Fix _add_pydantic_field_to_parser using cli annotation as parser args

pydantic_cli/__init__.py

@@ -8,12 +8,13 @@
 import typing
 from copy import deepcopy
 from typing import overload
-from typing import Any, Mapping, Callable
+from typing import Any, Mapping, Callable, get_origin
 
 
 import pydantic
-from pydantic import BaseModel, PydanticDeprecatedSince20
+from pydantic import BaseModel, TypeAdapter, PydanticDeprecatedSince20
 from pydantic.fields import FieldInfo
+from pydantic.config import JsonDict
 
 # This is not great. Pydantic >= 3 changing Field will require backward incompatible
 # changes to the API of Field(int, cli=('-m', '--max-records')) and pydantic-cli will
@@ -23,7 +24,6 @@
 
 from ._version import __version__
 
-from .core import M, Tuple1or2Type, Tuple1Type, Tuple2Type
 from .core import EpilogueHandlerType, PrologueHandlerType, ExceptionHandlerType
 from .core import (
     CliConfig,
@@ -91,33 +91,7 @@ def _is_sequence(annotation: Any) -> bool:
     SET_TYPES: list[type] = [set, typing.Set, collections.abc.MutableSet]
     FROZEN_SET_TYPES: list[type] = [frozenset, typing.FrozenSet, collections.abc.Set]
     ALL_SEQ = set(LIST_TYPES + SET_TYPES + FROZEN_SET_TYPES)
-
-    # what is exactly going on here?
-    return getattr(annotation, "__origin__", "NOTFOUND") in ALL_SEQ
-
-
-@pydantic.validate_call
-def __process_tuple(tuple_one_or_two: Tuple1or2Type, long_arg: str) -> Tuple1or2Type:
-    """
-    If the custom args are provided as only short, then
-    add the long version. Or just use the
-    """
-    lx: list[str] = list(tuple_one_or_two)
-
-    nx = len(lx)
-    if nx == 1:
-        if len(lx[0]) == 2:  # xs = '-s'
-            return lx[0], long_arg
-        else:
-            # this is the positional only case
-            return (lx[0],)
-    elif nx == 2:
-        # the explicit form is provided
-        return lx[0], lx[1]
-    else:
-        raise ValueError(
-            f"Unsupported format for `{tuple_one_or_two}` type={type(tuple_one_or_two)}. Expected 1 or 2 tuple."
-        )
+    return get_origin(annotation) in ALL_SEQ
 
 
 def _add_pydantic_field_to_parser(
@@ -151,15 +125,6 @@ def _add_pydantic_field_to_parser(
                                           it's not well-defined or supported. This should be List[T]
     """
 
-    default_long_arg = "".join([long_prefix, field_id])
-    # there's mypy type issues here
-    cli_custom_: Tuple1or2Type = (
-        (default_long_arg,)
-        if field_info.json_schema_extra is None  # type: ignore
-        else field_info.json_schema_extra.get("cli", (default_long_arg,))  # type: ignore
-    )
-    cli_short_long: Tuple1or2Type = __process_tuple(cli_custom_, default_long_arg)
-
     is_required = field_info.is_required()
     default_value = field_info.default
     is_sequence = _is_sequence(field_info.annotation)
@@ -170,25 +135,32 @@ def _add_pydantic_field_to_parser(
         default_value = override_value
         is_required = False
 
+    if cli := isinstance(field_info.json_schema_extra, dict) and field_info.json_schema_extra.get("cli"):
+        # use custom cli annotation if provided
+        args = TypeAdapter(tuple[str, ...]).validate_python(cli)
+    else:
+        # positional if required, else named optional
+        args = () if is_required else (f"{long_prefix}{field_id}",)
+
     # Delete cli and json_schema_extras metadata isn't in FieldInfo and won't be displayed
     # Not sure if this is the correct, or expected behavior.
     cfield_info = deepcopy(field_info)
     cfield_info.json_schema_extra = None
     # write this to keep backward compat with 3.10
-    help_ = "".join(["Field(", cfield_info.__repr_str__(", "), ")"])
-
-    # log.debug(f"Creating Argument Field={field_id} opts:{cli_short_long}, allow_none={field.allow_none} default={default_value} type={field.type_} required={is_required} dest={field_id} desc={description}")
-
-    # MK. I don't think there's any point trying to fight with argparse to get
-    # the types correct here. It's just a mess from a type standpoint.
-    shape_kw = {"nargs": "+"} if is_sequence else {}
+    help_ = field_info.description or "".join(["Field(", field_info.__repr_str__(", "), ")"])
+
+    kwargs: dict[str, Any] = {}
+    if is_sequence:
+        kwargs["nargs"] = "+"
+    if len(args) > 0:
+        # only provide required to non-positional options
+        kwargs["required"] = is_required
     parser.add_argument(
-        *cli_short_long,
+        *args,
         help=help_,
         default=default_value,
         dest=field_id,
-        required=is_required,
-        **shape_kw,  # type: ignore
+        **kwargs,
     )
 
