Skip to content

context-graph: sessions-graph recall(), the read path the eval benchmarks #424

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
antejavor merged 1 commit into from
Oct 5, 2026
Merged

context-graph: sessions-graph recall(), the read path the eval benchmarks #424

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 4 additions & 12 deletions context-graph/eval/src/context_graph_eval/cli.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,7 @@
from .convert.longmemeval import DEFAULT_REVISION, build_corpus, fetch, haystack_path, load_raw
from .corpus import write_corpus
from .hybrid import LANES as HYBRID_LANES
from .hybrid import HybridConfig
from .hybrid import RecallConfig
from .official_judge import OFFICIAL_JUDGE_MODEL
from .reconcile import EXTRACTION_BACKENDS
from .runner import RETRIEVAL_STRATEGIES
Expand Down Expand Up @@ -127,13 +127,8 @@ def main(argv: list[str] | None = None) -> int:
"full-text index over raw, UNRECONCILED turns -- no distillation, no LLM extraction cost). "
"'text-search' forces reconciliation off regardless of --skip-reconcile: there is no memory "
"for it to build that this strategy would read. 'hybrid' finds turns by vector and text "
"search and uses the typed graph to find facts and more turns (see hybrid.py).",
)
run.add_argument(
"--hybrid-user-fact-types-from",
choices=("facts", "names"),
default="names",
help="how the hybrid user_facts lane picks relation types (see hybrid.HybridConfig).",
"search and uses the typed graph to find facts and more turns: sessions-graph's recall, the "
"product read path (see hybrid.py).",
)
run.add_argument(
"--hybrid-lanes",
Expand Down Expand Up @@ -477,10 +472,7 @@ def _run(args) -> int:
extraction_backend=args.extraction_backend,
retrieval_strategy=args.retrieval_strategy,
text_search_limit=args.text_search_limit,
hybrid=HybridConfig(
lanes=tuple(lane for lane in args.hybrid_lanes.split(",") if lane),
user_fact_types_from=args.hybrid_user_fact_types_from,
),
hybrid=RecallConfig.from_mapping({"lanes": args.hybrid_lanes}),
official_judge_model=official_model,
),
)
Expand Down
Loading
Loading