Add db/ source files that were blocked by overly broad gitignore

The old `db/` pattern in .gitignore matched src/yc_bench/db/ too,
preventing all ORM models and session.py from being committed.
Previous commit fixed .gitignore to `/db/`; this adds the 10 missing files.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

src/yc_bench/db/session.py

@@ -0,0 +1,94 @@
+from __future__ import annotations
+
+import os
+from contextlib import contextmanager
+
+from sqlalchemy import create_engine
+from sqlalchemy.orm import Session, sessionmaker
+
+
+def _get_database_url() -> str:
+    default = "sqlite:///db/yc_bench.db"
+    url = os.environ.get("DATABASE_URL", default)
+    # Auto-create parent directory for sqlite:/// relative paths
+    if url.startswith("sqlite:///") and not url.startswith("sqlite:////"):
+        import pathlib
+        db_path = pathlib.Path(url[len("sqlite:///"):])
+        db_path.parent.mkdir(parents=True, exist_ok=True)
+    return url
+
+
+def _maybe_register_psycopg_enum_dumpers(url: str) -> None:
+    """Register psycopg3 enum value dumpers — only needed for PostgreSQL.
+
+    psycopg3 sends enum .name (uppercase) by default; the DB stores .value
+    (lowercase). This adapter fixes that mismatch. SQLite doesn't need it
+    because SQLAlchemy stores enums as plain VARCHAR using .value directly.
+    """
+    if not url.startswith("postgresql"):
+        return
+
+    import enum
+    import psycopg
+    from psycopg.adapt import Dumper
+    from psycopg.pq import Format
+
+    class _EnumValueDumper(Dumper):
+        format = Format.TEXT
+
+        def dump(self, obj):
+            return obj.value.encode("utf-8") if isinstance(obj, enum.Enum) else str(obj).encode("utf-8")
+
+    from .models.company import Domain
+    from .models.event import EventType
+    from .models.ledger import LedgerCategory
+    from .models.task import TaskStatus
+
+    for cls in (Domain, EventType, LedgerCategory, TaskStatus):
+        psycopg.adapters.register_dumper(cls, _EnumValueDumper)
+
+
+def build_engine(url: str | None = None):
+    db_url = url or _get_database_url()
+    _maybe_register_psycopg_enum_dumpers(db_url)
+
+    kwargs: dict = {"echo": False, "future": True}
+    if db_url.startswith("sqlite"):
+        # Required for SQLAlchemy's connection pool when used across threads
+        kwargs["connect_args"] = {"check_same_thread": False}
+
+    return create_engine(db_url, **kwargs)
+
+
+def build_session_factory(engine):
+    return sessionmaker(
+        bind=engine,
+        autoflush=False,
+        autocommit=False,
+        expire_on_commit=False,
+        class_=Session,
+    )
+
+
+@contextmanager
+def session_scope(session_factory):
+    session = session_factory()
+    try:
+        yield session
+        session.commit()
+    except Exception:
+        session.rollback()
+        raise
+    finally:
+        session.close()
+
+
+def init_db(engine) -> None:
+    """Create all tables that do not yet exist. Safe to call on every startup."""
+    from .base import Base
+    # Import all models so SQLAlchemy registers them with Base.metadata before create_all.
+    from .models import company, employee, event, ledger, scratchpad, session, sim_state, task  # noqa: F401
+    Base.metadata.create_all(engine)
+
+
+__all__ = ["build_engine", "build_session_factory", "session_scope", "init_db"]