Fix infra topology query-sensitive ranking

[yashkrishan]

Summary

• Wire InfraTopologyReader into the existing semantic similarity scoring path so infra claims are ranked by query relevance instead of a flat default score.
• Keep topology traversal query-free; stamp similarity after discovery so BFS/blast-radius behavior does not change.
• Allow graph read --query on infra_topology.service_neighborhood and extract shared claim_semantic_similarity across readers.

Test plan

• pytest tests/unit/test_p9_readers.py
• pytest tests/conformance/test_graph_surface_lite_e2e.py
• Contract suites: graph workbench ontology, surface-lite contract, CLI contract, views
• FalkorDB roundtrip integration test

Made with Cursor

[coderabbitai]

📝 Walkthrough

Summary by CodeRabbit

• New Features

  • Query-aware ranking is now supported in infrastructure topology results, helping the most relevant items appear first.
  • Search/query input is now accepted across the relevant graph surface for better filtering and ranking.

• Bug Fixes

  • Improved similarity scoring consistency across readers so rankings are more stable and predictable.
  • Boolean or malformed similarity values are now ignored instead of affecting results.

• Tests

  • Added end-to-end and unit coverage for query-sensitive ranking behavior.

Walkthrough

Adds a shared claim_semantic_similarity helper for normalizing semantic similarity values from claim rows, adopts it across seven readers, extends InfraTopologyReader to stamp query-derived similarity scores during traversal, updates ontology view filters to accept query, and adds corresponding unit and e2e tests.

Changes

Semantic similarity propagation

Layer / File(s)	Summary
Shared similarity helper potpie/context-engine/application/readers/_common.py	Adds claim_semantic_similarity(row) returning a normalized float or None (boolean/non-numeric values yield None), and exports it via __all__.
Reader adoption of shared helper application/readers/coding_preferences.py, decisions.py, docs.py, features.py, owners.py, prior_bugs.py, timeline_reader.py	Replaces manual row.properties["semantic_similarity"] extraction and float coercion with calls to claim_semantic_similarity(row) when building Candidate objects; timeline_reader.py also updates its internal sort_key.
Query-based similarity stamping in infra topology application/readers/infra_topology.py	Populates Candidate.semantic_similarity via the shared helper, passes fact_query=req.query on the unanchored traversal path, and adds _stamp_query_similarity to re-query discovered rows by claim_key and stamp similarity scores without changing traversal discovery.
Ontology filter support for query domain/graph_workbench_ontology.py	Adds "query" to optional_scope and supported_filters for the infra_topology.service_neighborhood view override.
Unit and e2e test coverage tests/unit/test_p9_readers.py, tests/conformance/test_graph_surface_lite_e2e.py	Adds tests verifying claim_semantic_similarity ignores booleans, anchored/unanchored/no-query semantic ranking behavior in InfraTopologyReader, and an e2e test confirming query-sensitive ranking through resolve and read surfaces.

Sequence Diagram(s)

sequenceDiagram
  participant Reader as InfraTopologyReader
  participant Traversal as BFS traversal
  participant Store as Claim store

  Reader->>Traversal: traverse anchors (query-free)
  Traversal->>Store: find_claims (structural filters)
  Store-->>Traversal: discovered rows
  Reader->>Reader: _stamp_query_similarity(rows, req.query)
  Reader->>Store: find_claims(claim_key_in, fact_query=req.query)
  Store-->>Reader: similarity-stamped rows
  Reader->>Reader: build Candidate.semantic_similarity via claim_semantic_similarity

Loading

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly summarizes the main change: making infra topology ranking sensitive to the query.
Description check	✅ Passed	The description is directly related to the changeset and matches the implemented infra ranking and query support work.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@potpie/context-engine/application/readers/infra_topology.py`:
- Around line 233-260: The stamping logic in _stamp_query_similarity currently
assumes limit=len(claim_keys) will return all requested rows, but
ClaimQuery.find_claims applies claim_key_in after vector ranking so some keys
can be dropped. Update _stamp_query_similarity to handle misses by doing a keyed
follow-up for any unstamped claim keys, or switch to a query shape that
prefilters by claim_key before similarity ranking. Make sure the fix preserves
the existing row set while still attaching semantic_similarity to every
discovered row.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Repository UI (base), Organization UI (inherited)

Review profile: ASSERTIVE

Plan: Pro

Run ID: 30262d1b-d85a-4160-94b5-8b52f30a0c6c

📥 Commits

Reviewing files that changed from the base of the PR and between a191a5e and 4aff175.

📒 Files selected for processing (12)

• potpie/context-engine/application/readers/_common.py
• potpie/context-engine/application/readers/coding_preferences.py
• potpie/context-engine/application/readers/decisions.py
• potpie/context-engine/application/readers/docs.py
• potpie/context-engine/application/readers/features.py
• potpie/context-engine/application/readers/infra_topology.py
• potpie/context-engine/application/readers/owners.py
• potpie/context-engine/application/readers/prior_bugs.py
• potpie/context-engine/application/readers/timeline_reader.py
• potpie/context-engine/domain/graph_workbench_ontology.py
• potpie/context-engine/tests/conformance/test_graph_surface_lite_e2e.py
• potpie/context-engine/tests/unit/test_p9_readers.py

[linear-code]

POT-1852