:PROPERTIES:
:ID:       e6a2c650-ff59-4b72-b073-970731796888
:END:
#+title: Legacy search segment construction

Thoughts on constructing a segment representation from legacy data in the
[[id:11edd6c9-b976-403b-a419-b5542ddedaae][Subscriber Search Service]], referencing a [[https://gitlab.aweber.io/CP/Services/subscriber-search/-/merge_requests/15/diffs?commit_id=4d203dfec800cf49825d32a7ea663c4b862693bb][WIP commit]] loading legacy segments from
the database.

- The legacy segment stores field and operator information encoded in a search
  box ID, and a match value in a list view criteria field.
- Legacy segments only support a top-level AND grouping of criteria

Therefore, given a list of criteria on a segment, a Search can be formed as
follows:
#+begin_src python
  import dataclasses

  from subscribersearch import legacy, search

  @dataclasses.dataclass
  class Segment:
      """A saved search on a list representing a segment of subscribers.

      Likely belongs in the search module.
      """
      id: typing.Optional[int]
      name: str
      search: search.Search

      @staticmethod
      def from_dict(value: dict) -> Search:
          return Segment(
              id=value.get("id"),
              name=value["name"],
              search=search.from_dict(value["search"]),
          )

      def to_dict(self) -> typing.Dict[str, typing.Any]:
          return {
              "id": self.id,
              "name": self.name,
              "search": self.search.to_dict(),
          }


  # segment = fetch legacy segment info
  # rows = fetch legacy segment criteria
  segment = Segment(
      id=segment["id"],
      name=segment["name"],
      search=search.Search(
          list_id=segment["list_id"],
          group=search.Group(
              group_type=search.GroupType.AND,
              conditions=[
                  search.Condition(
                      field=filter.field,
                      operator=filter.operator,
                      match=match,
                  )
                  for filter, match in [
                      (legacy.filter_from_id(row["sb_id"]), row["lvc_criteria"]) in rows
                  ]
              ],
          ),
      ),
  )
#+end_src

By having functions that can retrieve a Segment object from the database and
save one back to it, we have an interface that we can then copy later for
storing and retrieving segments in an updated schema, and migrate between the
two.