Add type checking to Continuous Integration for Snowfakery (#740)

* Type checking and some CI upgrades

.coveragerc

@@ -11,3 +11,5 @@ exclude_lines =
     raise AssertionError
     def __repr__
     def __str__
+    if T.TYPE_CHECKING
+    if TYPE_CHECKING

.github/workflows/ci.yml

@@ -36,7 +36,8 @@ jobs:
         run: |
           python setup.py install
           snowbench snowfakery/tools/benchmark_1.yml --num-records 10_000  --num-records-tablename Account --number-of-processes 4
-          snowbench
+          snowbench | tee bench.txt
+          tail -n 7 bench.txt >> $GITHUB_STEP_SUMMARY
 
   with_cci:
     name: With CumulusCI
@@ -69,6 +70,9 @@ jobs:
           diff-cover coverage.xml --fail-under 100 --compare-branch=_remote_main_ --diff-range-notation=.. --show-uncovered --markdown-report coverage.md
           cat coverage.md >> $GITHUB_STEP_SUMMARY
 
+      - name: Type Check
+        run: pyright
+
   faker_docs:
     name: Faker Docs
     runs-on: sfdc-ubuntu-latest

pyproject.toml

@@ -0,0 +1,14 @@
+[tool.pyright]
+include = [
+    "snowfakery/utils/*",
+    "snowfakery/tools/*", 
+    "snowfakery/__init__.py",
+    "snowfakery/__main__.py",
+    "snowfakery/cci_mapping_files/post_processes.py",
+    "snowfakery/data_gen_exceptions.py",
+    "snowfakery/salesforce.py",
+    "snowfakery/standard_plugins/SnowfakeryVersion.py",
+    "snowfakery/standard_plugins/__init__.py",
+    "snowfakery/standard_plugins/base64.py",
+    "snowfakery/object_rows.py",
+    ]

requirements/dev.in

@@ -6,9 +6,10 @@ diff-cover
 mkdocs<1.3.0      # need to change Snowfakery monkey-patching before upgrade
 mkdocs-exclude-search
 pre-commit
+pyright
 pytest
 pytest-cov
-typeguard==2.10.0   # do not upgrade until #181 is fixed
+typeguard
 faker-microservice
 tox
 tox-gh-actions      # needed for CI only

requirements/dev.txt

@@ -27,7 +27,7 @@ click==8.1.3
     #   -r requirements/prod.txt
     #   black
     #   mkdocs
-coverage[toml]==6.4.4
+coverage[toml]==6.5.0
     # via
     #   -r requirements/dev.in
     #   coveralls
@@ -58,14 +58,14 @@ greenlet==1.1.3
     #   sqlalchemy
 gvgen==1.0
     # via -r requirements/prod.txt
-identify==2.5.5
+identify==2.5.6
     # via pre-commit
 idna==3.4
     # via
     #   -r requirements/prod.txt
     #   requests
     #   yarl
-importlib-metadata==4.12.0
+importlib-metadata==5.0.0
     # via
     #   markdown
     #   mkdocs
@@ -99,7 +99,9 @@ multidict==6.0.2
 mypy-extensions==0.4.3
     # via black
 nodeenv==1.7.0
-    # via pre-commit
+    # via
+    #   pre-commit
+    #   pyright
 packaging==21.3
     # via
     #   mkdocs
@@ -128,14 +130,16 @@ pygments==2.13.0
     # via diff-cover
 pyparsing==3.0.9
     # via packaging
+pyright==1.1.274
+    # via -r requirements/dev.in
 pyrsistent==0.18.1
     # via jsonschema
 pytest==7.1.3
     # via
     #   -r requirements/dev.in
     #   pytest-cov
     #   pytest-vcr
-pytest-cov==3.0.0
+pytest-cov==4.0.0
     # via -r requirements/dev.in
 pytest-vcr==1.0.2
     # via -r requirements/dev.in
@@ -184,7 +188,7 @@ tox==3.26.0
     #   tox-gh-actions
 tox-gh-actions==2.10.0
     # via -r requirements/dev.in
-typeguard==2.10.0
+typeguard==2.13.3
     # via -r requirements/dev.in
 typing-extensions==4.3.0
     # via
@@ -216,5 +220,5 @@ zipp==3.8.1
     #   importlib-resources
 
 # The following packages are considered to be unsafe in a requirements file:
-setuptools==65.4.0
+setuptools==65.4.1
     # via nodeenv

snowfakery/backports/typeguard_context_manager_hack.py

@@ -0,0 +1,24 @@
+import contextlib
+from typing import Callable, ContextManager, Iterator, TypeVar
+
+T = TypeVar("T")
+
+# until https://github.com/agronholm/typeguard/issues/115 is fully fixed
+# or typeguard is removed, we need this hack.
+
+# The fix is on typeguard/master and should be merged in Typeguard 3.0
+#
+# We need this feature to be moved from UNRELEASED to RELEASED:
+#
+#    Changed the import hook to append @typechecked to the
+#    decorator list instead of inserting it as the first decorator
+#   (fixes type checking inconsistencies with mypy regarding at least
+#    @contextmanager, probably others too)
+
+
+def contextmanager(
+    func: Callable[..., Iterator[T]]
+) -> Callable[..., ContextManager[T]]:
+    result = contextlib.contextmanager(func)
+    result.__annotations__ = {**func.__annotations__, "return": ContextManager[T]}
+    return result

snowfakery/fakedata/fake_data_generator.py

@@ -164,8 +164,8 @@ class FakeData:
     def __init__(
         self,
         faker_providers: T.Sequence[object],
-        locale: str = None,
-        faker_context: PluginContext = None,
+        locale: T.Optional[str] = None,
+        faker_context: T.Optional[PluginContext] = None,
     ):
         # access to persistent state
         self.faker_context = faker_context

snowfakery/object_rows.py

@@ -1,11 +1,15 @@
 from enum import Enum, auto
+import typing as T
 
 import yaml
 import snowfakery  # noqa
 from .utils.yaml_utils import SnowfakeryDumper
 from contextvars import ContextVar
 
