Feature: Header Loading and Column Detection
  As a developer
  I want _load_headers to handle all edge cases
  So that column detection is robust

  Scenario: Load headers when sheet is None
    Given a tracker is created but not authenticated
    And sheet attribute is None
    When _load_headers is called internally
    Then it should return early without error
    And headers should remain empty
    And column indices should remain None

  Scenario: Load headers successfully identifies all columns
    Given an authenticated tracker
    And a sheet with headers "Dataset Name, Status, Error"
    When the tracker loads headers
    Then dataset_name_col should be 0
    And status_col should be 1
    And error_col should be 2
    And headers should contain lowercased column names

  Scenario: Load headers with alternate column names for dataset
    Given an authenticated tracker
    And a sheet with headers "ID, Status, Error"
    When the tracker loads headers
    Then dataset_name_col should be identified as 0
    And status_col should be 1
    And error_col should be 2

  Scenario: Load headers with alternate column names for status
    Given an authenticated tracker
    And a sheet with headers "Dataset Name, State, Error"
    When the tracker loads headers
    Then dataset_name_col should be 0
    And status_col should be identified as 1
    And error_col should be 2

  Scenario: Load headers with alternate column names for error
    Given an authenticated tracker
    And a sheet with headers "Dataset Name, Status, Error Message"
    When the tracker loads headers
    Then dataset_name_col should be 0
    And status_col should be 1
    And error_col should be identified as 2


  Scenario: Load headers with case insensitive matching
    Given an authenticated tracker
    And a sheet with headers "DATASET NAME, STATUS, ERROR"
    When the tracker loads headers
    Then all columns should be identified correctly
    And headers should be normalized to lowercase

  Scenario: Load headers with extra whitespace
    Given an authenticated tracker
    And a sheet with headers "  Dataset Name  , Status, Error  "
    When the tracker loads headers
    Then all columns should be identified correctly after trimming

  Scenario: Load headers when exception occurs during API call
    Given an authenticated tracker
    And the Google Sheets API raises exception during header fetch
    When the tracker attempts to load headers
    Then it should catch the exception
    And set default column indices (0, 1, 2)
    And headers should be empty list

  Scenario: Load headers with _find_column_index trying multiple possible names
    Given an authenticated tracker
    And a sheet with headers "dataset_id, processing status, error message"
    When the tracker loads headers using _find_column_index
    Then dataset_name_col should match "dataset_id"
    And status_col should match "processing status"
    And error_col should match "error message"

  Scenario: Load headers converts all header names to lowercase
    Given an authenticated tracker
    And a sheet with mixed case headers "DaTaSeT NaMe, StAtUs, ErRoR"
    When the tracker loads headers
    Then all headers should be stored in lowercase
    And column matching should work case-insensitively