You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Each of our Elasticsearch documents stores the original source ID from the upstream data in two places: as part of the GID in the _id field, and in the source_id field.
Conservatively estimating 10 bytes per record, this probably represents about 6GB of duplicated data in a full planet build, and thus it would make sense to get rid of it when possible.
There will be changes required to the API, pelias/model, and here in pelias/schema to do this.
We'll also have to ensure reasonable backwards compatibility is maintained. Especially in the API, where the source_id field in our GeoJSON responses will still be required.
The text was updated successfully, but these errors were encountered:
This change implements the 5th and final step in the GID migration plan
outlined in
pelias/pelias#672 (comment).
It adjusts the expectations of the API so that it requires the
full GID, rather than the source_id, to be stored in the `_id` field in
Elasticsearch.
In the process, the `source_id` field is also no longer used in
`helper/geojsonify.js`, making this the first step towards removing the
`source_id` field (pelias/schema#383).
Connects pelias/pelias#672
This change implements the 5th and final step in the GID migration plan
outlined in
pelias/pelias#672 (comment).
It adjusts the expectations of the API so that it requires the
full GID, rather than the source_id, to be stored in the `_id` field in
Elasticsearch.
In the process, the `source_id` field is also no longer used in
`helper/geojsonify.js`, making this the first step towards removing the
`source_id` field (pelias/schema#383).
Connects pelias/pelias#672
Each of our Elasticsearch documents stores the original source ID from the upstream data in two places: as part of the GID in the
_id
field, and in thesource_id
field.Conservatively estimating 10 bytes per record, this probably represents about 6GB of duplicated data in a full planet build, and thus it would make sense to get rid of it when possible.
There will be changes required to the API, pelias/model, and here in pelias/schema to do this.
We'll also have to ensure reasonable backwards compatibility is maintained. Especially in the API, where the
source_id
field in our GeoJSON responses will still be required.The text was updated successfully, but these errors were encountered: