Add completions, Add suggestions list to Argument

examples/argument_examples.py

@@ -21,11 +21,13 @@ async def test_args(
     service: str,
     place: Place = Place.NEW_YORK,
     region: str = "us-east-1",
+    tag: str | None = None,
     verbose: bool | None = None,
+    number: int | None = None,
 ) -> str:
     if verbose:
-        print(f"Deploying {service} to {region} at {place}...")
+        print(f"Deploying {service}:{tag}:{number} to {region} at {place}...")
-    return f"{service} deployed to {region} at {place}"
+    return f"{service}:{tag}:{number} deployed to {region} at {place}"
 
 
 def default_config(parser: CommandArgumentParser) -> None:
@@ -55,6 +57,17 @@ def default_config(parser: CommandArgumentParser) -> None:
         action="store_bool_optional",
         help="Enable verbose output.",
     )
+    parser.add_argument(
+        "--tag",
+        type=str,
+        help="Optional tag for the deployment.",
+        suggestions=["latest", "stable", "beta"],
+    )
+    parser.add_argument(
+        "--number",
+        type=int,
+        help="Optional number argument.",
+    )
 
 
 flx = Falyx("Argument Examples")

falyx/falyx.py

@@ -507,7 +507,6 @@ class Falyx:
                 message=self.prompt,
                 multiline=False,
                 completer=self._get_completer(),
-                reserve_space_for_menu=5,
                 validator=CommandValidator(self, self._get_validator_error_message()),
                 bottom_toolbar=self._get_bottom_bar_render(),
                 key_bindings=self.key_bindings,

falyx/parser/argument.py

@@ -26,6 +26,7 @@ class Argument:
         resolver (BaseAction | None):
             An action object that resolves the argument, if applicable.
         lazy_resolver (bool): True if the resolver should be called lazily, False otherwise
+        suggestions (list[str] | None): A list of suggestions for the argument.
     """
 
     flags: tuple[str, ...]
@@ -40,6 +41,7 @@ class Argument:
     positional: bool = False
     resolver: BaseAction | None = None
     lazy_resolver: bool = False
+    suggestions: list[str] | None = None
 
     def get_positional_text(self) -> str:
         """Get the positional text for the argument."""

falyx/parser/command_argument_parser.py

@@ -4,8 +4,8 @@ from __future__ import annotations
 
 from collections import defaultdict
 from copy import deepcopy
-from difflib import get_close_matches
+from dataclasses import dataclass
-from typing import Any, Iterable
+from typing import Any, Iterable, Sequence
 
 from rich.console import Console
 from rich.markup import escape
@@ -20,6 +20,12 @@ from falyx.parser.utils import coerce_value
 from falyx.signals import HelpSignal
 
 
+@dataclass
+class ArgumentState:
+    arg: Argument
+    consumed: bool = False
+
+
 class CommandArgumentParser:
     """
     Custom argument parser for Falyx Commands.
@@ -65,6 +71,8 @@ class CommandArgumentParser:
         self._flag_map: dict[str, Argument] = {}
         self._dest_set: set[str] = set()
         self._add_help()
+        self._last_positional_states: dict[str, ArgumentState] = {}
+        self._last_keyword_states: dict[str, ArgumentState] = {}
 
     def _add_help(self):
         """Add help argument to the parser."""
@@ -360,19 +368,19 @@ class CommandArgumentParser:
         )
 
         self._register_argument(argument)
-        self._register_argument(negated_argument)
+        self._register_argument(negated_argument, bypass_validation=True)
 
