TS Active Atoms Invalid Input

[MattBurn]

Describe the bug
In BlockGeom the ts_active_atoms is of type Hybrid which inherits from GeomWrapper requiring a mode string input.

ts_active_atoms does not require a mode as it is simply a list of integers as can be seen in the following documentation

To Reproduce
The naive approach of simply using a list[int]

from opi.core import Calculator
from opi.input.blocks import BlockGeom

calc = Calculator("job")
calc.input.add_blocks(BlockGeom(ts_active_atoms=[1, 2, 3]))
calc.write_input()

Traceback (most recent call last):
  File "opi_test.py", line 5, in <module>
    calc.input.add_blocks(BlockGeom(ts_active_atoms=[1, 2, 3]))
                          ~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File ".venv/lib/python3.14/site-packages/pydantic/main.py", line 263, in __init__
    validated_self = self.__pydantic_validator__.validate_python(data, self_instance=self)
pydantic_core._pydantic_core.ValidationError: 1 validation error for BlockGeom
ts_active_atoms
  Input should be a valid dictionary or instance of Hybrid [type=model_type, input_value=[1, 2, 3], input_type=list]
    For further information visit https://errors.pydantic.dev/2.13/v/model_type

Correctly wrapping the types with Hybrid and NumList

from opi.core import Calculator
from opi.input.blocks import BlockGeom, Hybrid, NumList

calc = Calculator("job")
calc.input.add_blocks(BlockGeom(ts_active_atoms=Hybrid(list=NumList([1, 2, 3]))))
calc.write_input()

Traceback (most recent call last):
  File "opi_test.py", line 5, in <module>
    calc.input.add_blocks(BlockGeom(ts_active_atoms=Hybrid(list=NumList([1, 2, 3]))))
                                                    ~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^
  File "venv/lib/python3.14/site-packages/pydantic/main.py", line 263, in __init__
    validated_self = self.__pydantic_validator__.validate_python(data, self_instance=self)
pydantic_core._pydantic_core.ValidationError: 1 validation error for Hybrid
mode
  Field required [type=missing, input_value={'list': NumList(numlist=[1, 2, 3])}, input_type=dict]
    For further information visit https://errors.pydantic.dev/2.13/v/missing

Workaround (setting mode="")

from opi.core import Calculator
from opi.input.blocks import BlockGeom, Hybrid, NumList

calc = Calculator("job")
calc.input.add_blocks(BlockGeom(ts_active_atoms=Hybrid(mode="", list=NumList([1, 2, 3]))))
calc.write_input()

Produces the following orca input:

%geom
    ts_active_atoms {  1,2,3 } end
end

%output
    jsonpropfile True
    jsongbwfile True
end

Which will work but is not intuitive.

Expected behavior
I am thinking that ts_active_atoms should not use Hybrid not sure of the correct type but we essentially need something that wraps NumList without the mode attribute.

Not sure what other attributes will also need the same fix.

[haneug]

Thanks for reporting this issue. I agree Hybrid is not correct here. Maybe IntGroupEnd is already sufficient? We will look into this and supply a fix.