feat: --skip-validation allows mismatched Augur versions

[corneliusroemer]

Description of proposed changes

Currently, files produced by Augur contain the version of the Augur instance used to produce the data file.

When there's a major version mismatch between the version a file was produced with and the version of Augur export, Augur throws an error complaining about version mismatch.

While it's a good thing to warn the user, there are situations where one may want to take the risk and allow Augur export to use output from different versions.

This PR extends the meaning of the --skip-validation argument to also not complain about version mismatch in input files.

--skip-validation was introduced in #865

Related issue(s)

Fixes #901
Related to #865

Testing

I added integration tests

[codecov]

Codecov Report

Merging #902 (2be8e8c) into master (a49914a) will increase coverage by 0.02%.
The diff coverage is 100.00%.

❗ Current head 2be8e8c differs from pull request most recent head 61f2ae0. Consider uploading reports for the commit 61f2ae0 to get more accurate results

@@            Coverage Diff             @@
##           master     #902      +/-   ##
==========================================
+ Coverage   59.64%   59.67%   +0.02%     
==========================================
  Files          53       53              
  Lines        6317     6319       +2     
  Branches     1586     1586              
==========================================
+ Hits         3768     3771       +3     
+ Misses       2286     2285       -1     
  Partials      263      263              

Impacted Files	Coverage Δ
augur/export_v2.py	69.04% <100.00%> (+0.15%)	⬆️
augur/util_support/node_data_file.py	100.00% <100.00%> (ø)
augur/util_support/node_data_reader.py	100.00% <100.00%> (ø)
augur/utils.py	64.25% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a49914a...61f2ae0. Read the comment docs.

[huddlej]

Thank you, @corneliusroemer! I had a couple minor comments about the tests, but otherwise good to merge. (Thank you for adding tests!)

[huddlej]

I agree it's annoying to pass around this variable through multiple classes. This isn't ideal OO design and an OO implementation is even what we need here. Your implementation here seems fine, in the current context. We should leave to another PR to reimplement the node data reader as a function again. 😄

[jameshadfield]

Yes!

[jameshadfield]

I really needed this today so I took it upon myself to implement John's PR suggestions & also rebased it onto current master (99 commits ahead!). I think this should be merged 🙏

[jameshadfield]

Test failure is the same as seen on another PR: #961 (comment) and seems unrelated to this PR 😕

[victorlin]

This was bumped in Slack. Force-pushed a rebase on latest master, then I'll merge once tests pass.