Query DSL representation about imbo-metadata-search

[kbrabrand]

My vote goes to number three. It looks a bit messy now, but adding a few use statements will shorten the code considerably. I think the slightly more rigid parsing and rule set is a good thing in this case.

We need to at least implement the $exists operator and a wildcard operator was mentioned, but I'm not sure what operator was used for wildcards in the end. Should check the implementation by @christeredvartsen and implement the same set of operators.

[fangel]

Yes. Mongo already has support for $exists in their Query DSL, so our subset will just have to include this too.

The Mongo Query-DSL does not have (as far as I know) support for a $wildcard operator, so technically speaking, our DSL won't be a subset of Mongos - but hey. $wildcard can however be emulated using $regex, so Mongo does support it, albeit with a different syntax. So it's fine for us to have it, as all of our search backends can support it (SQL would likely have to be some sort of LIKE, which is slow but functional).

[kbrabrand]

I don't know how you planned to implement the AST building and handling, but I looked into a few different alternatives and Dissect seems to provide a pretty straight forward way of building and working with ASTs. Just a thought.

[fangel]

Closing this issue, as we now have a a good start to an implementation that follows option 3.