sqlmesh 0.200.0__tar.gz → 0.200.1.dev2__tar.gz

{sqlmesh-0.200.0 → sqlmesh-0.200.1.dev2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sqlmesh
-Version: 0.200.0
+Version: 0.200.1.dev2
 Summary: Next-generation data transformation framework
 Author-email: "TobikoData Inc." <engineering@tobikodata.com>
 License:                                  Apache License

{sqlmesh-0.200.0 → sqlmesh-0.200.1.dev2}/sqlmesh/_version.py

@@ -17,5 +17,5 @@ __version__: str
 __version_tuple__: VERSION_TUPLE
 version_tuple: VERSION_TUPLE
 
-__version__ = version = '0.200.0'
-__version_tuple__ = version_tuple = (0, 200, 0)
+__version__ = version = '0.200.1.dev2'
+__version_tuple__ = version_tuple = (0, 200, 1, 'dev2')

{sqlmesh-0.200.0 → sqlmesh-0.200.1.dev2}/sqlmesh/core/console.py

@@ -912,6 +912,7 @@ class TerminalConsole(Console):
         self.table_diff_model_tasks: t.Dict[str, TaskID] = {}
         self.table_diff_progress_live: t.Optional[Live] = None
 
+        self.signal_progress_logged = False
         self.signal_status_tree: t.Optional[Tree] = None
 
         self.verbosity = verbosity
@@ -956,7 +957,8 @@ class TerminalConsole(Console):
     ) -> None:
         """Indicates that a new snapshot evaluation/auditing progress has begun."""
         # Add a newline to separate signal checking from evaluation
-        self._print("")
+        if self.signal_progress_logged:
+            self._print("")
 
         if not self.evaluation_progress_live:
             self.evaluation_total_progress = make_progress_bar(
@@ -1188,6 +1190,7 @@ class TerminalConsole(Console):
         if self.signal_status_tree is not None:
             self._print(self.signal_status_tree)
             self.signal_status_tree = None
+            self.signal_progress_logged = True
 
     def start_creation_progress(
         self,

{sqlmesh-0.200.0 → sqlmesh-0.200.1.dev2}/sqlmesh/core/scheduler.py

@@ -754,7 +754,7 @@ class Scheduler:
         """
         signals = snapshot.is_model and snapshot.model.render_signal_calls()
 
-        if not signals:
+        if not (signals and signals.signals_to_kwargs):
             return intervals
 
