Run area aggregate (stats) updates only for an area and its ancestors

[vnugent]

Frontend issue: OpenBeta/open-tacos#877

Background

• Currently, the only way to update area statistics is by manually running a Kubernetes job.
• The job only updates USA tree
• Slow ~ 25 minutes

# The Kubernetes job calls
yarn update-stats

which calls https://github.com/OpenBeta/openbeta-graphql/blob/develop/src/db/utils/jobs/UpdateStatsJob.ts

To do

Based on recent work for issue #213 we can introduce a similar function to only update the affected subtree

[siman4457]

I'll take a crack at this.

[siman4457]

Updating a climb's stats affects its parents.

Potential Solution:

I don't think we have a way of referencing parent nodes. This can be fixed by adding a reference to parent nodes:

{
  "_id": {
    "$oid": "6467ab2ae622317bdb3ef74c"
  },
  "area_name": "Indian Creek",
  "climbs": [],
  "children": [
    ...    
  ],
  "parent": ObjectId("<id of parent area/crag>"),
  ...
}

In this example, the parent would be the id of Southeast Utah since the pathTokens shows:

[
  0: "USA",
  1: "Utah",
  2: "Southeast Utah",
  3: "Indian Creek"
]

[vnugent]

ancestors is a CSV list of area ID's mirroring pathTokens

[siman4457]

Child:

_id: ObjectId("6467ab2ae622317bdb3ef74c")
area_name: "Indian Creek"
ancestors: "1db1e8ba-a40e-587c-88a4-64f5ea814b8e,ee1d37e7-dc41-5c26-a358-5d51af01f499,cbbfd7c6-5cf2-5063-8939-5f3233efa573,f166e672-4a52-56d3-94f1-14c876feb670"
pathTokens:[
  0: "USA",
  1: "Utah",
  2: "Southeast Utah",
  3: "Indian Creek"
]

Parent:

_id: ObjectId("6467ab2ae622317bdb3eef3f")
area_name: "Southeast Utah"

_id 6467ab2ae622317bdb3eef3f is not found in the ancestor list of the child. Is there some conversion between the _id and strings saved in ancestors?

I see that a mmuid is being generated from paths in genMUIDFromPaths() in src/model/MutableAreaDataSource.ts. Seems like the mmuid generated is based on the area name and not the area id.

const uuid = genMUIDFromPaths(parentPathTokens, areaName)

[vnugent]

For the Areas collection, _id is auto-generated by mongo. I believe we're not using it anywhere. What we use as the primary key is metadata.area_id (it's the uuid generated from genMUIDFromPaths())

Indian Creek

ancestors: "1db1e8ba-a40e-587c-88a4-64f5ea814b8e,ee1d37e7-dc41-5c26-a358-5d51af01f499,cbbfd7c6-5cf2-5063-8939-5f3233efa573,f166e672-4a52-56d3-94f1-14c876feb670"

cbbfd7c6-5cf2-5063-8939-5f3233efa573 is area_id of South East Utah

Looking it up in Mongo Express
{ metadata.area_id: UUID('cbbfd7c6-5cf2-5063-8939-5f3233efa573') }