chhecker / unify Goto Github PK

It would be nice if unicode exponents could be used to write units.
normal:
#qty("2.5", "J mol^-1 K^-1")
unicode:
#qty("2.5", "J mol⁻¹ K⁻¹")

It should be possible to set the default decimal separator. I tried doing

#set _format-float(decsep: ",")

$ num("2.0") $

as I prefer to leave the numbers in code with dots (yet want it to show up as 2,0), but it throws a syntax error. Any way to set that in another way?

(originally typst/packages#188)

Hi.
Sorry to disturb you.
When I tried to use the coulomb measurement ubit, it didn't work.
I realized, as in the attachement, that there was a typo in the unit.csv file, probabbly it is necessary to remove the character ".
By the way, in LaTeX SIunits, when you type the scientific notation the multiplication sign is a times and not a dot. Could it be possible to have the possibility tto choice between a dot or times symbol also in your package (I attach a file to be clearer).
Thank a lot and sorry for my bad English.
Best regards,
Roberto
units.csv.pdf
Generic.pdf

• #unit("hz") ⇒ h (hecto ?)
• #unit("Hz") ⇒ zH (zepto Henry)

Results of #qty is similar.

I am using v0.4.2 and I've verified this issue on current main (d2ab4c3).

Workaround

Don't use units-short: #unit("hertz") works as expected.

Reason

• h is a valid prefix (hecto or 100).
• H is a valid unit (Henry).
• z is a valid prefix (zepto or $10^{-21}$).

Therefore the checks pass, and we try to parse z and fail. However, we don't #panic and let it go…

Possible Fix

In #format-unit-short, remove the else if in the for loop…

    if acc.len() > 0 {
      panic("invalid unit: " + quantity)
    }

This will make #unit("hz") into error: panicked with: "invalid unit: hz".

In the same for loop, add and unit == none in this if, because the prefix must appear before the unit:

This will make #unit("Hz") into error: panicked with: "invalid unit: Hz".

A Bigger Refactor?

I suggest a bigger refactor of the parsing logic in #format-unit-short…

1. If quantity in units-short, unit = units-short.at(quantity), ✓.
2. If not, divide quantitiy into two parts, if the first part is a prefix and the second is a unit, ✓.
3. If not, try to divide at other positions.
4. If all division fail, panic.

How do you think so? If you agree but are busy, I can try to work out a PR.

There is an additional thin space at the beginning when using uncertainties, that shouldn't be there:

#import "@preview/unify:0.4.2": qty, num

#qty("200+-5", "m") \
#num("200+-5")

$(200 plus.minus 5)$ // How it should look

(Using web app, 0.8.0)

Typically, the only units you don't use spaces before are a lone-standing degree symbol and primes. This means there should be one if the unit is "°C", "°F" etc., since the "°" has a "suffix" (see e.g. [https://german.stackexchange.com/a/322]).

The Dyne is used in astronomy in the cgs unit system.

I'm looking for a way to add custom units (e.g., "hours") while importing unify from @preview.
According to the Readme:

The possible values of the three latter parts are loaded at runtime from prefixes.csv, units.csv, and postfixes.csv (in the library directory). There, you can also add your own units.

However, when importing from @preview, these CSVs aren't available to me.

Would it make sense to either

• maintain the corresponding dictionaries like units, units-short, etc., as state and add a function for easily adding a new unit, prefix, or postfix; or
• add parameters of the same names as those dictionaries to _format-unit defaulting to the existing dictionaries, which should allow users to use set rules to specify the extended dictionaries?