fix: prevent None file_path from propagating as unknown_source

[he-yufeng]

Summary

Fixes the root cause of #2764 where some documents show "unknown_source" instead of their actual filenames.

Root Cause

getattr(status_doc, "file_path", "unknown_source") only returns the fallback when the attribute doesn't exist. When file_path is present but None, it returns None — which then permanently overwrites valid filenames during reprocessing or consistency validation.

Same pattern in JsonDocStatusStorage: "file_path" not in data passes when the key exists with a null value, creating DocProcessingStatus(file_path=None).

Changes

lightrag/lightrag.py (4 sites):

# Before — returns None when file_path attribute is None
file_path = getattr(status_doc, "file_path", "unknown_source")

# After — catches both missing and None
file_path = getattr(status_doc, "file_path", None) or "unknown_source"

lightrag/kg/json_doc_status_impl.py (3 sites):

# Before — only catches missing key, not null value
if "file_path" not in data:
    data["file_path"] = "no-file-path"

# After — catches missing, None, and empty string
if not data.get("file_path"):
    data["file_path"] = "no-file-path"

Complements #2793 which fixed the API input layer — this PR fixes the processing pipeline and storage layer.

Test plan

• 275 tests pass, 33 skipped (pre-existing skip/xfail)
• 1 pre-existing failure in test_interactive_setup_outputs.py unrelated to this change

[danielaskdd]

@codex review

[danielaskdd]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. 🎉

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".