Make sure to setup and configure the infrastructure before using it.
| Warning | ||
|---|---|---|
| ||
This page is under construction. Do not trust it until this box is removed |
| Note |
|---|
In current setup any request to HMI data will transparently load data from JSOC. This may take a while for large time ranges. |
Startup Infrastructure
| Info | ||
|---|---|---|
| ||
|
| Info | ||
|---|---|---|
| ||
The Docker daemon should have been started automatically. |
| Code Block |
|---|
# startup docker-composeinfrastructure cli> docker-compose up./infrastructure.sh run |
Update Infrastructure
Occasionally you may want to upgrade to the latest version of the infrastructure.
| Note |
|---|
This may erase the content of your database and you will have to reprocess the data. This is certainly something we need to improve. |
| Code Block |
| cli> curl -o docker-compose.yml https://dev.flarecast.eu/stash/projects/INFRA/repos/dev-infra/browse/docker-compose.yml?raw
cli> docker-compose pull Sometimes it may be necessary to upgrade the whole database (e.g. schema changed). In such a case you may get an error about the database was not able to start. You have then to delete your old database to use the most current version and your old data will be lost. |
| Code Block |
|---|
cli> ./infrastructure.sh update |
Open FLARECAST Pipeline Management Console
| Info | ||||
|---|---|---|---|---|
| ||||
Switch to "Kitematic" Select Container "devinfrastest_webapp_1" On the right side panel click on the arrow icon ( |
| Info | ||
|---|---|---|
| ||
Open http://120127.0.0.1:8000/ in your browser |
Now you can
- Inspect registered algorithms (tab "Algorithms")
- Manually start a new instance of an algorithm in a dedicated container (tab "Pipeline")
- Manage running and stopped containers (tab "Containers")
read the descriptions of the services and access all the services via web interfaces
TODO: Update section below, according to interface of infra-viewer (no more webapp).
Inspect registered Algorithms
In tab "Algorithms" you can get information about registered algorithms.
Start new instance of an Algorithm (Pipeline)
- Open tab "Pipeline"
- Select pipeline step
- Download: download remote data and ingest into db (e.g SWPC, typically maintained by FHNW)
- Feature extraction: extract features and ingest into property_database (e.g. SMART)
- Machine learning (training)
- Machine learning (execution)
- Visualisation (experimental): algorithms to visualise data.
- Select algorithm to run
- "Setup new instance"
- Modify input parameters (if needed, please note that modifications will not be persisted over multiple sessions)
- "Run new instance"
Intermezzo: add existing algorithm container to infrastructure
Use this commands if a new algorithm container gets added to the FLARECAST Docker hub infrastructure
...