Skip to content

Basic Configuration

All configuration of Daiquiri is done using yaml files

Tip

An example of each configuration file can be found in daiquiri/resources/example_config

Info

For general usage configuration files should be stored in the local beamline project i.e. daiquiri_idxx/daiquiri_idxx/resources/config. This should be cookiecutter'd from the https://gitlab.esrf.fr/ui/daiquiri-local project. This project can provide a set a preconfigured files using the sensible-defaults branch

app.yml

This file defines the core configuration of the application. The example can be copied from daiquiri/resources/example_config to daiquiri_idxx/daiquiri_idxx/resources/config/app.yml and updated accordingly.

Secrets

Unique secrets should be generated for iosecret, secret, and url_secret, this can be done with: 

cat /dev/urandom | env LC_CTYPE=C tr -dc 'a-zA-Z0-9' | fold -w 64 | head -n 1

SSL

SSL must be enabled and for this to work a local ssl key / cert pair must be available which matche the machine name, then can be defined with: 

ssl: true
ssl_cert: my_cert.bundle
ssl_key: my_key.key

Authentication

Authentication is modular, currently supporting ldap 

auth_type: ldap
auth_server: ldap.esrf.fr

Metadata

The metadata handler should be set to ispyalchemy to make use of the mysql database. For worse case scenarios it is possible to drop back to a mock metadata handler. The meta_beamline should match the beamline session information, usually the beamline in upper case. meta_staff denotes the permission for which daiquiri deems the current user to be a staff member 

meta_type: ispyalchemy
meta_user: daiquiri
meta_pass: pass
meta_url: host:3306/daiquiri
meta_beamline: ID21
meta_staff: id21_staff

By default sessions are visible one hour either side of their start time. This can be modified for users and staff independently:

meta_session_grace: 2

# A different value for staff members
meta_session_grace_staff: 24

Saving

Two saving types are currently supported bliss_basic and bliss_esrf. The basic saving makes use of Bliss's standard scan saving, where as the bliss_esrf saving plugin requires the bliss esrf saving policy to be enabled and will ingest datasets automatically into ICAT using the metadata manager, where a single actor execution will produce a dataset.

To configure the bliss_basic scan saving the follow yml properties are required: 

saving_type: bliss_basic
saving_session: bliss_session
saving_arguments:
    base_path: "/data/{meta_beamline}/tmp/{sessionid.proposal}"
    template: "{sampleid.name}"
    data_filename: "{datacollectionid}"

Using bliss_esrf the data policy can be automatically applied by using daiquiri's saving, an example config is as follows: 

saving_type: bliss_esrf
saving_session: bliss_session
saving_arguments:
    proposal: "{sessionid.proposal}"
    sample: "{sampleid.name}"
    dataset: "{datacollectionid}"

Extra metadata can be provided to ICAT for each dataset by providing the saving_metadata property. For example: 

saving_metadata:
    sample_name: "{sampleid.name}"
    sample_description: "{componentid.description}"
    sample_composition: "{componentid.sequence}"
    dataset_definition: "{metadata_definition}"

These properties will be passed into ICAT to define details about the sample and the definition field

Without Bliss

A standalone saving type is provided in order to use daiquiri without any controls system saving integration. This allows daiquiri to set the actor directory and file template, which can then be consumed directly within the actor.

saving_type: simple
saving_arguments:
    base_path: "/data/{meta_beamline}/tmp/{sessionid.proposal}"
    template: "{sampleid.name}"
    data_filename: "{datacollectionid}"

It supports the same keys as the bliss_basic saving type, primarily base_path, template, and data_filename.

Controls Session

A controls session can be injected automatically into actors to make available session globals: 

controls_session_type: bliss
controls_session_name: bliss_session

Implementors

The implementors directive should point to the local project implementors package name. This will be automatically set by the daiquiri-server-idxx shell script 

implementors: daiquiri_id21.implementors

Components

Components can be loaded by creating a yml file with a component key in the resources/config folder. The daiquiri/resources folder comes with a set of sensible defaults. A list of available components can be found in Components

For example config/scans.yml 

component: scans
...
some_value: true
...

Swagger-UI

Automatic documentation of the API is provided by Swagger-UI. It is disabled by default but can be enabled via:

swagger: true

Once enabled it can be accessed at http://localhost:8080/swagger-ui/