[talk-au] vicmap import (was Re: How to efficiently improve AU address coverage?)

[Yuchen Pei]

On Mon 2023-10-02 12:22:46 +1100, Yuchen Pei wrote:

> On Mon 2023-10-02 11:45:05 +1100, Yuchen Pei wrote:

>> [... 10 lines elided]

>> Maybe we can go stage by stage.

> Stage 2 - Set unit from housenumber

> Judging from the command I guess the full info of the changesets are
> encoded in dist/conflate/mr_explodeUnitFromNumber.geojson. I did a
> search in that file for the occurrence of the string `"type":"Feature"`,
> but only found 212 of them, rather than ~9832 mentioned in the README. I
> wonder why that is the case?

Similarly, I grepped files for Stages 3 and 4.

For Stage 3, the README does not mention which geojson file to use.
Given the name of the stage ("New addresses in blocks without any
existing addresses"), my guess was
dist/conflate/noOSMAddressWithinBlock.geojson, and I got 1.3M lines in
grepping which is closer the to number in the README (~1,843,909
features) than Stage 2.

$ grep -o '"type":"Feature"' dist/conflate/noOSMAddressWithinBlock.geojson | wc -l
1335627

For Stage 4, I got only 79, compared to 3360 in the README

$ grep -o '"type":"Feature"' dist/conflate/mr_exactMatchSetFlats.geojson | wc -l
79

Is my method of counting wrong?

> [... 14 lines elided]

Best,
Yuchen

--
Timezone: UTC+11
PGP Key: 47F9 D050 1E11 8879 9040  4941 2126 7E93 EF86 DFD0
          <https://ypei.org/assets/ypei-pubkey.txt>