-    def _register_argument(self, argument: Argument):
+    def _register_argument(
+        self, argument: Argument, bypass_validation: bool = False
+    ) -> None:
 
         for flag in argument.flags:
-            if (
+            if flag in self._flag_map and not bypass_validation:
-                flag in self._flag_map
-                and not argument.action == ArgumentAction.STORE_BOOL_OPTIONAL
-            ):
                 existing = self._flag_map[flag]
                 raise CommandArgumentError(
                     f"Flag '{flag}' is already used by argument '{existing.dest}'"
                 )
+
         for flag in argument.flags:
             self._flag_map[flag] = argument
             if not argument.positional:
@@ -397,6 +405,7 @@ class CommandArgumentParser:
         dest: str | None = None,
         resolver: BaseAction | None = None,
         lazy_resolver: bool = True,
+        suggestions: list[str] | None = None,
     ) -> None:
         """Add an argument to the parser.
         For `ArgumentAction.ACTION`, `nargs` and `type` determine how many and what kind
@@ -416,6 +425,8 @@ class CommandArgumentParser:
             help: A brief description of the argument.
             dest: The name of the attribute to be added to the object returned by parse_args().
             resolver: A BaseAction called with optional nargs specified parsed arguments.
+            lazy_resolver: If True, the resolver is called lazily when the argument is accessed.
+            suggestions: A list of suggestions for the argument.
         """
         expected_type = type
         self._validate_flags(flags)
@@ -446,6 +457,10 @@ class CommandArgumentParser:
                 f"Default value '{default}' not in allowed choices: {choices}"
             )
         required = self._determine_required(required, positional, nargs, action)
+        if not isinstance(suggestions, Sequence) and suggestions is not None:
+            raise CommandArgumentError(
+                f"suggestions must be a list or None, got {type(suggestions)}"
+            )
         if not isinstance(lazy_resolver, bool):
             raise CommandArgumentError(
                 f"lazy_resolver must be a boolean, got {type(lazy_resolver)}"
@@ -466,6 +481,7 @@ class CommandArgumentParser:
                 positional=positional,
                 resolver=resolver,
                 lazy_resolver=lazy_resolver,
+                suggestions=suggestions,
             )
             self._register_argument(argument)
 
@@ -491,6 +507,27 @@ class CommandArgumentParser:
             )
         return defs
 
+    def raise_remaining_args_error(
+        self, token: str, arg_states: dict[str, ArgumentState]
+    ) -> None:
+        consumed_dests = [
+            state.arg.dest for state in arg_states.values() if state.consumed
+        ]
+        remaining_flags = [
+            flag
+            for flag, arg in self._keyword.items()
+            if arg.dest not in consumed_dests and flag.startswith(token)
+        ]
+
+        if remaining_flags:
+            raise CommandArgumentError(
+                f"Unrecognized option '{token}'. Did you mean one of: {', '.join(remaining_flags)}?"
+            )
+        else:
+            raise CommandArgumentError(
+                f"Unrecognized option '{token}'. Use --help to see available options."
+            )
+
     def _consume_nargs(
         self, args: list[str], start: int, spec: Argument
     ) -> tuple[list[str], int]:
@@ -536,6 +573,7 @@ class CommandArgumentParser:
         result: dict[str, Any],
         positional_args: list[Argument],
         consumed_positional_indicies: set[int],
+        arg_states: dict[str, ArgumentState],
         from_validate: bool = False,
     ) -> int:
         remaining_positional_args = [
@@ -581,17 +619,7 @@ class CommandArgumentParser:
             except Exception as error:
                 if len(args[i - new_i :]) == 1 and args[i - new_i].startswith("-"):
                     token = args[i - new_i]
-                    valid_flags = [
+                    self.raise_remaining_args_error(token, arg_states)
-                        flag for flag in self._flag_map if flag.startswith(token)
-                    ]
-                    if valid_flags:
-                        raise CommandArgumentError(
-                            f"Unrecognized option '{token}'. Did you mean one of: {', '.join(valid_flags)}?"
-                        ) from error
-                    else:
-                        raise CommandArgumentError(
-                            f"Unrecognized option '{token}'. Use --help to see available options."
-                        ) from error
                 else:
                     raise CommandArgumentError(
                         f"Invalid value for '{spec.dest}': {error}"
@@ -607,6 +635,7 @@ class CommandArgumentParser:
                         raise CommandArgumentError(
                             f"[{spec.dest}] Action failed: {error}"
                         ) from error
+                arg_states[spec.dest].consumed = True
             elif not typed and spec.default:
                 result[spec.dest] = spec.default
             elif spec.action == ArgumentAction.APPEND:
@@ -619,8 +648,10 @@ class CommandArgumentParser:
                 assert result.get(spec.dest) is not None, "dest should not be None"
                 result[spec.dest].extend(typed)
             elif spec.nargs in (None, 1, "?"):
+                arg_states[spec.dest].consumed = True
                 result[spec.dest] = typed[0] if len(typed) == 1 else typed
             else:
+                arg_states[spec.dest].consumed = True
                 result[spec.dest] = typed
 
