Flatten output line numbering to a single 1..N sequence (HPE, HP, Lenovo)

[regalen]

Summary

Flattens the output Item column (line_sequence) so every emitted line takes the next number in a single running sequence (1, 2, 3, …) instead of sub-numbering bundle/config components as parent.NN (1.01, 1.02, …). Applied across all parsers that previously nested:

• HPE Bid (XLSX) — BundleDetails flatten under the running counter.
• HP Bid (XLSX) — Bundle Detail rows flatten.
• HP OneConfig (XLSX) — components flatten to 2, 3, … after the parent 1.
• Lenovo BRDA DCG (PDF + XLS) — child/component rows flatten.

Zebra was already flat and is unchanged.

Notes

• Comment placement is unchanged — Max Qty still appears only on Part Number/Bundle lines.
• Because the sequence no longer encodes hierarchy, parent vs child is now distinguished by cost (parents priced > 0; children dropped to 0, exported as the 0.0001 sentinel). Lenovo tests were reworked accordingly.
• Downstream consumers that relied on the dotted Item numbers to identify child lines should now use cost instead.

Verification

• Updated all affected parser tests and regenerated the affected goldens.
• Full parsing suite passes: dotnet test → 178 passed.
• Docs refreshed: per-format specs, output_mapping.md, project_memory.md.