This client process the data from the CNM polybot experiment and store it in a globus portal.
The simple polly flow:
- Data input : pollybot.json
- Transfer to ALCF Eagle
- Proc creates thermo result.
tools/thermo.py
- Proc creates images based on the json.
tools/create_images.py
- Proc gathers metadata and structure it for the portal.
tools/gather_metadata.py
- Publish metadata and images into the portal
conda create -n gladier python pip
conda activate gladier
git clone https://github.com/AD-SDL/gladier-polly
cd gladier-polly
pip install -r requirements.txt
On polly_client.py
change local_endpoint_id = 'cde22510-5de7-11ec-9b5c-f9dfb1abb183'
with your globus endpoint id.
For single client:
polly_client.py path/to/file.json
is a single shot client that creates a flow based on one single json file
For file watcher:
polly_watcher.py --dir path/to/folder
creates a filesystem watchdog on a given folder and invokes polly_client
when new json files arrive.
-
Create GCP at polly machine.
-
update
local_polly_endpoint
-
create
sample_folder
per json (done) -
update 'proc' functions
-
create polly_group at globus for access permission
-
add polly_group on globus_endpoint, flow_permissions
-
start new index
-
create new portal and facets
-
typo on the json
'ressitance_data'