updates for mongo embedded #22
Conversation
| @@ -118,14 +118,14 @@ def get_header_field(field): | |||
| if field[0:6] == 'count_': | |||
| return 0 | |||
| else: | |||
| return None | |||
| return {} | |||
There was a problem hiding this comment.
Under what circumstances would this codepath be hit?
There was a problem hiding this comment.
example could be. If a run doesn't have a stop document and you ask for the stop document
intake_bluesky/mongo_embedded.py
Outdated
| @@ -278,7 +278,8 @@ def search(self, query): | |||
| MongoDB query. | |||
| """ | |||
| if query: | |||
| query = {f"start.{key}": val for key, val in query.items()} | |||
| query = {f"start.{key}".replace("start.$", "$", 1): | |||
There was a problem hiding this comment.
This is sufficient to pass the tests, but our test coverage is lacking. What about a query like {'$and', [{'detector': 'pe1}, {'motor': {'$in': ['motor_A', 'motor_B']}}]}? I think the general solution here will either be to have separate RunStart collection that we treat as a disposable cache or discover a general strategy for transforming queries that are constructed to operate on a collection into queries that operate on a sub-field of documents in that collection.
There was a problem hiding this comment.
I think I added a decent general solution.
…use volatile_db anymore
|
Power-cycled to re-run tests. |
No description provided.