         self.console.start_signal_progress(

sqlmesh-0.200.1.dev2/sqlmesh/migrations/v0085_deterministic_repr.py

@@ -0,0 +1,134 @@
+"""
+When serializing some objects, like `__sqlmesh__vars__`, the order of keys in the dictionary were not deterministic
+and therefore this migration applies deterministic sorting to the keys of the dictionary.
+"""
+
+import json
+import typing as t
+from dataclasses import dataclass
+
+from sqlglot import exp
+
+from sqlmesh.utils.migration import index_text_type, blob_text_type
+
+
+# Make sure `SqlValue` is defined so it can be used by `eval` call in the migration
+@dataclass
+class SqlValue:
+    """A SQL string representing a generated SQLGlot AST."""
+
+    sql: str
+
+
+def _deterministic_repr(obj: t.Any) -> str:
+    """
+    This is a copy of the function from utils.metaprogramming
+    """
+
+    def _normalize_for_repr(o: t.Any) -> t.Any:
+        if isinstance(o, dict):
+            sorted_items = sorted(o.items(), key=lambda x: str(x[0]))
+            return {k: _normalize_for_repr(v) for k, v in sorted_items}
+        if isinstance(o, (list, tuple)):
+            # Recursively normalize nested structures
+            normalized = [_normalize_for_repr(item) for item in o]
+            return type(o)(normalized)
+        return o
+
+    try:
+        return repr(_normalize_for_repr(obj))
+    except Exception:
+        return repr(obj)
+
+
+def migrate(state_sync, **kwargs):  # type: ignore
+    import pandas as pd
+
+    engine_adapter = state_sync.engine_adapter
+    schema = state_sync.schema
+    snapshots_table = "_snapshots"
+    if schema:
+        snapshots_table = f"{schema}.{snapshots_table}"
+
+    migration_needed = False
+    new_snapshots = []
+
+    for (
+        name,
+        identifier,
+        version,
+        snapshot,
+        kind_name,
+        updated_ts,
+        unpaused_ts,
+        ttl_ms,
+        unrestorable,
+    ) in engine_adapter.fetchall(
+        exp.select(
+            "name",
+            "identifier",
+            "version",
+            "snapshot",
+            "kind_name",
+            "updated_ts",
+            "unpaused_ts",
+            "ttl_ms",
+            "unrestorable",
+        ).from_(snapshots_table),
+        quote_identifiers=True,
+    ):
+        parsed_snapshot = json.loads(snapshot)
+        python_env = parsed_snapshot["node"].get("python_env")
+
+        if python_env:
+            for key, executable in python_env.items():
+                if isinstance(executable, dict) and executable.get("kind") == "value":
+                    old_payload = executable["payload"]
+                    try:
+                        # Try to parse the old payload and re-serialize it deterministically
+                        parsed_value = eval(old_payload)
+                        new_payload = _deterministic_repr(parsed_value)
+
+                        # Only update if the representation changed
+                        if old_payload != new_payload:
+                            executable["payload"] = new_payload
+                            migration_needed = True
+                    except Exception:
+                        # If we still can't eval it, leave it as-is
+                        pass
+
+        new_snapshots.append(
+            {
+                "name": name,
+                "identifier": identifier,
+                "version": version,
+                "snapshot": json.dumps(parsed_snapshot),
+                "kind_name": kind_name,
+                "updated_ts": updated_ts,
+                "unpaused_ts": unpaused_ts,
+                "ttl_ms": ttl_ms,
+                "unrestorable": unrestorable,
+            }
+        )
+
+    if migration_needed and new_snapshots:
+        engine_adapter.delete_from(snapshots_table, "TRUE")
+
+        index_type = index_text_type(engine_adapter.dialect)
+        blob_type = blob_text_type(engine_adapter.dialect)
+
+        engine_adapter.insert_append(
+            snapshots_table,
+            pd.DataFrame(new_snapshots),
+            columns_to_types={
+                "name": exp.DataType.build(index_type),
+                "identifier": exp.DataType.build(index_type),
+                "version": exp.DataType.build(index_type),
+                "snapshot": exp.DataType.build(blob_type),
+                "kind_name": exp.DataType.build("text"),
+                "updated_ts": exp.DataType.build("bigint"),
+                "unpaused_ts": exp.DataType.build("bigint"),
+                "ttl_ms": exp.DataType.build("bigint"),
+                "unrestorable": exp.DataType.build("boolean"),
+            },
+        )

{sqlmesh-0.200.0 → sqlmesh-0.200.1.dev2}/sqlmesh/utils/metaprogramming.py

@@ -425,7 +425,9 @@ class Executable(PydanticModel):
 
     @classmethod
     def value(cls, v: t.Any, is_metadata: t.Optional[bool] = None) -> Executable:
-        return Executable(payload=repr(v), kind=ExecutableKind.VALUE, is_metadata=is_metadata)
+        return Executable(
+            payload=_deterministic_repr(v), kind=ExecutableKind.VALUE, is_metadata=is_metadata
+        )
 
 
 def serialize_env(env: t.Dict[str, t.Any], path: Path) -> t.Dict[str, Executable]:
@@ -633,6 +635,38 @@ def print_exception(
     out.write(tb)
 
 
+def _deterministic_repr(obj: t.Any) -> str:
+    """Create a deterministic representation by ensuring consistent ordering before repr().
+
+    For dictionaries, ensures consistent key ordering to prevent non-deterministic
+    serialization that affects fingerprinting. Uses Python's native repr() logic
+    for all formatting to handle edge cases properly.
+
+    Note that this function assumes list/tuple order is significant and therefore does not sort them.
+
+    Args:
+        obj: The object to represent as a string.
+
+    Returns:
+        A deterministic string representation of the object.
+    """
+
+    def _normalize_for_repr(o: t.Any) -> t.Any:
+        if isinstance(o, dict):
+            sorted_items = sorted(o.items(), key=lambda x: str(x[0]))
+            return {k: _normalize_for_repr(v) for k, v in sorted_items}
+        if isinstance(o, (list, tuple)):
+            # Recursively normalize nested structures
+            normalized = [_normalize_for_repr(item) for item in o]
+            return type(o)(normalized)
+        return o
+
+    try:
+        return repr(_normalize_for_repr(obj))
+    except Exception:
+        return repr(obj)
+
+
 def import_python_file(path: Path, relative_base: Path = Path()) -> types.ModuleType:
     relative_path = path.absolute().relative_to(relative_base.absolute())
     module_name = str(relative_path.with_suffix("")).replace(os.path.sep, ".")

