ENH: Add subclassing for trips #518
Conversation
for more information, see https://pre-commit.ci
Codecov ReportPatch coverage:
Additional details and impacted files@@ Coverage Diff @@
## master #518 +/- ##
==========================================
+ Coverage 93.05% 93.31% +0.25%
==========================================
Files 34 34
Lines 2074 2153 +79
Branches 388 409 +21
==========================================
+ Hits 1930 2009 +79
Misses 130 130
Partials 14 14
☔ View full report in Codecov by Sentry. |
NinaWie
left a comment
There was a problem hiding this comment.
I had a look, thanks for the great work on this complicated issue due to the DataFrame / GeoDataFrame fallbacks. In response to your comment, I would still prefer to have a dedicated class Trips, so that the user does not need to care about the dataframe / geodataframe stuff. Some of my comments are related to this. Before changing everything to having a Trips class, I would however wait whether @hongyeehh or @henrymartin1 agree to that.
NinaWie
left a comment
There was a problem hiding this comment.
Thanks for the explanations and changes, from my side, it is fine now :)
hongyeehh
left a comment
There was a problem hiding this comment.
Thanks for the additional explanation. This is also ready from my side. Great effort!
Adding the Trips classes to trackintel.
I changed the accessor to the
tripsfunction to create either theTripsDataFrameorTripsGeoDataFrameclass, depending on whether ageocolumnis present or not.When I implemented these changes, I sometimes thought it would be easier to implement a single class, aka a
Tripsclass with no fallback mechanism, as Nina suggested in a meeting. So that it remains aTripsclass even if it loses the geometry column. I would be open to discuss if this alternative implementation would be more appropriate.In another PR, I should add documentation explaining this split into two classes and the complicated fallback pattern.