-IdManager = "snowfakery.data_generator_runtime.IdManager"
+if T.TYPE_CHECKING:
+    from snowfakery.data_generator_runtime import IdManager
+else:
+    IdManager = "snowfakery.data_generator_runtime.IdManager"
 RowHistoryCV = ContextVar("RowHistory")
 
 
@@ -105,7 +109,7 @@ class NicknameSlot(ObjectReference):
 
     _tablename: str
     id_manager: IdManager
-    allocated_id: int = None
+    allocated_id: T.Union[T.Optional[int], SlotState] = None
 
     def __init__(self, tablename: str, id_manager: IdManager):
         self._tablename = tablename

snowfakery/tools/storage_bench.py

@@ -1,69 +1,4 @@
 """ Benchmarking tool for the RowHistory backing database"""
 
-from snowfakery.row_history import RowHistory
-from snowfakery.object_rows import RowHistoryCV
-from time import time
-
-COUNT = 200_000
-
-
-def main(count):
-    rh: RowHistory = RowHistory()
-    RowHistoryCV.set(rh)
-    start = time()
-    for i in range(0, COUNT, 4):
-        rh.save_row(
-            "table",
-            None,
-            {
-                "id": i,
-                "blah": "blah",
-            },
-        )
-
-        rh.save_row(
-            "table",
-            "Nicknamed",
-            {
-                "id": i + 1,
-                "blah": "blah",
-            },
-        )
-
-        rh.save_row(
-            "table",
-            "Nicknamed2",
-            {
-                "id": i + 2,
-                "blah": "blah",
-            },
-        )
-
-        if i % 1000 == 0:
-            nickname = "Sparse"
-        else:
-            nickname = "Unused"
-
-        rh.save_row(
-            "table",
-            nickname,
-            {
-                "id": i + 3,
-                "blah": "blah",
-            },
-        )
-
-    end = time()
-    print("Saved", end - start)
-    start = time()
-    for i in range(0, COUNT):
-        rh.load_row("table", i)
-        assert rh.random_row_reference("table", "current-iteration", False).blah
-        assert rh.random_row_reference("Nicknamed", "current-iteration", False).blah
-        assert rh.random_row_reference("Nicknamed2", "current-iteration", False).blah
-        assert rh.random_row_reference("Sparse", "current-iteration", False).blah
-    end = time()
-    print("Loaded", end - start)
-
-
-main(COUNT)
+# Deleted after 2f589a4a44f7ff8e9a4650e978d24839d47f61d9
+# Go back to that PR if you want that code back.

snowfakery/utils/files.py

@@ -1,7 +1,12 @@
 import typing as T
-from contextlib import contextmanager
 from pathlib import Path
 
+if T.TYPE_CHECKING:
+    from contextlib import contextmanager
+else:
+    # until https://github.com/agronholm/typeguard/issues/115 is fixed in typeguard 3
+    from snowfakery.backports.typeguard_context_manager_hack import contextmanager
+
 from click.utils import LazyFile
 
 OpenFileLike = T.Union[T.TextIO, LazyFile]
@@ -11,7 +16,7 @@
 @contextmanager
 def open_file_like(
     file_like: T.Optional[FileLike], mode, **kwargs
-) -> T.ContextManager[T.Tuple[str, OpenFileLike]]:
+) -> T.Generator[T.Tuple[T.Optional[Path], T.Optional[OpenFileLike]], None, None]:
     """Look at a file-like or path-like object and open it
 
     Returns a) the path it was given OR a best-guess of the path and
@@ -30,7 +35,7 @@ def open_file_like(
             yield file_like, f
 
     elif hasattr(file_like, "name"):
-        yield file_like.name, file_like
+        yield Path(file_like.name), file_like  # type: ignore
 
     elif hasattr(file_like, "read"):
-        yield None, file_like
+        yield None, T.cast(OpenFileLike, file_like)

snowfakery/utils/randomized_range.py

@@ -4,7 +4,7 @@
 
 
 class UpdatableRandomRange:
-    def __init__(self, start: int, stop: int = None):
+    def __init__(self, start: int, stop: int):
         assert stop > start
         self.min = start
         self._set_new_range_immediately(start, stop)

snowfakery/utils/template_utils.py

@@ -1,4 +1,4 @@
-from typing import Sequence
+import typing as T
 import string
 from snowfakery.fakedata.fake_data_generator import FakeData
 
@@ -46,9 +46,9 @@ class FakerTemplateLibrary:
 
     def __init__(
         self,
-        faker_providers: Sequence[object],
-        locale: str = None,
-        context: PluginContext = None,
+        faker_providers: T.Sequence[object],
+        locale: T.Optional[str] = None,
+        context: T.Optional[PluginContext] = None,
     ):
         self.locale = locale
         self.context = context