feat(agents): semantic task clustering + focus-area inferred preferred_areas (#97, #113)

- New ml/agents/clustering.py: embed task content via nomic-embed-text
  (Ollama), greedy cosine clustering (threshold 0.72, max 6 clusters),
  graceful fallback to project-id grouping when Ollama is unreachable
- focus_area v2.0.0: compute() uses semantic clusters as focus areas;
  adds preferred_areas InferredParam inferred from top-2 projects by
  task_completion count
- 135 tests, all passing

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
2026-05-06 06:54:46 +00:00
parent 336644a90a
commit 26fc67776f
5 changed files with 404 additions and 41 deletions

View File

@@ -662,4 +662,51 @@ class TestFocusAreaPreferredAreas:
def test_version_bumped(self):
from ml.agents.focus_area import MANIFEST as FA_MANIFEST
assert FA_MANIFEST.version == "1.1.0"
assert FA_MANIFEST.version == "2.0.0"
def test_snapshot_uses_cluster_keys(self):
tasks = [self._task("T", "work")]
out = self.agent.compute(_inp(tasks=tasks))
assert "top_cluster_label" in out.signals_snapshot
assert "cluster_count" in out.signals_snapshot
assert "strategy" in out.signals_snapshot
# ── focus-area: preferred_areas inference from task_completions (#113) ────────
class TestFocusAreaPreferredAreasInference:
from ml.agents.focus_area import MANIFEST as _FA_MANIFEST
def _completion(self, project_id: str) -> TaskCompletion:
return _completion(project_id, lateness_days=0.0)
def test_cold_start_no_completions(self):
history = _history(completions=[])
from ml.agents.focus_area import MANIFEST as FA_MANIFEST
result = run_inference(FA_MANIFEST, history)
assert result["preferred_areas"] == []
def test_top_two_projects_returned(self):
completions = (
[_completion("p1", 0)] * 8
+ [_completion("p2", 0)] * 5
+ [_completion("p3", 0)] * 2
)
history = _history(completions=completions)
from ml.agents.focus_area import MANIFEST as FA_MANIFEST
result = run_inference(FA_MANIFEST, history)
assert result["preferred_areas"] == ["p1", "p2"]
def test_single_project_returns_one(self):
completions = [_completion("work", 0)] * 6
history = _history(completions=completions)
from ml.agents.focus_area import MANIFEST as FA_MANIFEST
result = run_inference(FA_MANIFEST, history)
assert result["preferred_areas"] == ["work"]
def test_none_project_id_ignored(self):
completions = [_completion(None, 0)] * 5 + [_completion("real", 0)] * 3
history = _history(completions=completions)
from ml.agents.focus_area import MANIFEST as FA_MANIFEST
result = run_inference(FA_MANIFEST, history)
assert result["preferred_areas"] == ["real"]