,
    timestamp = os.time(),
    recoverable = error_def.recoverable,
    category = error_def.category,
    user_message = error_def.user_message,
    retry_strategy = error_def.retry_strategy,
    max_retries = error_def.max_retries,
    error_id = M.generate_error_id()
  }

  -- Track error
  M.track_error(error_obj)

  return error_obj
end

-- Generate unique error ID
function M.generate_error_id()
  return string.format("ERR_%d_%s", os.time(), math.random(1000, 9999))
end

-- Track error occurrence
function M.track_error(error_obj)
  -- Add to recent errors
  table.insert(error_state.recent_errors, error_obj)

  -- Keep only last 50 errors
  if #error_state.recent_errors > 50 then
    table.remove(error_state.recent_errors, 1)
  end

  -- Update error counts
  local key = error_obj.type
  error_state.error_counts[key] = (error_state.error_counts[key] or 0) + 1

  -- Log the error
  logging.handle_error(error_obj.message, error_obj.category, error_obj)
end

-- Check if error should be retried
function M.should_retry(error_obj, current_attempt)
  if not error_obj.recoverable then
    return false, "Error is not recoverable"
  end

  if current_attempt >= error_obj.max_retries then
    return false, "Maximum retries exceeded"
  end

  -- Check if we recently attempted recovery for this error type
  local last_attempt = error_state.last_recovery_attempt[error_obj.type]
  if last_attempt and (os.time() - last_attempt) < 5 then
    return false, "Recovery attempt too recent"
  end

  return true, "Retry allowed"
end

-- Execute operation with error handling and recovery
function M.safe_execute(operation, error_type, context, func, ...)
  local current_attempt = 0
  local max_attempts = (ERROR_TYPES[error_type] and ERROR_TYPES[error_type].max_retries or 0) + 1

  while current_attempt < max_attempts do
    local success, result = pcall(func, ...)

    if success then
      -- Reset recovery state on success
      error_state.last_recovery_attempt[error_type] = nil
      error_state.recovery_in_progress[error_type] = nil
      return true, result
    else
      current_attempt = current_attempt + 1
      local error_obj = M.create_error(error_type, result, context)

      local should_retry, retry_reason = M.should_retry(error_obj, current_attempt)

      if should_retry and current_attempt < max_attempts then
        error_state.last_recovery_attempt[error_type] = os.time()
        error_state.recovery_in_progress[error_type] = true

        -- Show user message
        if error_obj.user_message then
          vim.notify(error_obj.user_message, vim.log.levels.WARN)
        end

        -- Apply retry strategy
        M.apply_retry_strategy(error_obj.retry_strategy, current_attempt)

        logging.info("Retrying operation", {
          operation = operation,
          attempt = current_attempt,
          error_type = error_type,
          reason = retry_reason
        })
      else
        -- Final failure
        error_state.recovery_in_progress[error_type] = nil

        -- Show final error message
        M.show_final_error(error_obj, current_attempt)

        return false, error_obj
      end
    end
  end

  return false, "Operation failed after all retry attempts"
end

-- Apply retry strategy
function M.apply_retry_strategy(strategy, attempt)
  if strategy == "immediate" then
    -- No delay
  elseif strategy == "exponential_backoff" then
    local delay = math.min(2 ^ attempt, 10) -- Cap at 10 seconds
    vim.defer_fn(function() end, delay * 1000)
  elseif strategy == "linear_backoff" then
    local delay = attempt * 1000 -- 1 second per attempt
    vim.defer_fn(function() end, delay)
  end
end

-- Show final error to user
function M.show_final_error(error_obj, attempt_count)
  local message = string.format("%s (%d attempts made)", error_obj.user_message or error_obj.message, attempt_count)

  if error_obj.category == "validation" or error_obj.category == "query" then
    vim.notify(message, vim.log.levels.ERROR)
  elseif error_obj.category == "filesystem" or error_obj.category == "database" then
    vim.notify(message, vim.log.levels.ERROR)
  else
    vim.notify(message, vim.log.levels.WARN)
  end
end

-- Wrap function for safe execution
function M.wrap(operation_name, error_type, func)
  return function(...)
    return M.safe_execute(operation_name, error_type, {operation = operation_name}, func, ...)
  end
end