             if spec.nargs not in ("*", "+"):
@@ -629,15 +660,7 @@ class CommandArgumentParser:
         if i < len(args):
             if len(args[i:]) == 1 and args[i].startswith("-"):
                 token = args[i]
-                valid_flags = [flag for flag in self._flag_map if flag.startswith(token)]
+                self.raise_remaining_args_error(token, arg_states)
-                if valid_flags:
-                    raise CommandArgumentError(
-                        f"Unrecognized option '{token}'. Did you mean one of: {', '.join(valid_flags)}?"
-                    )
-                else:
-                    raise CommandArgumentError(
-                        f"Unrecognized option '{token}'. Use --help to see available options."
-                    )
             else:
                 plural = "s" if len(args[i:]) > 1 else ""
                 raise CommandArgumentError(
@@ -671,6 +694,7 @@ class CommandArgumentParser:
         positional_args: list[Argument],
         consumed_positional_indices: set[int],
         consumed_indices: set[int],
+        arg_states: dict[str, ArgumentState],
         from_validate: bool = False,
     ) -> int:
         if token in self._keyword:
@@ -680,6 +704,7 @@ class CommandArgumentParser:
             if action == ArgumentAction.HELP:
                 if not from_validate:
                     self.render_help()
+                arg_states[spec.dest].consumed = True
                 raise HelpSignal()
             elif action == ArgumentAction.ACTION:
                 assert isinstance(
@@ -692,24 +717,29 @@ class CommandArgumentParser:
                     raise CommandArgumentError(
                         f"Invalid value for '{spec.dest}': {error}"
                     ) from error
-                try:
+                if not spec.lazy_resolver or not from_validate:
-                    result[spec.dest] = await spec.resolver(*typed_values)
+                    try:
-                except Exception as error:
+                        result[spec.dest] = await spec.resolver(*typed_values)
-                    raise CommandArgumentError(
+                    except Exception as error:
-                        f"[{spec.dest}] Action failed: {error}"
+                        raise CommandArgumentError(
-                    ) from error
+                            f"[{spec.dest}] Action failed: {error}"
+                        ) from error
+                arg_states[spec.dest].consumed = True
                 consumed_indices.update(range(i, new_i))
                 i = new_i
             elif action == ArgumentAction.STORE_TRUE:
                 result[spec.dest] = True
+                arg_states[spec.dest].consumed = True
                 consumed_indices.add(i)
                 i += 1
             elif action == ArgumentAction.STORE_FALSE:
                 result[spec.dest] = False
+                arg_states[spec.dest].consumed = True
                 consumed_indices.add(i)
                 i += 1
             elif action == ArgumentAction.STORE_BOOL_OPTIONAL:
                 result[spec.dest] = spec.type(True)
+                arg_states[spec.dest].consumed = True
                 consumed_indices.add(i)
                 i += 1
             elif action == ArgumentAction.COUNT:
@@ -779,19 +809,11 @@ class CommandArgumentParser:
                     )
                 else:
                     result[spec.dest] = typed_values
+                arg_states[spec.dest].consumed = True
                 consumed_indices.update(range(i, new_i))
                 i = new_i
         elif token.startswith("-"):
-            # Handle unrecognized option
+            self.raise_remaining_args_error(token, arg_states)
-            valid_flags = [flag for flag in self._flag_map if flag.startswith(token)]
-            if valid_flags:
-                raise CommandArgumentError(
-                    f"Unrecognized option '{token}'. Did you mean one of: {', '.join(valid_flags)}?"
-                )
-            else:
-                raise CommandArgumentError(
-                    f"Unrecognized option '{token}'. Use --help to see available options."
-                )
         else:
             # Get the next flagged argument index if it exists
             next_flagged_index = -1
@@ -806,6 +828,7 @@ class CommandArgumentParser:
                 result,
                 positional_args,
                 consumed_positional_indices,
+                arg_states=arg_states,
                 from_validate=from_validate,
             )
             i += args_consumed
@@ -818,6 +841,14 @@ class CommandArgumentParser:
         if args is None:
             args = []
 
