Fix JSONL loader handling of blank/invalid lines

[natoverse]

Summary

• Skip blank/whitespace-only lines in JSONL input loader
• Skip malformed JSON rows safely
• Skip non-object JSON rows
• Add regression test and fixture coverage
• Add semversioner patch entry

Fixes #2424

[natoverse]

This replaces #2423 to include more handling of edge cases

[Copilot]

Pull request overview

Improves robustness of the JSONL input reader so that a single blank or invalid line doesn’t cause the entire file to be skipped, aligning behavior with the reported issue (#2424).

Changes:

• Updated the JSONL reader to ignore blank/whitespace-only lines, skip malformed JSON rows, and skip non-object JSON values.
• Added a regression test and a JSONL fixture covering blank, invalid, and non-object rows.
• Added a semversioner patch note documenting the behavior change.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

File	Description
packages/graphrag-input/graphrag_input/jsonl.py	Makes JSONL parsing resilient by skipping blank/invalid/non-object rows and continuing through the file.
tests/unit/indexing/input/test_jsonl_loader.py	Adds regression coverage to ensure only valid object rows become documents.
tests/unit/indexing/input/data/jsonl-with-invalid-and-blank-lines/input.jsonl	Provides a fixture with blank, invalid JSON, and non-object rows interleaved with valid objects.
.semversioner/next-release/patch-20260710221345787890.json	Records the patch-level change in JSONL loader behavior.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.