     return parser

pydantic_cli/examples/simple.py

@@ -5,7 +5,7 @@
 and can be called
 
 ```bash
-my-tool --input_file file.fasta --max_records 10
+my-tool file.fasta 10
 ```
 """
 

pydantic_cli/examples/simple_with_enum.py

@@ -23,8 +23,8 @@ class State(str, Enum):
 
 
 class Options(Cmd):
-    states: Set[State]
     mode: Mode
+    states: Set[State]  # sequence positionals must be after other positionals
     max_records: int = 100
 
     def run(self) -> None:

pydantic_cli/examples/simple_with_json_config_not_found.py

@@ -1,6 +1,8 @@
 import sys
 import logging
 
+from pydantic import Field
+
 from pydantic_cli import run_and_exit, CliConfig, Cmd
 
 log = logging.getLogger(__name__)
@@ -19,7 +21,7 @@ class Options(Cmd):
         cli_json_validate_path=False,
     )
 
-    input_file: str
+    input_file: str = Field(cli=("--input_file",))
     max_records: int = 10
 
     def run(self) -> None:

pydantic_cli/examples/simple_with_list.py

@@ -9,13 +9,14 @@
 ```
 """
 
+from pydantic import Field
 from pydantic_cli import run_and_exit, Cmd
 
 
 class Options(Cmd):
-    input_file: list[str]
-    filters: set[str]
-    max_records: int
+    input_file: list[str] = Field(cli=("--input_file",))
+    filters: set[str] = Field(cli=("--filters",))
+    max_records: int = Field(cli=("--max_records",))
 
     def run(self) -> None:
         print(f"Mock example running with {self}")

pydantic_cli/tests/test_examples_simple.py

@@ -8,4 +8,4 @@ class TestExamples(_TestHarness[Options]):
     CONFIG = HarnessConfig(Options)
 
     def test_simple_01(self):
-        self.run_config(["--input_file", "/path/to/file.txt", "--max_record", "1234"])
+        self.run_config(["/path/to/file.txt", "1234"])

pydantic_cli/tests/test_examples_simple_boolean_and_config.py

@@ -7,14 +7,14 @@ class TestExamples(_TestHarness[Options]):
     CONFIG = HarnessConfig(Options)
 
     def test_simple_01(self):
-        self.run_config(["--input_file", "/path/to/file.txt"])
+        self.run_config(["/path/to/file.txt"])
 
     def test_simple_02(self):
         self.run_config(
-            ["--input_file", "/path/to/file.txt", "--run-training", "false"]
+            ["/path/to/file.txt", "--run-training", "false"]
         )
 
     def test_simple_03(self):
         self.run_config(
-            ["--input_file", "/path/to/file.txt", "-r", "false", "--dry-run", "false"]
+            ["/path/to/file.txt", "-r", "false", "--dry-run", "false"]
         )

pydantic_cli/tests/test_examples_simple_required.py

@@ -0,0 +1,10 @@
+from . import _TestHarness, HarnessConfig
+
+from pydantic_cli.examples.simple import Options
+
+class TestExamples(_TestHarness[Options]):
+
+    CONFIG = HarnessConfig(Options)
+
+    def test_simple_01(self):
+        self.run_config(["/path/to/file.txt", "1234"])

pydantic_cli/tests/test_examples_simple_schema.py

@@ -7,12 +7,12 @@ class TestExamples(_TestHarness):
     CONFIG = HarnessConfig(Options)
 
     def test_01(self):
-        args = "-f /path/to/file.txt --max_records 1234 -s 1.234 --max-filter-score 10.234 -n none"
+        args = "-f /path/to/file.txt -m 1234 -s 1.234 --max-filter-score 10.234 -n none"
         self.run_config(args.split())
 
     def test_02(self):
         args = "-f /path/to/file.txt -m 1234 -s 1.234 -S 10.234 --filter-name alphax"
         self.run_config(args.split())
 
     def test_03(self):
-        self.run_config(["-f", "/path/to/file.txt", "-s", "1.234", "-n", "beta.v2"])
+        self.run_config(["-f", "/path/to/file.txt", "-m", "1234", "-s", "1.234", "-n", "beta.v2"])

pydantic_cli/tests/test_examples_simple_with_boolean.py

@@ -7,15 +7,14 @@ class TestExamples(_TestHarness[Options]):
     CONFIG = HarnessConfig(Options)
 
     def test_simple_01(self):
-        self.run_config(["--input_file", "/path/to/file.txt"])
+        self.run_config(["/path/to/file.txt"])
 
     def test_simple_02(self):
-        self.run_config(["--input_file", "/path/to/file.txt", "--run_training", "y"])
+        self.run_config(["/path/to/file.txt", "--run_training", "y"])
 
     def test_simple_03(self):
         self.run_config(
             [
