fractal-analytics-platform / fractal-containers

Containers for testing Fractal
BSD 3-Clause "New" or "Revised" License
1 stars 0 forks source link

proof of principle: run a full example #9

Closed tcompa closed 11 months ago

tcompa commented 11 months ago

E.g. example 01 from fractal-demos.

tcompa commented 11 months ago

First glitch: I had no write access to /tmp/fractal_test/, and then I had to enter the container to download images.

tcompa commented 11 months ago

Second glitch: the workflow failed for a known issue of fractal-tasks-core (https://github.com/fractal-analytics-platform/fractal-tasks-core/issues/508).

tcompa commented 11 months ago

Second glitch: the workflow failed for a known issue of fractal-tasks-core (fractal-analytics-platform/fractal-tasks-core#508).

This will be fixed in the next fractal-tasks-core release, but for now I fixed this by hand:

cd /tmp/fractal_test/tasks/.fractal/fractal-tasks-core0.10.1/venv/bin
./pip3 install fsspec==2023.6
tcompa commented 11 months ago

Finally the workflow execution went through, great!

Note: the workflow was the JSON one stored in https://zenodo.org/record/8287221.

Screenshot from 2023-09-04 14-58-26

Screenshot from 2023-09-04 14-58-35

tcompa commented 11 months ago

Re-cap:

  1. I started the containers with make run
  2. I entered the fractal-server container to download images into the right folder (docker exec -it fractal-server bash, and then roughly this script to download images in /tmp/fractal_test/images).
  3. I triggered collection of fractal-tasks-core 0.10.1 from the web client. This takes a while, as usual.
  4. From the web client, I created the I/O datasets, with resource paths like /tmp/fractal_test/images/10.5281_zenodo.8287221 and /tmp/fractal_test/output.
  5. I downloaded the JSON workflow from https://zenodo.org/record/8287221.
  6. I imported the workflow into the DB, via the web client.
  7. I ran the workflow and (after a few attempts) it went through.
tcompa commented 11 months ago

This worked as expected - closing