Forced photometry update pipeline - missing match stage

[Theodlz]

That had been hot patched early on in production (without it, it just got stuck and didn't even finish running any query anyway, so none of the resulting data in the DB has been affected), but adding it here to not forget later on.

[mcoughlin]

Might be worth a quick comment on what drove the allowDiskUse change.

[Theodlz]

Might be worth a quick comment on what drove the allowDiskUse change.

It actually shouldn't be necessary. Basically, before I noticed the missing stage (which I had on a local change, didn't commit it to the original PR I guess...) the sort stage wouldn't run because well it tried to grab the whole table. So I added the allowDiskUse that basically means that if you need more than the 100MB limit for an aggregation pipeline stage, you can use some of the disk.

Now that the missing stage is added it shouldn't be necessary. However, I will still leave it so that if in the future some objects accumulate enough data to go over that limit, it won't crash like it did.

This is already documented in multiple places in the code, so I'd rather not add that here (especially given that I am not sure it will ever be needed in this context now that the pipeline has been fixed).

[mcoughlin]

Sounds good Theo