18 Vectorise matched glass descriptions

[georgerichardson]

closes #18

---

Checklist:

• [x] I have refactored my code out from notebooks/
• [x] I have run flake8 and addressed any linter erors
• [x] I have checked the code runs
• [ ] I have tested the code
• [x] I have run pre-commit and addressed any issues not automatically fixed
• [x] I have rebased onto dev (or merged any new changes from dev)
• [x] I have documented the code
  • [x] Major functions have docstrings
  • [x] Appropriate information has been added to READMEs
• [ ] I have explained the feature in this PR or (better) in output/reports/
• [x] I have requested a code review

[georgerichardson]

I haven't run any sanity checks on the results besides checking shape and that order is retained. Will do a manual check of the results before we merge. Small but not insignificant chance another model might work noticeably better

[georgerichardson]

There is now a quality assurance flow flow_qa.py which produces a chart and number that show the percent of descriptions that are not truncated by the maximum input length of the sentence encoder. It also produces a random sample of 100 company descriptions that are matched to their nearest neighbour by cosine distance according to the embeddings produced.