DiamondLightSource / hyperion

Unattended Data Collection using BlueSky / Ophyd
BSD 3-Clause "New" or "Revised" License
5 stars 5 forks source link

hyperion

Tests codecov

Repository for the Hyperion project to implement Unattended Data Collections on the Diamond MX beamlines using the BlueSky / Ophyd framework from BNL.

Currently the software is able to:

Left to do is:

Development Installation

This project supports only the most recent Python version for which our dependencies are available - currently Python 3.11.

Run ./utility_scripts/dls_dev_env.sh (This assumes you're on a DLS machine. If you are not, you should be able to just run a subset of this script)

Note that because Hyperion makes heavy use of Dodal this will also pull a local editable version of dodal to the parent folder of this repo.

Controlling the Gridscan Externally (e.g. from GDA)

Starting the bluesky runner

You can start the bluesky runner by running run_hyperion.sh. Note that this will fail on a developer machine unless you have a simulated beamline running, instead you should do run_hyperion.sh --dev --skip-startup-connection, which will give you a running instance (note that without hardware trying to run a plan on this will fail). The --dev flag ensures that logging will not be sent to the production Graylog.

This script will determine whether you are on a beamline or a production machine based on the BEAMLINE environment variable. If on a beamline Hyperion will run with INFO level logging, sending its logs to both production graylog and to the beamline/log/bluesky/hyperion.log on the shared file system.

If in a dev environment Hyperion will log to a local graylog instance instead and into a file at ./tmp/dev/hyperion.log. A local instance of graylog will need to be running for this to work correctly. To set this up and run up the containers on your local machine run the setup_graylog.sh script.

This uses the generic defaults for a local graylog instance. It can be accessed on localhost:9000 where the username and password for the graylog portal are both admin.

The hyperion python module can also be run directly without the startup script. It takes the same command line options, including:

INFO level logging of the Bluesky event documents can be enabled with the flag

python -m hyperion --dev --verbose-event-logging

Lastly, you can choose to skip running the hardware connection scripts on startup with the flag

python -m hyperion --skip-startup-connection

Testing

Unit tests can be run with python -m pytest -m "not s03" --random-order. To see log output from tests you can turn on logging with the --logging command line option and then use the -s command line option to print logs into the console. So to run the unit tests such that all logs are at printed to the terminal, you can use python -m pytest -m "not s03" --random-order --logging -s. Note that this will likely overrun your terminal buffer, so you can narrow the selection of tests with the -k "<test name pattern>" option.

To be able to run the system tests, or a complete fake scan, we need the simulated S03 beamline. This can be found at: https://gitlab.diamond.ac.uk/controls/python3/s03_utils

To fake interaction and processing with Zocalo, you can run fake_zocalo/dls_start_fake_zocalo.sh, and make sure to run module load dials/latest before starting hyperion (in the same terminal).

Tracing

Tracing information (the time taken to complete different steps of experiments) is collected by an OpenTelemetry tracer, and currently we export this information to a local Jaeger monitor (if available). To see the tracing output, run the Jaeger all-in-one container, and go to the web interface at http://localhost:16686.

Starting a scan

To start a scan you can do the following:

curl -X PUT http://127.0.0.1:5005/flyscan_xray_centre/start --data-binary "@tests/test_data/parameter_json_files/test_parameters.json" -H "Content-Type: application/json"

Getting the Runner Status

To get the status of the runner:

curl http://127.0.0.1:5005/status

Stopping the Scan

To stop a scan that is currently running:

curl -X PUT http://127.0.0.1:5005/stop

Writing out DEBUG logs

To make the app write the DEBUG level logs stored in the CircularMemoryHandler:

curl -X PUT http://127.0.0.1:5005/flush_debug_log