Draft: parse prefix as scale factor

Siehe #17 (closed)

Klappt jetzt fast alles. Es kann sein, dass nicht alle Ausgabenmodi die Scale-Faktoren enthalten. Ich habe auch noch nicht die Integration mit dccQuantities getestet.

Die Umrechnung in Base-Einheiten hat auch noch ein paar Probleme, siehe Test-case unten. Bin aber nicht sicher, ob das in den gleichen PR gehört.

Test-Code

from dsiUnits import dsiUnit

u_cm = dsiUnit(r"\centi\metre")
u_m = dsiUnit(r"\metre")
u_s = dsiUnit(r"\second")


u_cm2 = dsiUnit(r"\centi\metre\tothe{2}")
u_mcm = dsiUnit(r"\metre\centi\metre")
u_cmm = dsiUnit(r"\centi\metre\metre")

u_cm1 = dsiUnit(r"\centi\metre\tothe{-1}")
u_pcm = dsiUnit(r"\one\per\centi\metre")


print(f"s * cm = {u_s*u_cm} = {(u_s*u_cm).toBaseUnitTree()}")
print(f"cm * s = {u_cm*u_s} = {(u_cm*u_s).toBaseUnitTree()}")
print(f"cm * m = {u_cm*u_m} = {(u_cm*u_m).toBaseUnitTree()}")
print(f"m * cm = {u_m*u_cm} = {(u_m*u_cm).toBaseUnitTree()}")
print("")

print(f"cm = {u_cm} = {u_cm.toBaseUnitTree()}" )
print(f"cm2 = {u_cm2} = {u_cm2.toBaseUnitTree()}" )
print("")

print(f"cm-1 = {u_cm1} = {u_cm1.toBaseUnitTree()}" )
print(f"1/cm = {u_pcm} = {u_pcm.toBaseUnitTree()}" )
print("")

print(f"cm2^-1 = {u_cm2**-1} = {(u_cm2**-1).toBaseUnitTree()} " )
print(f"cm / (cm2) = {u_cm/u_cm2} = {(u_cm/u_cm2).toBaseUnitTree()}" )

print("")
print(f"cm2 / m = {u_cm2/u_m} = {(u_cm2/u_m).toBaseUnitTree()}" )
print(f"m*cm / (cm2) = {u_mcm/u_cm2} = {(u_mcm/u_cm2).toBaseUnitTree()}" )
print("")

u_eV = dsiUnit(r"\electronvolt")
u_J = dsiUnit(r"\joule")
print(f"eV = {u_eV} = {u_eV.toBaseUnitTree()}")
print(f"J = {u_J} = {u_J.toBaseUnitTree()}")
print(f"eV/J = {(u_eV/u_J)} = {(u_eV/u_J).toBaseUnitTree()}")
print(f"J*eV = {u_J*u_eV} = {(u_J*u_eV).toBaseUnitTree()}")

Ausgabe

s * cm = 0.01*\metre\second = 0.01*\metre\second
cm * s = 0.01*\metre\second = 0.01*\metre\second
cm * m = 0.01*\metre\tothe{2} = 0.01*\metre\tothe{2}
m * cm = 0.01*\metre\tothe{2} = 0.01*\metre\tothe{2}

cm = 0.01*\metre = 0.01*\metre
cm2 = 0.0001*\metre\tothe{2} = 0.0001*\metre\tothe{2}

cm-1 = 100.0*\metre\tothe{-1} = 100.0*\metre\tothe{-1}
1/cm = \one\per0.01*\metre = 100.0*\metre\tothe{-1}

cm2^-1 = 10000.0*\metre\tothe{-2} = 10000.0*\metre\tothe{-2}
cm / (cm2) = \one\per0.01*\metre = 100.0*\metre\tothe{-1}

cm2 / m = 0.0001*\metre = 0.0001*\metre
m*cm / (cm2) = 100.0*\one = 100.0*\one

eV = \electronvolt = 1.602176634e-19*\joule
J = \joule = \kilogram\metre\tothe{2}\second\tothe{-2}
eV/J = \electronvolt\per\joule = 1.602176634e-19*\joule\kilogram\tothe{-1}\metre\tothe{-2}\second\tothe{2}
J*eV = \electronvolt\joule = 1.602176634e-19*\joule\kilogram\metre\tothe{2}\second\tothe{-2}

requested review from @Seeger

assigned to @Seeger

added 2 commits

• 3bb02b52 - prefix in latex output and bug fix in node scaling
• eb8f1b65 - force test pass

Compare with previous version

We were able to reproduce the error and will have a look at this fix. In the next version (next week) we will change the behavior as suggested here (calculating scale factors instead of arithmetic on the Prefixes). We will ad a function to convert a scale Factor to an SI prefix if this is possible as well as an option to rest the scale factor.

The use case for the scale factor reset is: The scale factor was applied to the data so it dosn't need to be stored in the unit any longer.

The baseUnitConversion errors are strongly related to this bug and will be fixed with this.

eV needs extra treatment (eV is converted to J in the base unit conversion) but (J wasn't equal in the first round so its converted to its base unit representation leading to inequality in the next conversion)

Many thanks for your support

mentioned in commit c919e272

mentioned in commit 6e4ba046

closed