Add fixes for ecoinvent units

[BenPortner]

This PR adds a couple of fixes, which are necessary to use the new PintInterpreter with ecoinvent databases. Ecoinvent uses some specific units (like EUR2005, Sm3, unit), which cannot be interpreter by pint out of the box. I added a ecoinvent_units.txt file to easily add these definitions to pint. I also wrote tests to make sure they are imported correctly.

@cmutel I am not sure if the definition file should be in bw2parameters or rather in bw2data. What do you think?

[BenPortner]

Successful workflow result here: https://github.com/brightway-lca/brightway2-parameters/actions/runs/7966380898/job/21747522619 (dunno why Github doesn't show it on the MR...)

[cmutel]

This is needed for ecoinvent but makes me nervous for general application...

[BenPortner]

You mean because "short ton" is no longer defined? If you want, I can add it back under a different name

[cmutel]

I thought that these prefixes were already defined in Pint, but I guess I was dreaming? But if we need to define prefixes that this shouldn't be in "ecoinvent specific stuff".

[BenPortner]

Yes and no. Pint can interpret stuff like this:

>>> from pint import UnitRegistry
>>> ureg = UnitRegistry()
>>> ureg("kilobecquerel").to("becquerel")
<Quantity(1000.0, 'becquerel')>

Unfortunately, ecoinvent does not use "kilobecquerel". They write "kilo Becquerel" and that causes two errors in pint: it tries to interpret "kilo" (which it doesn't know out of the box) and it tries to interpret "Becquerel" (which it doesn't recognize because of the capital letter). So these lines really are necessary just to make ecoinvent work.

[cmutel]

You can add normal cubic meter as well, see https://github.com/cmutel/flowmapper/blob/main/flowmapper/data/units.txt#L7C1-L7C34

[BenPortner]

I think this requires some clarification.

First off, "normal conditions" and "standard conditions" are very problematic terms. NIST defines "normal conditions" as 20°C and 1 atm. DIN defines it as 0°C and 1 atm. IUPAC defines "standard conditions" as 0°C and 1000 bar, however, before 1982 it defined it as 0°C and 1 atm (making it effectively equivalent to DIN normal conditions). Apart from these four, there are a range of industry-specific definitions, which may take different values altogether. Even if we assume naming consistency within all datasets of one database (I doubt it), assuming consistency between different databases is a leap of faith. This begs the question: do we really want to allow these conversions?

If we say "yes, we definitely want to convert between these", I see another problem. Taking the NIST definition for "normal conditions" and the current IUPAC definition for "standard conditions", I calculate the conversion factor as follows:

$V_n = V_s \cdot \dfrac{p_s}{p_n} \cdot \dfrac{T_n}{T_s} = V_s \cdot \dfrac{100000 Pa}{101325 Pa} \cdot \dfrac{293.15 K}{273.15 K} = 1.0592 \cdot V_s$

You propose a conversion factor of 1.0732 instead. Could you explain where it comes from?

[cmutel]

This name should reflect that it is ecoinvent-specific.

[BenPortner]

Yeah, I was a bit sloppy here. Confession: not all the units in ecoinvent_units.txt are really from ecoinvent. I added some of my own. In general I think it's a good idea to have a place where the user can add custom definitions. I'm just not sure if we should have an extra file for them. What do you think?

[BenPortner]

@cmutel not pushing, just making sure you've noticed my answer.