feat: v0.2.0 sprint — ground truth eval, crossover/mutation, checkpointing, similarity guards, dataset loader, CLI commands, extended test coverage

Aggregates all v0.2.0 sprint work (GARAA-30 through GARAA-40) and fixes
2 integration tests that broke when the codebase went async (DSPyLLMAdapter
and full pipeline tests now properly await coroutines).

277 tests pass (260 unit + 17 integration).

Co-Authored-By: Paperclip <noreply@paperclip.ing>

tests/unit/test_logging.py

@@ -0,0 +1,189 @@
+"""Unit tests for structured logging configuration."""
+from __future__ import annotations
+
+import json
+import logging
+from pathlib import Path
+
+from prometheus.cli.logging_setup import configure_logging, get_logger
+
+
+class TestConfigureLogging:
+    def _count_handlers(self, name: str = "prometheus") -> int:
+        return len(logging.getLogger(name).handlers)
+
+    def test_default_creates_console_handler(self) -> None:
+        configure_logging(level=logging.INFO)
+        prom = logging.getLogger("prometheus")
+        assert len(prom.handlers) == 1
+        assert isinstance(prom.handlers[0], logging.StreamHandler)
+        prom.handlers.clear()
+
+    def test_json_format_produces_valid_json(self, capsys) -> None:
+        configure_logging(level=logging.INFO, log_format="json")
+        logger = get_logger("test_json")
+        logger.info("hello", extra={"structured": {"key": "value"}})
+
+        captured = capsys.readouterr()
+        # Output goes to stderr
+        line = captured.err.strip().split("\n")[-1]
+        data = json.loads(line)
+        assert data["message"] == "hello"
+        assert data["structured"]["key"] == "value"
+        assert data["level"] == "INFO"
+        assert "timestamp" in data
+
+        logging.getLogger("prometheus").handlers.clear()
+
+    def test_text_format_includes_structured_extras(self, capsys) -> None:
+        configure_logging(level=logging.INFO, log_format="text")
+        logger = get_logger("test_text")
+        logger.info("msg", extra={"structured": {"foo": "bar"}})
+
+        captured = capsys.readouterr()
+        assert "foo=bar" in captured.err
+
+        logging.getLogger("prometheus").handlers.clear()
+
+    def test_debug_level_shows_debug_messages(self, capsys) -> None:
+        configure_logging(level=logging.DEBUG)
+        logger = get_logger("test_debug")
+        logger.debug("debug msg")
+
+        captured = capsys.readouterr()
+        assert "debug msg" in captured.err
+
+        logging.getLogger("prometheus").handlers.clear()
+
+    def test_warning_level_hides_debug_messages(self, capsys) -> None:
+        configure_logging(level=logging.WARNING)
+        logger = get_logger("test_warn")
+        logger.debug("should not appear")
+        logger.info("also hidden")
+
+        captured = capsys.readouterr()
+        assert "should not appear" not in captured.err
+        assert "also hidden" not in captured.err
+
+        logging.getLogger("prometheus").handlers.clear()
+
+    def test_file_handler_writes_to_file(self, tmp_path: Path) -> None:
+        log_file = tmp_path / "test.log"
+        configure_logging(level=logging.INFO, log_file=str(log_file))
+        logger = get_logger("test_file")
+        logger.info("file message")
+
+        prom = logging.getLogger("prometheus")
+        # Flush handlers
+        for h in prom.handlers:
+            h.flush()
+        prom.handlers.clear()
+
+        content = log_file.read_text()
+        assert "file message" in content
+
+    def test_json_file_output(self, tmp_path: Path) -> None:
+        log_file = tmp_path / "test.json.log"
+        configure_logging(level=logging.INFO, log_format="json", log_file=str(log_file))
+        logger = get_logger("test_json_file")
+        logger.info("json file msg", extra={"structured": {"x": 1}})
+
+        prom = logging.getLogger("prometheus")
+        for h in prom.handlers:
+            h.flush()
+        prom.handlers.clear()
+
+        content = log_file.read_text().strip()
+        data = json.loads(content)
+        assert data["message"] == "json file msg"
+        assert data["structured"]["x"] == 1
+
+    def test_reconfigure_clears_old_handlers(self) -> None:
+        configure_logging(level=logging.INFO)
+        configure_logging(level=logging.DEBUG)
+        prom = logging.getLogger("prometheus")
+        assert len(prom.handlers) == 1
+        prom.handlers.clear()
+
+    def test_propagate_false_prevents_duplicate_output(self, capsys) -> None:
+        configure_logging(level=logging.INFO)
+        prom = logging.getLogger("prometheus")
+        assert prom.propagate is False
+        prom.handlers.clear()
+
+
+class TestGetLogger:
+    def test_returns_child_of_prometheus(self) -> None:
+        logger = get_logger("mymodule")
+        assert logger.name == "prometheus.mymodule"
+
+    def test_inherits_level_from_parent(self) -> None:
+        configure_logging(level=logging.DEBUG)
+        logger = get_logger("child")
+        assert logger.getEffectiveLevel() <= logging.DEBUG
+        logging.getLogger("prometheus").handlers.clear()
+
+
+class TestJsonFormatter:
+    def test_exception_included(self, capsys) -> None:
+        configure_logging(level=logging.ERROR, log_format="json")
+        logger = get_logger("test_exc")
+        try:
+            raise ValueError("boom")
+        except ValueError:
+            logger.error("failed", exc_info=True)
+
+        captured = capsys.readouterr()
+        line = captured.err.strip().split("\n")[-1]
+        data = json.loads(line)
+        assert "ValueError: boom" in data["exception"]
+
+        logging.getLogger("prometheus").handlers.clear()
+
+
+class TestLoggingCLIIntegration:
+    """Tests for CLI flags that configure logging."""
+
+    def test_verbose_flag_enables_info(self, tmp_path: Path) -> None:
+        """Simulate what -v does — configure_logging at INFO level."""
+        configure_logging(level=logging.INFO)
+        logger = get_logger("evolution")
+        logger.info("test message")
+
+        prom = logging.getLogger("prometheus")
+        assert len(prom.handlers) == 1
+        prom.handlers.clear()
+
+    def test_debug_flag_enables_debug(self) -> None:
+        """Simulate what --debug does — configure_logging at DEBUG level."""
+        configure_logging(level=logging.DEBUG)
+        logger = get_logger("evolution")
+        logger.debug("debug message")
+
+        prom = logging.getLogger("prometheus")
+        assert prom.level == logging.DEBUG
+        prom.handlers.clear()
+
+    def test_log_format_invalid_rejected(self) -> None:
+        """Invalid log_format should be caught by OptimizationConfig validator."""
+        from pydantic import ValidationError
+        from prometheus.application.dto import OptimizationConfig
+
+        import pytest
+
+        with pytest.raises(ValidationError, match="log_format must be one of"):
+            OptimizationConfig(
+                seed_prompt="a",
+                task_description="b",
+                log_format="xml",
+            )
+
+    def test_log_format_text_and_json_accepted(self) -> None:
+        """Both text and json log_format values should be valid."""
+        from prometheus.application.dto import OptimizationConfig
+
+        for fmt in ("text", "json"):
+            config = OptimizationConfig(
+                seed_prompt="a", task_description="b", log_format=fmt,
+            )
+            assert config.log_format == fmt