-                "--input_file",
                 "/path/to/file.txt",
                 "--run_training",
                 "true",

pydantic_cli/tests/test_examples_simple_with_boolean_custom.py

@@ -9,28 +9,20 @@ class TestExamples(_TestHarness[Options]):
     def test_simple_01(self):
         self.run_config(
             [
-                "--input_file",
                 "/path/to/file.txt",
-                "--input_file2",
                 "/path/2.txt",
                 "-f",
                 "/path/to/file.h5",
-                "--report_json",
                 "output.json",
                 "--fasta",
                 "output.fasta",
-                "--gamma",
                 "true",
-                "--alpha",
                 "true",
-                "--zeta_mode",
                 "true",
                 "--epsilon",
                 "true",
-                "--states",
                 "RUNNING",
                 "FAILED",
-                "--filter_mode",
                 "1",
             ]
         )

pydantic_cli/tests/test_examples_simple_with_enum.py

@@ -8,17 +8,15 @@ class TestExamples(_TestHarness[Options]):
     CONFIG = HarnessConfig(Options)
 
     def test_simple_01(self):
-        args = ["--states", "RUNNING", "FAILED", "--max_records", "1234", "--mode", "1"]
+        args = ["--max_records", "1234", "1", "RUNNING", "FAILED"]
         self.run_config(args)
 
     def test_bad_enum_value(self):
         args = [
-            "--states",
-            "RUNNING",
-            "BAD_STATE",
             "--max_records",
             "1234",
-            "--mode",
             "1",
+            "RUNNING",
+            "BAD_STATE",
         ]
         self.run_config(args, exit_code=1)

pydantic_cli/tests/test_examples_simple_with_enum_by_name.py

@@ -8,18 +8,18 @@ class TestExamples(_TestHarness[Options]):
     CONFIG = HarnessConfig(Options)
 
     def test_simple_01(self):
-        args = ["--states", "RUNNING", "FAILED", "--mode", "alpha"]
+        args = ["--states", "RUNNING", "FAILED", "--", "alpha"]
         self.run_config(args)
 
     def test_case_insensitive(self):
-        args = ["--states", "successful", "failed", "--mode", "ALPHA"]
+        args = ["--states", "successful", "failed", "--", "ALPHA"]
         self.run_config(args)
 
     def test_bad_enum_by_value(self):
         args = [
             "--states",
             "RUNNING",
-            "--mode",
+            "--",
             "1",
         ]
         self.run_config(args, exit_code=1)
@@ -28,7 +28,7 @@ def test_bad_enum_value(self):
         args = [
             "--states",
             "RUNNING",
-            "--mode",
+            "--",
             "DRAGON",
         ]
         self.run_config(args, exit_code=1)

pydantic_cli/tests/test_examples_simple_with_list.py

@@ -12,7 +12,7 @@ def test_simple_01(self):
             "--input_file",
             "/path/to/file.txt",
             "/and/another/file.txt",
-            "--max_record",
+            "--max_records",
             "1234",
             "--filters",
             "alpha",

pydantic_cli/tests/test_examples_with_json_config.py

@@ -9,9 +9,9 @@ class TestExample(_TestHarness[Opts]):
 
     CONFIG = HarnessConfig(Opts)
 
-    def _util(self, d, more_args):
+    def _run_cmd_with_config(self, config, more_args):
         with NamedTemporaryFile(mode="w", delete=True) as f:
-            json.dump(d, f)
+            json.dump(config, f)
             f.flush()
             args = ["--json-training", str(f.name)] + more_args
             self.run_config(args)
@@ -24,9 +24,9 @@ def test_simple_json(self):
             alpha=1.234,
             beta=9.854,
         )
-        self._util(opt.model_dump(), [])
+        self._run_cmd_with_config(opt.model_dump(), [])
 
     def test_simple_partial_json(self):
-        d = dict(max_records=12, min_filter_score=1.024, alpha=1.234, beta=9.854)
+        config = dict(max_records=12, min_filter_score=1.024, alpha=1.234, beta=9.854)
 
-        self._util(d, ["--hdf_file", "/path/to/file.hdf5"])
+        self._run_cmd_with_config(config, ["/path/to/file.hdf5"])