streamlit-quantity-text-input 0.1.0a0

Streamlit input field that uses the pint library for parsing quantities

streamlit_quantity_text_input

This Python module gives you a Streamlit text input field (like st.text_input) for entering and processing concentrations, volumes, amounts ect. in free-text, yet in an input-safe manner.

The quantity_text_input field uses the pint library to parse the user's input as a quantity (volume, speed, concentration etc.).

Features

• The input field can be configured to accept only inputs of a given dimensionality (e.g. volume) and will accept inputs in a wide range of relevant units.
• The input field can be configured to return the input value in your preferred unit (such as milliliters, barrels, in³ etc.).
• In case of user input error, the input field displays a helpful error message to the user immediately below the field.
• Developers can specify a callable for pre-processing the user's input before it is handed over to the input field's parsing algorithm.
• The return value, a pint.Quantity instance, can be used downstream for unit-aware calculations.

Getting started

$ pip install streamlit-quantity-text-input

from streamlit_quantity_text_input import quantity_text_input

vol = quantity_text_input(
    "Please input a volume:",
    mandatory_dimension="volume",
    output_unit="ml",
    lower_limit="0 l")

print(f"You input volume is: {vol if vol else 'pending'}")

Check out the accompanying demo.py for a more detailed example and documentation.

Roadmap

Wanted but not yet planned:

• Test cases: Unit tests, perhaps via the doctest module.

• Flag: relaxed_time_input. Implement a flag to allow a much broader range of types of time inputs to be parsed correctly, consistently (e.g. 1h45, 1hrs3minutes, 2wks60'45" etc.). Bonus points for offering a solution that also addresses Danish spellings (perhaps as a pre-parser).

• Fool-profing the math parser: The pint library treats scalars and units equally, which can lead to some unexpected parsing when mixing scalars and units (1 m / 2 s parses into 1*m/2*s=½ m*s, not ½ m/s). Presumably, this can be fixed by tokenizing the input string and 'gluing' scalars and their abutting unit(s) together with parenthesis (((1*m)/(2*s))). However, writing the tokenizer could be non-trivial.

License

See LICENSE.txt

Project status

New as of April 2024.