session-memory Phase 0: budget-based retention sweep (T05)

- session_memory/core/retention.py: RetentionConfig + sweep() with backstop,
  budget (oldest-analyzed-first, never touches un-analyzed), and hard-cap
  overflow (analyze-now then reported last-resort data_loss); EvictionReport
- tests/test_retention.py covers all four branches

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

tests/test_retention.py

@@ -0,0 +1,107 @@
+"""Retention tests (T05): each pass of the budget-based eviction, with tiny caps."""
+
+import os
+import sys
+from datetime import datetime, timedelta, timezone
+
+sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+
+from session_memory.adapters.claude import Normalized  # noqa: E402
+from session_memory.core import digest as digest_mod  # noqa: E402
+from session_memory.core.retention import RetentionConfig, sweep  # noqa: E402
+from session_memory.core.schema import Cost, Session, SessionEvent  # noqa: E402
+from session_memory.core.store import Store  # noqa: E402
+
+NOW = datetime(2026, 6, 6, tzinfo=timezone.utc)
+
+
+def _ingest(st, native, *, body_bytes=1000, ended=None, analyze=False):
+    uid = Session.make_uid("claude", native)
+    s = Session(session_uid=uid, flavor="claude", native_session_id=native,
+                ended_at=(ended or NOW).strftime("%Y-%m-%dT%H:%M:%SZ"),
+                ingested_at=NOW.strftime("%Y-%m-%dT%H:%M:%SZ"))
+    ref = f"blob://{native}/0"
+    events = [SessionEvent(session_uid=uid, seq=0, kind="assistant_msg", payload_ref=ref)]
+    st.ingest(Normalized(session=s, events=events, blobs={ref: "x" * body_bytes}))
+    if analyze:
+        digest_mod.analyze(st, uid)
+    return uid
+
+
+def _store(tmp_path):
+    return Store(str(tmp_path / "m.db"), str(tmp_path / "blobs"))
+
+
+def test_backstop_evicts_old_analyzed_only(tmp_path):
+    st = _store(tmp_path)
+    old = _ingest(st, "old", ended=NOW - timedelta(days=60), analyze=True)
+    young = _ingest(st, "young", ended=NOW - timedelta(days=1), analyze=True)
+    unanalyzed_old = _ingest(st, "oldraw", ended=NOW - timedelta(days=60), analyze=False)
+
+    cfg = RetentionConfig(raw_soft_cap_bytes=10**12, raw_hard_cap_bytes=10**12, raw_max_age_days=45)
+    rep = sweep(st, cfg, now=NOW)
+
+    assert old in rep.backstop_evicted
+    assert young not in rep.backstop_evicted          # too recent
+    assert unanalyzed_old not in rep.backstop_evicted  # not analyzed -> protected
+    assert st.get_session(old).evicted_at is not None
+    assert st.get_session(unanalyzed_old).evicted_at is None
+
+
+def test_budget_pass_evicts_oldest_analyzed_first(tmp_path):
+    st = _store(tmp_path)
+    a = _ingest(st, "a", body_bytes=2000, ended=NOW - timedelta(days=3), analyze=True)
+    b = _ingest(st, "b", body_bytes=2000, ended=NOW - timedelta(days=2), analyze=True)
+    c = _ingest(st, "c", body_bytes=2000, ended=NOW - timedelta(days=1), analyze=True)
+
+    # soft cap that forces evicting ~two of the three
+    cfg = RetentionConfig(raw_soft_cap_bytes=2500, raw_hard_cap_bytes=10**9, raw_max_age_days=10**6)
+    rep = sweep(st, cfg, now=NOW)
+
+    assert rep.budget_evicted[:2] == [a, b]            # oldest-first
+    assert st.get_session(c).evicted_at is None        # newest survives
+    assert st.tier1_usage_bytes() <= cfg.raw_soft_cap_bytes
+
+
+def test_budget_pass_never_touches_unanalyzed(tmp_path):
+    st = _store(tmp_path)
+    raw1 = _ingest(st, "r1", body_bytes=5000, analyze=False)
+    raw2 = _ingest(st, "r2", body_bytes=5000, analyze=False)
+
+    cfg = RetentionConfig(raw_soft_cap_bytes=100, raw_hard_cap_bytes=10**9, raw_max_age_days=10**6)
+    rep = sweep(st, cfg, now=NOW)
+
+    # over soft cap but nothing analyzed -> no eviction, no data loss
+    assert rep.budget_evicted == []
+    assert rep.lost_data is False
+    assert st.get_session(raw1).evicted_at is None
+    assert st.get_session(raw2).evicted_at is None
+    assert st.tier1_usage_bytes() > cfg.raw_soft_cap_bytes  # tolerated, not destroyed
+
+
+def test_overflow_analyzes_then_evicts_without_data_loss(tmp_path):
+    st = _store(tmp_path)
+    r1 = _ingest(st, "r1", body_bytes=4000, ended=NOW - timedelta(days=2), analyze=False)
+    r2 = _ingest(st, "r2", body_bytes=4000, ended=NOW - timedelta(days=1), analyze=False)
+
+    cfg = RetentionConfig(raw_soft_cap_bytes=3000, raw_hard_cap_bytes=5000, raw_max_age_days=10**6)
+    rep = sweep(st, cfg, now=NOW, analyze_fn=digest_mod.analyze)
+
+    # overflow path analyzed the un-analyzed sessions, then budget-evicted
+    assert set(rep.overflow_analyzed) == {r1, r2}
+    assert rep.lost_data is False                       # analysis avoided data loss
+    assert st.tier1_usage_bytes() <= cfg.raw_soft_cap_bytes
+
+
+def test_overflow_last_resort_reports_data_loss(tmp_path):
+    st = _store(tmp_path)
+    # one un-analyzed session bigger than the hard cap, analysis disabled (no fn)
+    big = _ingest(st, "big", body_bytes=20000, analyze=False)
+
+    cfg = RetentionConfig(raw_soft_cap_bytes=1000, raw_hard_cap_bytes=2000, raw_max_age_days=10**6)
+    rep = sweep(st, cfg, now=NOW, analyze_fn=None)
+
+    assert big in rep.overflow_data_loss
+    assert rep.lost_data is True
+    assert any("data_loss" in w for w in rep.warnings)
+    assert st.get_session(big).evicted_at is not None