search

pm4py / pm4py-core

Public repository for the PM4Py (Process Mining for Python) project.
https://pm4py.fit.fraunhofer.de
GNU General Public License v3.0
722 stars 286 forks source link

"tuple has no attribute week" in view_events_distribution_graph #409

Closed AndreasKarasenko closed 1 year ago

AndreasKarasenko commented 1 year ago

When using view_events_distribution_graph with dist_type="week" pm4py executes "x.isocalendar().week" where x is a pandas.Timestamp object.

However pandas.Timestamp.isocalendar() returns [int, int, int] which then throws the above error.

Suggestion: change line 117 in pm4py\statistics\attributes\log\get.py from: values = Counter(x.isocalendar().week for x in timestamp_values) to: values = Counter(x.isocalendar()[1] for x in timestamp_values)

fit-alessandro-berti commented 1 year ago

Dear @AndreasKarasenko

Thank you for signaling. We will fix this in the next release.

fit-alessandro-berti commented 1 year ago

Errata corrige: the allowed distribution type are just days_month, months, years, hours, days_week, weeks

With weeks it works

AndreasKarasenko commented 1 year ago

Sorry, there was a typo in my initial comment, when I encountered the problem I did use the correct distr_type.

In this sense the problem persists on my end (without the mentioned fix). I suppose that there might be a mismatch in the type of timestamp in my logs and the ones expected by pm4py (can you tell me which one you expect?). I will come back to this issue next week with a reproducible example. For now changing the line above as I mentioned fixes it for me locally. With the type of timestamps I have (see below) I could also drop the call to isocalendar(), which also fixes it.

Additional Information: pandas version: 2.0.1 Timestamp type: pandas._libs.tslibs.timestamps.Timestamp