Raise HeadersNeeded instead of TypeError when subscripting a headerless Dataset by column name

[gaoflow]

Problem

Dataset.__getitem__ and Dataset.__delitem__ both evaluate key in self.headers when given a string key. When no headers have been set (self.headers is None), Python raises an opaque:

TypeError: argument of type 'NoneType' is not iterable

instead of the library's own HeadersNeeded exception, which every other header-dependent operation (e.g. sort, filter, stack_cols) already raises.

Minimal reproduction:

import tablib
d = tablib.Dataset()
d.append([1, 2, 3])

d["col"]    # TypeError, should be HeadersNeeded
del d["col"]  # TypeError, should be HeadersNeeded

Fix

Add a self.headers is None guard at the top of the isinstance(key, str) branch in both dunder methods that raises HeadersNeeded() before the membership test is attempted.

Tests

Two new tests added to DatasetTests:

• test_getitem_str_key_no_headers_raises — verifies d["col"] raises HeadersNeeded
• test_delitem_str_key_no_headers_raises — verifies del d["col"] raises HeadersNeeded

All 155 tests pass (test_cli_export_github deselected as it is a pre-existing failure unrelated to this change).

[claudep]

Looks good, many thanks for your contribution!

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 93.22%. Comparing base (64f0201) to head (ab10a5f).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #648      +/-   ##
==========================================
+ Coverage   93.19%   93.22%   +0.02%     
==========================================
  Files          29       29              
  Lines        3247     3261      +14     
==========================================
+ Hits         3026     3040      +14     
  Misses        221      221              

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.