+        arg_states = {arg.dest: ArgumentState(arg) for arg in self._arguments}
+        self._last_positional_states = {
+            arg.dest: arg_states[arg.dest] for arg in self._positional.values()
+        }
+        self._last_keyword_states = {
+            arg.dest: arg_states[arg.dest] for arg in self._keyword_list
+        }
+
         result = {arg.dest: deepcopy(arg.default) for arg in self._arguments}
         positional_args: list[Argument] = [
             arg for arg in self._arguments if arg.positional
@@ -839,6 +870,7 @@ class CommandArgumentParser:
                 positional_args,
                 consumed_positional_indices,
                 consumed_indices,
+                arg_states=arg_states,
                 from_validate=from_validate,
             )
 
@@ -863,6 +895,7 @@ class CommandArgumentParser:
                 )
 
             if spec.choices and result.get(spec.dest) not in spec.choices:
+                arg_states[spec.dest].consumed = False
                 raise CommandArgumentError(
                     f"Invalid value for '{spec.dest}': must be one of {{{', '.join(spec.choices)}}}"
                 )
@@ -924,72 +957,79 @@ class CommandArgumentParser:
         Returns:
             A list of possible completions based on the current input.
         """
-        consumed_positionals = []
+
-        positional_choices = [
+        # Case 1: Next positional argument
-            str(choice)
+        next_non_consumed_positional: Argument | None = None
-            for arg in self._positional.values()
+        for state in self._last_positional_states.values():
-            for choice in arg.choices
+            if not state.consumed:
-            if arg.choices
+                next_non_consumed_positional = state.arg
-        ]
+                break
-        if not args:
+        if next_non_consumed_positional:
-            # Nothing entered yet: suggest all top-level flags and positionals
+            if next_non_consumed_positional.choices:
-            if positional_choices:
+                return sorted(
-                return sorted(set(positional_choices))
+                    (str(choice) for choice in next_non_consumed_positional.choices)
-            return sorted(
-                set(
-                    flag
-                    for arg in self._arguments
-                    for flag in arg.flags
-                    if not arg.positional
                 )
-            )
+            if next_non_consumed_positional.suggestions:
+                return sorted(next_non_consumed_positional.suggestions)
+
+        consumed_dests = [
+            state.arg.dest
+            for state in self._last_keyword_states.values()
+            if state.consumed
+        ]
+
+        remaining_flags = [
+            flag for flag, arg in self._keyword.items() if arg.dest not in consumed_dests
+        ]
 
         last = args[-1]
+        next_to_last = args[-2] if len(args) > 1 else ""
         suggestions: list[str] = []
 
