WSDataset: added an ignore_index option; ws_indexer: improved error handling; ws_indexer: ensure we use relative shard paths when possible

[jpc]

📚 PR Stack

#	Branch	Base	PR
1	jpc/preloading	main	#16
2	jpc/validate-shards-in-sql	jpc/preloading	#18
3	jpc/shard-pipe	jpc/validate-shards-in-sql	#19
4	jpc/sql-without-index	jpc/shard-pipe	#20
5	jpc/special-shard-columns	jpc/sql-without-index	#21
6	jpc/remove-in-progress	jpc/special-shard-columns	#22
7	jpc/fix-jupyter-repr	jpc/remove-in-progress	#23
8	jpc/sql-select-dotted	jpc/fix-jupyter-repr	#24
9	jpc/optimize-durations	jpc/sql-select-dotted	#25
10	jpc/source-links	jpc/optimize-durations	#26
11	jpc/audio-qol	jpc/source-links	#27
12	jpc/warn-subsampling-in-dataloader	jpc/audio-qol	#28
13	jpc/wsds-inspect-head	jpc/warn-subsampling-in-dataloader	#29
14	jpc/feather-index	jpc/wsds-inspect-head	#30
15	jpc/indexer	jpc/feather-index	#31
16	jpc/keys-diff	jpc/indexer	#32
17	jpc/fields-cleanup	jpc/keys-diff	#33
18	jpc/s3-shards	jpc/fields-cleanup	#34
19	jpc/indexing-fixes	jpc/s3-shards	#35 ◀️
20	jpc/validate-keys-on-the-fly	jpc/indexing-fixes	#36
21	jpc/repr-missing-last	jpc/validate-keys-on-the-fly	#37
22	jpc/drop-slots	jpc/repr-missing-last	#38

---

• WSDataset: added an ignore_index option
• ws_indexer: improved error handling
• ws_indexer: ensure we use relative shard paths when possible