I'm confused about MeterStatDict

[NorthIsUp]

This is mainly about naming things, but tell me if I'm way wrong on this.

So I have the following stats:

STATS = scales.collection('/api',
    meter.MeterStat('api_v1'),
    meter.MeterStat('multiplexed'),
    scales.IntDictStat('status_code'),
    scales.PmfStat('latency'),
    )

I can go STATS.status_code[200] += 1, but what I really want to do is have a default dictionary of meters so I could do STATS.status_code[200].mark(), but MeterStatDict doesn't seem to be anything like IntDictStat.

Can somebody explain/document the thoughts behind some of these data structures?

p.s. you guys going to py con? I'd love to meet up.

[PeterScott]

From looking at the code -- and I didn't write this part, so don't quote me on this -- it looks like MeterStatDict is part of the internal workings of MeterStat, and not meant to be used directly. The best workaround I can think of here is to create a bunch of MeterStats, one for each HTTP status code. I'm sure there's a way to do this automatically (see "Class Stats" in the readme for my best guess as to how), but there's a lot of magic in the scales internals, and I forget how most of it works.

If you do end up making something like a default dictionary of meters, that would be really cool.

[NorthIsUp]

As MeterStatDict is internal, I may submit a patch re-naming it to avoid confusion. Something like _MeterStatImpl. Or something.

[PeterScott]

That sounds reasonable; just adding the leading underscore could help. Of course, there are a bunch of other classes whose names could use changing like that, such as IntDict, which is internal to IntDictStat.

In general, I wish that scales had more documentation. It's also due for a code cleanup one of these days.