The following diagrams can be helpful to understand the GeoHub structure
The following sequence diagram shows how GeoHub works with other softwares in frontend and backend.
sequenceDiagram
autonumber
actor user
actor developer
participant client
participant maplibre
participant server
participant titiler
participant blob as Azure Blob
participant pg as pg_tileserv
participant mspc as Microsoft PC
participant db as PostGIS
user->>client: open GeoHub
alt is style parameter in URL
client->>+server: /style - request style info
server->>db: get style.json from PostGIS
server->>-client: return style data
else default
client->>client: initialise
end
user->>client: search datasets
client->>+server: /datasets - search datasets
server->>db: search
db->>server: return datasets
server->>-client: return datasets as GeoJSON
user->>+client: add datasets to map
client->>+server: /vector/{source}/metadata.json
server->>server: compute metadata
server->>-client: return metadata.json
client->>+server: /vector/{source}/tile.json
server->>server: compute tile
server->>-client: return tile.json
client->>maplibre: add layer in source
maplibre->>blob: request pbf
maplibre->>pg: request pbf
pg->>db: get pbf from PostGIS
maplibre->>+titiler: request raster tiles
titiler->>blob: read COG
titiler->>-maplibre: return png
maplibre->>maplibre: render map
client->>-user: provide rendered map
user->>client: operate to save map
maplibre->>client: style.json
client->>+server: /style - save style.json
server->>db: store style data
server->>-client: return style ID
user->>client: close GeoHub
In terms of Authentication with Azure Active Directory, the following figure shows how authentication works.
sequenceDiagram
autonumber
actor user
participant frontend
participant backend
participant authjs as Auth.js
participant azuread as Azure AD
user->>frontend: Sign in action
frontend->>authjs: Call authentication API
authjs->>azuread: Move to Azure AD login page
azuread->>user: Request user to login to Microsoft
user->>azuread: Login to Microsoft
azuread->>authjs: move to callback URL
authjs->>backend: store login info in session $page.data.session
authjs->>frontend: back to original geohub page
fronend->>user: Sign in complete
For the data upload pipeline which is managed by geohub-data-pipeline, the workflow is shown as the following diagram.
sequenceDiagram
autonumber
actor user
participant upload as /data/upload
participant portal as /data
participant publish as /data/publish
participant blob as Azure Blob Container
participant queue as Service Bus Queue
participant pipeline as GeoHub pipeline (AKS)
participant db as PostGIS
user->>+upload: Upload GIS data
upload->>blob: Upload blob to /raw folder
upload->>queue: Register message (Blob URL & token)
upload->>-user: Upload complete
queue->>pipeline: trigger pipeline, and receive message
blob->>pipeline: Download dataset from /raw folder
pipeline->>pipeline: Ingest dataset
alt if the dataset has a problem
pipeline->>blob: create .error file at /raw folder
else if the dataset is ingested successfully
pipeline->>blob: Upload ingested dataset to /datasets folder together with .ingesting file
end
user->>portal: Check the status of uploaded data
portal->>blob: scan /raw & /datasets folder to gather info
portal->>user: show the status of uploaded data
alt if the dataset is ready to publish
portal->>publish: Input metadata, click publish
publish->>db: register metadata in DB (it become searchable by /datasets api)
publish->>db: delete .ingesting file
publish->>portal: back to /data portal
else if the dataset has an error
portal->>portal: Check error message
portal->>blob: Delete uploaded dataset if user want
end
the diagram was created by mermaid online editor. Please read syntax of mermaid from the documentation
The following ER diagram was generated by ERD Editor extension for VS code.