-        # Case 1: Mid-flag (e.g., "--ver")
+        # Case 2: Mid-flag (e.g., "--ver")
-        if last.startswith("-") and not last in self._flag_map:
+        if last.startswith("-") and last not in self._keyword:
-            possible_flags = [flag for flag in self._flag_map if flag.startswith(last)]
+            if (
-            suggestions.extend(possible_flags)
+                len(args) > 1
-
+                and next_to_last in self._keyword
-        # Case 2: Flag that expects a value (e.g., ["--tag"])
+                and next_to_last in remaining_flags
-        elif last in self._flag_map:
+            ):
-            arg = self._flag_map[last]
+                # If the last token is a mid-flag, suggest based on the previous flag
-            if arg.choices:
+                arg = self._keyword[next_to_last]
-                suggestions.extend(arg.choices)
-
-        # Case 3: Just completed a flag, suggest next
-        elif len(args) >= 2 and args[-2] in self._flag_map:
-            # Just gave value for a flag, now suggest next possible
-            used_dests = {
-                self._flag_map[arg].dest for arg in args if arg in self._flag_map
-            }
-            remaining_args = [
-                a
-                for a in self._arguments
-                if not a.positional
-                and a.dest not in used_dests
-                and a.action != ArgumentAction.HELP
-            ]
-            for arg in remaining_args:
-                suggestions.extend(arg.flags)
-
-        # Case 4: Positional values not yet filled
-        else:
-            consumed_positionals = [arg for arg in self._arguments if arg.positional][
-                : len(args)
-            ]
-            remaining_positionals = [
-                arg
-                for arg in self._arguments
-                if arg.positional and arg not in consumed_positionals
-            ]
-            if remaining_positionals:
-                arg = remaining_positionals[0]
                 if arg.choices:
                     suggestions.extend(arg.choices)
-                else:
+                elif arg.suggestions:
-                    suggestions.append(f"<{arg.dest}>")  # generic placeholder
+                    suggestions.extend(arg.suggestions)
+            else:
+                possible_flags = [
+                    flag
+                    for flag, arg in self._keyword.items()
+                    if flag.startswith(last) and arg.dest not in consumed_dests
+                ]
+                suggestions.extend(possible_flags)
+        # Case 3: Flag that expects a value (e.g., ["--tag"])
+        elif last in self._keyword:
+            arg = self._keyword[last]
+            if arg.choices:
+                suggestions.extend(arg.choices)
+            elif arg.suggestions:
+                suggestions.extend(arg.suggestions)
+        # Case 4: Last flag with choices mid-choice (e.g., ["--tag", "v"])
+        elif next_to_last in self._keyword:
+            arg = self._keyword[next_to_last]
+            if arg.choices and last not in arg.choices:
+                suggestions.extend(arg.choices)
+            elif (
+                arg.suggestions
+                and last not in arg.suggestions
+                and not any(last.startswith(suggestion) for suggestion in arg.suggestions)
+                and any(suggestion.startswith(last) for suggestion in arg.suggestions)
+            ):
+                suggestions.extend(arg.suggestions)
+            else:
+                suggestions.extend(remaining_flags)
+        # Case 5: Suggest all remaining flags
+        else:
+            suggestions.extend(remaining_flags)
 
         return sorted(set(suggestions))
 

falyx/parser/signature.py

@@ -54,8 +54,10 @@ def infer_args_from_func(
         if arg_type is bool:
             if param.default is False:
                 action = "store_true"
-            else:
+                default = None
+            elif param.default is True:
                 action = "store_false"
+                default = None
 
         if arg_type is list:
             action = "append"
@@ -75,6 +77,7 @@ def infer_args_from_func(
                 "action": action,
                 "help": metadata.get("help", ""),
                 "choices": metadata.get("choices"),
+                "suggestions": metadata.get("suggestions"),
             }
         )
 

falyx/version.py

@@ -1 +1 @@
-__version__ = "0.1.62"
+__version__ = "0.1.63"

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "falyx"
-version = "0.1.62"
+version = "0.1.63"
 description = "Reliable and introspectable async CLI action framework."
 authors = ["Roland Thomas Jr <roland@rtj.dev>"]
 license = "MIT"

tests/test_parsers/test_completions.py

@@ -0,0 +1,18 @@
+import pytest
+
+from falyx.parser.command_argument_parser import CommandArgumentParser
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize(
+    "input_tokens, expected",
+    [
+        ([""], ["--help", "--tag", "-h"]),
+        (["--ta"], ["--tag"]),
+        (["--tag"], ["analytics", "build"]),
+    ],
+)
+async def test_suggest_next(input_tokens, expected):
+    parser = CommandArgumentParser(...)
+    parser.add_argument("--tag", choices=["analytics", "build"])
+    assert sorted(parser.suggest_next(input_tokens)) == sorted(expected)