xchem / fragalysis-stack Goto Github PK

• Simple view of 2D ligands - click and see protein.
• Wonka - but without analysis

I'm starting an xchem/fragalysis-stack:2.0.0 container image and there are (imo) too many warnings relating to the data model and subsequent migrations.

So that important warnings can be seen (and fixed) can we adjust the model so that the stack start-up proceeds without any warnings?

The log of the initial startup is attached.

stack-2.0.0.log

• Loader from output from Proasis
• Include ability to load from scratch

I'm launching xchem/fragalysis-stack:2.0.0"cold", i.e fresh deployment, empty database.

Getting django.contrib.auth.models.DoesNotExist: User matching query does not exist exception during the initial migration, which takes place as the stack container starts.

Although the stack container images starts there are exceptions during the data migration stage which look serious and probably should be fixed. The full stack log is attached but the first exception is shown below: -

2020-09-22T13:04:56.701971214Z Operations to perform:
2020-09-22T13:04:56.702040357Z   Apply all migrations: viewer
2020-09-22T13:04:56.722079828Z Running migrations:
2020-09-22T13:04:56.982359684Z   Applying contenttypes.0001_initial... OK
2020-09-22T13:04:57.795490666Z   Applying auth.0001_initial... OK
2020-09-22T13:05:02.496417457Z   Applying viewer.0001_initial... OK
2020-09-22T13:05:02.75391939Z   Applying viewer.0002_target_zip_archive... OK
2020-09-22T13:05:03.200757414Z   Applying viewer.0003_auto_20200804_1241... OK
2020-09-22T13:05:03.344376673Z   Applying viewer.0004_auto_20200804_1411... OK
2020-09-22T13:05:03.633734733Z Traceback (most recent call last):
2020-09-22T13:05:03.635918826Z   File "/usr/local/lib/python3.7/dist-packages/guardian/management/__init__.py", line 32, in create_anonymous_user
2020-09-22T13:05:03.635958956Z     User.objects.using(kwargs['using']).get(**lookup)
2020-09-22T13:05:03.635971365Z   File "/usr/local/lib/python3.7/dist-packages/django/db/models/query.py", line 431, in get
2020-09-22T13:05:03.636128109Z     self.model._meta.object_name
2020-09-22T13:05:03.636159122Z django.contrib.auth.models.DoesNotExist: User matching query does not exist.

The exception causes other exceptions, as you see During handling of the above exception, another exception occurred:

• psycopg2.errors.NotNullViolation: null value in column "last_login" violates not-null constraint
• django.db.utils.IntegrityError: null value in column "last_login" violates not-null constraint

There are also a rather large number of data warnings so an inspection of the data migrations looks necessary.

stack-2.0.0.log

Migrate/convert the CI actions away from Travis and introduce GitLab Action CI Workflows.

Launching a cold/fresh stack (v2.0.0)

Although the stack appears to start, some way through the initialisation (migration) stage, the following can be seen in the log (the full log is attached)...

2020-09-22T13:05:09.179452667Z usage: manage.py migrate [-h] [--noinput] [--database DATABASE] [--fake]
2020-09-22T13:05:09.179489471Z                          [--fake-initial] [--plan] [--run-syncdb] [--check]
2020-09-22T13:05:09.179495954Z                          [--version] [-v {0,1,2,3}] [--settings SETTINGS]
2020-09-22T13:05:09.179500455Z                          [--pythonpath PYTHONPATH] [--traceback] [--no-color]
2020-09-22T13:05:09.17950488Z                          [--force-color] [--skip-checks]
2020-09-22T13:05:09.179509212Z                          [app_label] [migration_name]
2020-09-22T13:05:09.179513261Z manage.py migrate: error: unrecognized arguments: hotspots

If hotspots has no migrations then maybe the migration command should be removed?

stack-2.0.0.log

• Working on RHEL7 basic Django app

Running the latest stack (2.0.4) I noticed the following in the migration log: -

2020-09-25T08:27:31.686131672Z CommandError: Cannot find a migration matching 'hypothesis' from app 'scoring'.

xchem/fragalysis#10

Hi,

I installed docker and executing sudo docker-compose up and it shows me errors ...

web_1 | File "/rdkit/rdkit/Chem/init.py", line 26, in
web_1 | from rdkit.Chem.inchi import *
web_1 | ImportError: No module named inchi

Can you point out what is going on ?

When building the fragalysis stack for the second and subsequent times an error in the Dockerfile becomes apparent.
It fails trying to run a mkdir command because the directory already exists.

The following workaround around line 34 fixed the problem:
RUN mkdir -p ${APP_ROOT}/static

NB: Suggestion from Alan: While fixing this, add to the end of lines && \ to link all the run commands together for efficiency.

With the new "Staging Replica" (based on Production media and DB) the automated 'latest' deployment does not work. This now needs to be disabled as staging deployments are now handled by individual users with access to the Production AWX server.

Launching a fresh 2.0.0 stack

Some way into the initialisation the following error is observed (full log attached): -

2020-09-22T13:05:08.167392945Z CommandError: App 'xchem_db' does not have migrations.

I suspect that, for a clean stack startup, we should not see any Errors.

stack-2.0.0.log

• Models and View for exit vectors (2D and 3D) and visualise in NGL