{sqlmesh-0.200.0 → sqlmesh-0.200.1.dev2}/sqlmesh.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sqlmesh
-Version: 0.200.0
+Version: 0.200.1.dev2
 Summary: Next-generation data transformation framework
 Author-email: "TobikoData Inc." <engineering@tobikodata.com>
 License:                                  Apache License

{sqlmesh-0.200.0 → sqlmesh-0.200.1.dev2}/sqlmesh.egg-info/SOURCES.txt

@@ -455,6 +455,7 @@ sqlmesh/migrations/v0081_update_partitioned_by.py
 sqlmesh/migrations/v0082_warn_if_incorrectly_duplicated_statements.py
 sqlmesh/migrations/v0083_use_sql_for_scd_time_data_type_data_hash.py
 sqlmesh/migrations/v0084_normalize_quote_when_matched_and_merge_filter.py
+sqlmesh/migrations/v0085_deterministic_repr.py
 sqlmesh/schedulers/__init__.py
 sqlmesh/utils/__init__.py
 sqlmesh/utils/aws.py

{sqlmesh-0.200.0 → sqlmesh-0.200.1.dev2}/tests/cli/test_cli.py

@@ -2175,3 +2175,18 @@ WHERE ds::DATE BETWEEN @start_ds AND @end_ds
 
     # Only one model was executed
     assert "100.0% • 1/1 • 0:00:00" in result.output
+
+    rmtree(tmp_path)
+    tmp_path.mkdir(parents=True, exist_ok=True)
+
+    create_example_project(tmp_path)
+
+    # Example project models have start dates, so there are no date prompts
+    # for the `prod` environment.
+    # Input: `y` to apply and backfill
+    result = runner.invoke(
+        cli, ["--log-file-dir", str(tmp_path), "--paths", str(tmp_path), "plan"], input="y\n"
+    )
+    assert_plan_success(result)
+
+    assert "Checking signals" not in result.output

{sqlmesh-0.200.0 → sqlmesh-0.200.1.dev2}/tests/utils/test_metaprogramming.py

@@ -22,6 +22,7 @@ from sqlmesh.utils.errors import SQLMeshError
 from sqlmesh.utils.metaprogramming import (
     Executable,
     ExecutableKind,
+    _deterministic_repr,
     build_env,
     func_globals,
     normalize_source,
@@ -48,7 +49,7 @@ def test_print_exception(mocker: MockerFixture):
     except Exception as ex:
         print_exception(ex, test_env, out_mock)
 
-    expected_message = r"""  File ".*?.tests.utils.test_metaprogramming\.py", line 47, in test_print_exception
+    expected_message = r"""  File ".*?.tests.utils.test_metaprogramming\.py", line 48, in test_print_exception
     eval\("test_fun\(\)", env\).*
 
   File '/test/path.py' \(or imported file\), line 2, in test_fun
@@ -457,3 +458,162 @@ def test_serialize_env_with_enum_import_appearing_in_two_functions() -> None:
     }
 
