Use effective identity for DryRunEvaluator node_key

[ptomecek]

Summary

DryRunEvaluator builds a plan of the nodes that would run and emits a node_key identifying each one. That key was computed with the plain structural cache_key(...), while the actual runtime — the in-process MemoryCacheEvaluator and the dependency-graph dedup — identifies the same nodes with cache_key(..., effective=True).

So the dry run could label a node differently than the real run caches and dedupes it. For models that use a narrower "effective" identity, two nodes that the real run treats as distinct could share a node_key in the plan (or vice-versa) — making the dry-run report an unreliable preview of what actually happens.

Fix

Compute the dry-run node_key with cache_key(..., effective=True), the same identity the memory cache and graph dedup already use. The dry-run plan now matches runtime behavior node-for-node.

This is a no-op for ordinary models (which use the structural key), so existing plans are unchanged; it only affects models that opt into effective identity.

Test

Adds test_node_key_uses_effective_identity: a model that ignores an unused context field. Two contexts that differ only in that ignored field previously produced different node_keys in the plan; now they collapse to the same key, matching how the runtime caches/dedupes them — and the emitted key equals cache_key(..., effective=True).

Validation

ccflow/tests/evaluators/ and ccflow/tests/test_effective_key_characterization.py pass (75 tests), including all pre-existing dry-run tests.

[github-actions]

Test Results

    1 files  ±0      1 suites  ±0   3m 17s ⏱️ +17s
1 329 tests +1  1 327 ✅ +1  2 💤 ±0  0 ❌ ±0 
1 335 runs  +1  1 333 ✅ +1  2 💤 ±0  0 ❌ ±0 

Results for commit b80b76a. ± Comparison against base commit 6a788a6.

♻️ This comment has been updated with latest results.

[codecov]

Codecov Report

❌ Patch coverage is 95.23810% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 93.48%. Comparing base (6a788a6) to head (b80b76a).
⚠️ Report is 3 commits behind head on main.

Files with missing lines	Patch %	Lines
ccflow/tests/evaluators/test_reporting.py	95.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #262   +/-   ##
=======================================
  Coverage   93.48%   93.48%           
=======================================
  Files         176      176           
  Lines       20327    20347   +20     
  Branches     1350     1350           
=======================================
+ Hits        19002    19021   +19     
- Misses       1052     1053    +1     
  Partials      273      273           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.