Fix amounts and add tests

[TimoDiepers]

The amounts just happened to work out because some exchange amounts were set to 1, so the interpolated shares had no influence on the total. We should also add tests checking if the Medusalca calculations are correct.

[TimoDiepers]

• Fixed amounts
• Added tests

[TimoDiepers]

Amounts fixed in 23ea70c

The problem was that we had to use the "unscaled" amounts from the original edge here, as we explode to new unit processes. In contrast, the amounts from the timeline are scaled to match the FU.

[TimoDiepers]

Additional problem came to my attention:
The setting: A foreground node (H2) links to a background node (el. mix) via a TD.

What happens in our code atm is that we don't link from H2 to exploded versions of el-mix, but link directly to the different available version of this process in the background dbs via interpolation. However, if two of these interpolations point towards the same background process, the matrix entry we add gets overwritten each time.

I'm pretty sure I talked about this with @cmutel at the Autumn School and he said that the default behavior of Datapackage.add_persistent_vector is to add the values instead of overwriting them - which seems to not be the case.
@jakobsarthur do you happen to know if changing this behavior is possible? Going through Datapackage.add_persistent_vector() I couldn't find what we need..

[TimoDiepers]

Nvm, found the sum_inter_duplicates flag in bwp.create_datapackage()

[TimoDiepers]

From the datapackage-creation-side, this should be fixed in b84644f
However, the results are still wrong. But I think the problem might be in the timestamps in the timeline df.

[jakobsarthur]

Great! I will check the timestamps of the timeline. Still working on that, also in issue #5.