Automate entry quality assignment

[mmzdouc]

Problem: the MITE schema has a "quality" field that currently defaults to "medium" and is not included in the html generation. It would be nice to track the quality of the entry transparently and see how additional contributions/fixes improve the entry. This could be also shown on the html page.

Proposal: Define criteria to access the quality of the entry. The quality could be shown on the html page as stars. The criteria could be shown as tooltip to quickly indicate why an entry is in this quality level.

Potential criteria:

• validated reaction SMARTS
• balanced reaction (see also the discussion of https://github.com/mite-standard/mite_extras/pull/17)
• all cross-links present
• Evidence quality is sufficiently high (e.g. should be at least heterologeous expression to qualify for quality "high")

[mmzdouc]

There are two levels to quality: quality of the entry itself, and quality of the underlying primary literature. The quality should ideally only access the entry itself i.e. the reaction:

• does the reaction lead to all the product or are all products registered (e.g. 5 resulting products but only one was written down)
• are there any forbidden products stated
• the specificity of the reaction SMARTS (if it is so generic that it hits 10% of PubChem, that would be bad)

Currently not a priority to decide but could be something to implement in the future, e.g. to define quality levels to subset dataset for machine learning purposes