Investigate if Django 3.2 is possible for 1.0 release

[Ed McDonagh]

Django 2.2 is now out of long term support. See if it is easy to update to 3.2 before release.

Would involve updating other libraries too, and ensuring OpenREM works with Python 3.10. Will probably target Python 3.7 to 3.10

[Ed McDonagh]

Updated requirements to django 3.2, everything else free to get latest. Made some required model changes. Refs #940 Tests will fail, mostly on xlrd and pydicom.

→ <<cset 9b21ea3650c6>>

[Ed McDonagh]

xlrd was used to read exported xlsx files. Later versions are xls only, recommending openpyxl. Converted first test. Refs #940

→ <<cset addab7f9b421>>

[Ed McDonagh]

Fixed other xlrd tests. Added python version to pipelines. Refs #940

→ <<cset 6415ed699eed>>

[Ed McDonagh]

Hopefully fixing the last of the tests. Still need to address warnings from pydicom. Refs #940

→ <<cset 451d740574ff>>

[Ed McDonagh]

Removing remaining references to xlrd. Refs #940

→ <<cset ad7c5c36a8f5>>

[Ed McDonagh]

Black. Refs #940

→ <<cset 49b72e5bf1b6>>

[Ed McDonagh]

Pydicom warnings. End angles aren't DICOM tags, looks like included on assumption, not examples. Referring Physician Identification is actually a sequence, with info we are unlikely to want (address/telephone etc). Removed code wouldn't have worked. Refs #940

→ <<cset f9d3e14d8c33>>

[Ed McDonagh]

Pydicom warnings. Plural corrections, removing tags that don't exist in image headers. Refs #940

→ <<cset 578cc2535322>>

[Ed McDonagh]

Fixed various illegal VRs in the test DICOM files. Now using conformant UIDs when testing DICOM QR. Refs #940

→ <<cset 17d2e9843ddb>>

[Ed McDonagh]

Removing django-qssats-magic as no longer needed. Removing execute permissions on three dicom test files. Refs #940

→ <<cset 8665f54617cf>>

[David Platten]

Pinned plotly python package to 5.8.2 (the most up-to-date) and updated plotly javascript library to the correct version (2.12.1) for the python package. Switched rendering mode for scatter charts from webgl to svg due to rendering problem when there are many subplots. Refs issue #940

→ <<cset 5d2b344da2c6>>

[Ed McDonagh]

Fixed versions. Removed python-dateutil - I think it is only dependency of matplotlib and pandas, so will install anyway. Refs #940

→ <<cset 0ff76ff5165b>>

[Ed McDonagh]

Update of django-js-reverse file. Seems to work... Refs #940

→ <<cset e36501451e8e>>

[Ed McDonagh]

Adding ref #940 to changes.

→ <<cset a278afc8f920>>

[Ed McDonagh]

Duplicating tests but with Python 3.8 and 3.9. Refs #940

→ <<cset e3ab7f299772>>

[Ed McDonagh]

Add testing against 3.8 and 3.9 as custom steps. Refs #940

→ <<cset 24c275178fea>>

[Ed McDonagh]

Updating Docker to use Python 3.10. Also removing median function migration file - no longer used.. Refs #940

→ <<cset e2102834f634>>

[Ed McDonagh]

Docker didn't complete. Reduced pip installs to one. Reduces size. Refs #940

→ <<cset 65824aa67758>>

[Ed McDonagh]

Docker build was completing, issue was with not specifying tag. Refs #940

→ <<cset 13078d7f1e1c>>

[Ed McDonagh]

Removed the venv from Docker. Think it works, smaller, simpler. Refs #940

→ <<cset e9c92ed7bf18>>

[Ed McDonagh]

Merged in testDjango3-2Python3-10 (pull request #514)

TestDjango3 2Python3 10

Approved-by: David Platten

Fixes #940

→ <<cset cebfe3e8ab66>>

[Ed McDonagh]

Issue #907 was marked as a duplicate of this issue.