Support passing in pValue

[arj03]

This is what I have used to test ssbc/jitdb#208. So not really ready yet.

Also we can do the same trick here for level indexes :)

[arj03]

Some numbers:

Only base + keys index & query for 1 author empty indexes

base: 41.1s
no decrypt: 30.2
no decrypt + jitdb pValue: 25.6

Base, keys, ebt, full mentions index & query for 1 author empty indexes

base: 56.3s
no decrypt + jitdb pValue: 37.7

So the more level plugins you have, the better it gets. It might also be a good idea to pass things like pContent to level plugins. Maybe even jitdb.

[github-actions]

Benchmark results

Part	Duration
add 1000 elements	524ms
add 1000 private box1 elements	1044ms
unbox 1000 private box1 elements first run	119ms
unbox 1000 private box1 elements second run	79ms
add 1000 private box1 elements	1047ms
query 1000 elements first run	31ms
query 1000 elements second run	16ms
add 1000 private box2 elements	574ms
unbox 1000 private box2 elements first run	322ms
unbox 1000 private box2 elements second run	320ms
Migrate (+db1)	10547ms
Migrate (alone)	3232ms
Migrate (+db1 +db2)	6954ms
Migrate (+db2)	5419ms
Migrate continuation (+db2)	815ms
Memory usage without indexes	721.51 MB = 32.24 MB + etc
Initial indexing	544ms
Initial indexing maxCpu=86	3641ms
Initial indexing compat	634ms
Two indexes updating concurrently	795ms
key one initial	41ms
key two	1ms
key one again	2ms
reboot and key one again	40ms
latest root posts	537ms
latest posts	11ms
votes one initial	457ms
votes again	2ms
hasRoot	301ms
hasRoot again	1ms
author one posts	226ms
author two posts	16ms
dedicated author one posts	264ms
dedicated author one posts again	2ms
Maximum memory usage	808.48 MB = 46.82 MB + etc
Indexes folder size	12.76mb

[staltz]

Oh silly me, now I see why we can't move the pValue calculation. Sorry for that :)

[staltz]

@arj03 By the way, I just had an idea for using WeakMap so that we don't need to recalculate any of these pointers. It could be put into BIPF and would work transparently. Then we could revert these pValue changes and we would still have the same performance benefits

[arj03]

Uhh nice!

[github-actions]

Benchmark results

Part	Duration
add 1000 elements	509ms
add 1000 private box1 elements	1412ms
unbox 1000 private box1 elements first run	166ms
unbox 1000 private box1 elements second run	124ms
add 1000 private box1 elements	1372ms
query 1000 elements first run	34ms
query 1000 elements second run	16ms
add 1000 private box2 elements	781ms
unbox 1000 private box2 elements first run	442ms
unbox 1000 private box2 elements second run	435ms
Migrate (+db1)	14705ms
Migrate (alone)	5230ms
Migrate (+db1 +db2)	10758ms
Migrate (+db2)	7441ms
Migrate continuation (+db2)	1232ms
Memory usage without indexes	736.76 MB = 31.09 MB + etc
Initial indexing	755ms
Initial indexing maxCpu=86	4685ms
Initial indexing compat	968ms
Two indexes updating concurrently	1081ms
key one initial	59ms
key two	4ms
key one again	2ms
reboot and key one again	51ms
latest root posts	729ms
latest posts	17ms
votes one initial	587ms
votes again	2ms
hasRoot	461ms
hasRoot again	1ms
author one posts	361ms
author two posts	18ms
dedicated author one posts	344ms
dedicated author one posts again	2ms
Maximum memory usage	799.99 MB = 44.66 MB + etc
Indexes folder size	12.76mb