-- Handle specific error types with custom recovery
local error_handlers = {}

function M.register_error_handler(error_type, handler)
  error_handlers[error_type] = handler
end

function M.handle_specific_error(error_obj)
  local handler = error_handlers[error_obj.type]
  if handler then
    local success, result = pcall(handler, error_obj)
    if success then
      return result
    else
      logging.error("Error handler failed", {
        error_type = error_obj.type,
        handler_error = result
      })
    end
  end

  return nil
end

-- Register default error handlers
M.register_error_handler("DATABASE_CONNECTION", function(error_obj)
  -- Try to reinitialize database connection
  local database = require('notex.database.init')
  local ok, err = database.reconnect()
  if ok then
    vim.notify("Database connection restored", vim.log.levels.INFO)
    return true
  end
  return false
end)

M.register_error_handler("FILE_NOT_FOUND", function(error_obj)
  -- Remove from index if file no longer exists
  if error_obj.context and error_obj.context.file_path then
    local indexer = require('notex.index')
    local ok, err = indexer.remove_document_by_path(error_obj.context.file_path)
    if ok then
      vim.notify("Removed missing file from index", vim.log.levels.INFO)
      return true
    end
  end
  return false
end)

M.register_error_handler("UI_ERROR", function(error_obj)
  -- Try to cleanup UI state
  local ui = require('notex.ui')
  ui.cleanup_all()
  vim.notify("UI state reset", vim.log.levels.INFO)
  return true
end)

-- Get error statistics
function M.get_error_statistics()
  local stats = {
    total_errors = 0,
    by_type = vim.deepcopy(error_state.error_counts),
    recent_errors = vim.list_slice(error_state.recent_errors, -10), -- Last 10 errors
    recovery_in_progress = vim.deepcopy(error_state.recovery_in_progress)
  }

  -- Calculate total
  for _, count in pairs(error_state.error_counts) do
    stats.total_errors = stats.total_errors + count
  end

  -- Get error rate in last hour
  local one_hour_ago = os.time() - 3600
  local recent_count = 0
  for _, error in ipairs(error_state.recent_errors) do
    if error.timestamp > one_hour_ago then
      recent_count = recent_count + 1
    end
  end
  stats.errors_per_hour = recent_count

  return stats
end

-- Clear error history
function M.clear_error_history()
  error_state.recent_errors = {}
  error_state.error_counts = {}
  error_state.last_recovery_attempt = {}
  error_state.recovery_in_progress = {}

  logging.info("Error history cleared")
end

-- Check system health based on errors
function M.check_system_health()
  local stats = M.get_error_statistics()
  local health = {
    status = "healthy",
    issues = {},
    recommendations = {}
  }

  -- Check error rate
  if stats.errors_per_hour > 10 then
    health.status = "degraded"
    table.insert(health.issues, "High error rate: " .. stats.errors_per_hour .. " errors/hour")
    table.insert(health.recommendations, "Check system logs for recurring issues")
  end

  -- Check for stuck recovery operations
  local stuck_recoveries = 0
  for error_type, in_progress in pairs(error_state.recovery_in_progress) do
    if in_progress then
      stuck_recoveries = stuck_recoveries + 1
    end
  end

  if stuck_recoveries > 0 then
    health.status = "degraded"
    table.insert(health.issues, stuck_recoveries .. " recovery operations in progress")
    table.insert(health.recommendations, "Consider restarting the plugin")
  end

  -- Check for specific error patterns
  local db_errors = error_state.error_counts["DATABASE_CONNECTION"] or 0
  local file_errors = error_state.error_counts["FILE_NOT_FOUND"] or 0

  if db_errors > 5 then
    health.status = "unhealthy"
    table.insert(health.issues, "Frequent database connection errors")
    table.insert(health.recommendations, "Check database file permissions and disk space")
  end

  if file_errors > 10 then
    table.insert(health.issues, "Many file not found errors")
    table.insert(health.recommendations, "Consider reindexing the workspace")
  end

  return health
end

-- Create user-friendly error messages
function M.format_error_for_user(error_obj)
  local message = error_obj.user_message or error_obj.message

  -- Add contextual information
  if error_obj.context.operation then
    message = message .. " (during: " .. error_obj.context.operation .. ")"
  end

  if error_obj.context.file_path then
    message = message .. " (file: " .. vim.fn.fnamemodify(error_obj.context.file_path, ":t") .. ")"
  end

  -- Add error ID for support
  message = message .. " [ID: " .. error_obj.error_id .. "]"

  return message
end

-- Export error types for use in other modules
M.ERROR_TYPES = ERROR_TYPES

return M