nw9663644-eng commented on code in PR #370:
URL: https://github.com/apache/hugegraph-ai/pull/370#discussion_r3612350023
##########
hugegraph-llm/src/hugegraph_llm/config/models/base_prompt_config.py:
##########
@@ -140,6 +141,7 @@ def to_literal(val):
"gremlin_generate_prompt":
to_literal(self.gremlin_generate_prompt),
"doc_input_text": to_literal(self.doc_input_text),
"graph_extract_split_type":
to_literal(self.graph_extract_split_type),
+ "graph_extract_max_workers": self.graph_extract_max_workers,
Review Comment:
Thanks for the detailed review. I addressed the remaining coverage and
failure-context gaps.
Changes included:
- Moved per-chunk parsing, normalization, schema filtering, and
`filter_item()` into the chunk-level exception wrapper so item-level failures
now preserve `Graph extraction failed for chunk x/y` context.
- Added validation for malformed item fields such as non-mapping
`properties`.
- Strengthened the in-flight lifecycle test by running `extractor.run()` in
a separate thread and asserting it does not return before the slow in-flight
call finishes.
- Added a REST success-path regression test to verify a non-default
`graph_extract_max_workers=4` reaches `schedule_flow()`, plus default
forwarding coverage.
- Reworked the concurrency-limit test to use deterministic overlap
coordination instead of relying on sleep timing.
- Added prompt-config persistence coverage for non-default
`graph_extract_max_workers` round-trip and old-config default behavior.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