     assert serialized_env == expected_env
+
+
+def test_deterministic_repr_basic_types():
+    """Test _deterministic_repr with basic Python types."""
+    # Test basic types that should use standard repr
+    assert _deterministic_repr(42) == "42"
+    assert _deterministic_repr("hello") == "'hello'"
+    assert _deterministic_repr(True) == "True"
+    assert _deterministic_repr(None) == "None"
+    assert _deterministic_repr(3.14) == "3.14"
+
+
+def test_deterministic_repr_dict_ordering():
+    """Test that _deterministic_repr produces consistent output for dicts with different key ordering."""
+    # Same dict with different key ordering
+    dict1 = {"c": 3, "a": 1, "b": 2}
+    dict2 = {"a": 1, "b": 2, "c": 3}
+    dict3 = {"b": 2, "c": 3, "a": 1}
+
+    repr1 = _deterministic_repr(dict1)
+    repr2 = _deterministic_repr(dict2)
+    repr3 = _deterministic_repr(dict3)
+
+    # All should produce the same representation
+    assert repr1 == repr2 == repr3
+    assert repr1 == "{'a': 1, 'b': 2, 'c': 3}"
+
+
+def test_deterministic_repr_mixed_key_types():
+    """Test _deterministic_repr with mixed key types (strings and numbers)."""
+    dict1 = {42: "number", "string": "text", 1: "one"}
+    dict2 = {"string": "text", 1: "one", 42: "number"}
+
+    repr1 = _deterministic_repr(dict1)
+    repr2 = _deterministic_repr(dict2)
+
+    # Should produce consistent ordering despite mixed key types
+    assert repr1 == repr2
+    # Numbers come before strings when sorting by string representation
+    assert repr1 == "{1: 'one', 42: 'number', 'string': 'text'}"
+
+
+def test_deterministic_repr_nested_structures():
+    """Test _deterministic_repr with deeply nested dictionaries."""
+    nested1 = {"outer": {"z": 26, "a": 1}, "list": [3, {"y": 2, "x": 1}], "simple": "value"}
+
+    nested2 = {"simple": "value", "list": [3, {"x": 1, "y": 2}], "outer": {"a": 1, "z": 26}}
+
+    repr1 = _deterministic_repr(nested1)
+    repr2 = _deterministic_repr(nested2)
+
+    assert repr1 == repr2
+    # Verify structure is maintained with sorted keys
+    expected = "{'list': [3, {'x': 1, 'y': 2}], 'outer': {'a': 1, 'z': 26}, 'simple': 'value'}"
+    assert repr1 == expected
+
+
+def test_deterministic_repr_lists_and_tuples():
+    """Test _deterministic_repr preserves order for lists/tuples but sorts nested dicts."""
+    # Lists should maintain their order
+    list_with_dicts = [{"b": 2, "a": 1}, {"d": 4, "c": 3}]
+    list_repr = _deterministic_repr(list_with_dicts)
+    expected_list = "[{'a': 1, 'b': 2}, {'c': 3, 'd': 4}]"
+    assert list_repr == expected_list
+
+    # Tuples should maintain their order
+    tuple_with_dicts = ({"z": 26, "a": 1}, {"y": 25, "b": 2})
+    tuple_repr = _deterministic_repr(tuple_with_dicts)
+    expected_tuple = "({'a': 1, 'z': 26}, {'b': 2, 'y': 25})"
+    assert tuple_repr == expected_tuple
+
+
+def test_deterministic_repr_empty_containers():
+    """Test _deterministic_repr with empty containers."""
+    assert _deterministic_repr({}) == "{}"
+    assert _deterministic_repr([]) == "[]"
+    assert _deterministic_repr(()) == "()"
+
+
+def test_deterministic_repr_special_characters():
+    """Test _deterministic_repr handles special characters correctly."""
+    special_dict = {
+        "quotes": "text with 'single' and \"double\" quotes",
+        "unicode": "unicode: ñáéíóú",
+        "newlines": "text\nwith\nnewlines",
+        "backslashes": "path\\to\\file",
+    }
+
+    result = _deterministic_repr(special_dict)
+
+    # Should be valid Python that can be evaluated
+    reconstructed = eval(result)
+    assert reconstructed == special_dict
+
+    # Should be deterministic - same input produces same output
+    result2 = _deterministic_repr(special_dict)
+    assert result == result2
+
+
+def test_deterministic_repr_executable_integration():
+    """Test that _deterministic_repr works correctly with Executable.value()."""
+    # Test the integration with Executable.value which is the main use case
+    variables1 = {"env": "dev", "debug": True, "timeout": 30}
+    variables2 = {"timeout": 30, "debug": True, "env": "dev"}
+
+    exec1 = Executable.value(variables1)
+    exec2 = Executable.value(variables2)
+
+    # Should produce identical payloads despite different input ordering
+    assert exec1.payload == exec2.payload
+    assert exec1.payload == "{'debug': True, 'env': 'dev', 'timeout': 30}"
+
+    # Should be valid Python
+    reconstructed = eval(exec1.payload)
+    assert reconstructed == variables1
+
+
+def test_deterministic_repr_complex_example():
+    """Test _deterministic_repr with a complex real-world-like structure."""
+    complex_vars = {
+        "database_config": {
+            "host": "localhost",
+            "port": 5432,
+            "credentials": {"username": "admin", "password": "secret"},
+        },
+        "feature_flags": ["flag_b", "flag_a"],
+        "metadata": {
+            "version": "1.0.0",
+            "environment": "production",
+            "tags": {"team": "data", "project": "analytics"},
+        },
+        42: "numeric_key",
+        "arrays": [{"config": {"nested": True, "level": 2}}, {"simple": "value"}],
+    }
+
+    expected_structure = {
+        42: "numeric_key",
+        "arrays": [{"config": {"level": 2, "nested": True}}, {"simple": "value"}],
+        "database_config": {
+            "credentials": {"password": "secret", "username": "admin"},
+            "host": "localhost",
+            "port": 5432,
+        },
+        "feature_flags": ["flag_b", "flag_a"],
+        "metadata": {
+            "environment": "production",
+            "tags": {"project": "analytics", "team": "data"},
+            "version": "1.0.0",
+        },
+    }
+
+    actual_repr = _deterministic_repr(complex_vars)
+    expected_repr = repr(expected_structure)
+    assert actual_repr == expected_repr
+
+    # Should be valid Python
+    reconstructed = eval(actual_repr)
+    assert isinstance(reconstructed, dict)
+    assert reconstructed == complex_vars