Loan statistics #1256
Loan statistics #1256
Conversation
|
|
||
| if creation_date and start_date: | ||
| waiting_time = (start_date - creation_date).days | ||
| stats["waiting_time"] = waiting_time if waiting_time >= 0 else None |
There was a problem hiding this comment.
Isn't it better to return 0 than None?
If we want to distinguish self-checkout from normal loans, this won't be enough (since a normal loan can be retrieved on the same day).
There was a problem hiding this comment.
I just wanted to make sure there are no negative values, which would skew the statistics. Like when doing a sum over the waiting time and there is one loan where the start_date was manually set before the creation_date I don't think it should influence the waiting time statistic. But it might be good to have a discussion about how to handle such cases?
Regarding distinguishing self-checkout and normal loans: one can filter with the q parameter for the delivery option.
| description="Each group_by item must be a dict with 'field' key" | ||
| ) | ||
|
|
||
| _validate_os_field_name(grouping["field"]) |
There was a problem hiding this comment.
Would it be possible to have an allow list instead of a regexp?
There was a problem hiding this comment.
That would surely be possible, but might limit in how the endpoint can be used and it requires more work from our side to keep track of all possible fields.
The histogram will be implemented for multiple record types and adding+maintaining an allow list might be tedious.
Also the endpoint is behind the backoffice permission, so I do not think it is too critical?
| actions = [action for action in self.actionsiter()] | ||
| res = search.helpers.bulk(self.client, actions, stats_only=True, chunk_size=50) |
There was a problem hiding this comment.
Could you maybe clarify what this part does?
There was a problem hiding this comment.
I added a comment. Hope that makes it more clear
JakobMiesner
force-pushed
the
feature/stats/loan-stats
branch
2 times, most recently
from
5a988e2 to
6bce595
Compare
|
|
||
| # Document availability during loan request | ||
| stat_events_index_name = "events-stats-loan-transitions" | ||
| if current_search_client.indices.exists(index=stat_events_index_name): |
There was a problem hiding this comment.
I think it would be better to fail rather than silently not index things if this conditional statatement is not fulfiled. ping @ntarocco WDYT?
There was a problem hiding this comment.
for context: the check was added because if there is no event at the moment a loan is indexed, there will be no events index and the loan indexer will throw an error in the lines below the if, leading to not/wrongly indexed loans.
I could manually create the index in our systems during deploy and add it to the release notes that the stats index needs to be created for the system to work. But then the setup script should also create the index for new systems.
There was a problem hiding this comment.
Agree, if it is required before running this code (weird, shouldn't it be created automatically if not existing?), then I would fail instead.
| try: | ||
| parsed_args = schema.load(request.args.to_dict()) | ||
| except ValidationError as e: | ||
| raise InvalidParameterError(description=e.messages) |
There was a problem hiding this comment.
is this returning json error in REST?
There was a problem hiding this comment.
It is. InvalidParameterException inherits from class RESTException.
* the endpoint returns a histogram for loans where requested metrics are grouped by and aggregated * extend the loans with a new stats object * the stats object contains `waiting_time`, `loan_duration` and `available_items_during_request`
JakobMiesner
force-pushed
the
feature/stats/loan-stats
branch
from
6bce595 to
b82b815
Compare
4 participants
waiting_time,loan_durationandavailable_items_during_request(closes: Library Statistics - 2. & 3. - Average duration of loan & Availability of requested documents CERNDocumentServer/cds-ils#1029)tests are failing because of dependency on this PR