search

bptlab / TracEX

This bachelorproject focuses on event log extraction from patient journeys using large-language models.
5 stars 5 forks source link

Feature/85 choose an activity key #111

Closed nils-schmitt closed 7 months ago

nils-schmitt commented 7 months ago

This PR will allow for the key after which the xes and dfg is structured to be selectable. I modified the handling of the dataframe in views.py, especially the confusing way of back and forth formatting of column names and the way of accessing the dataframe. Additionally, I removed the need for any .csv file and I fixed the bug, that the result view crashed when at any point the filters where so selective that no dfg is resulting and afterwards selecting again enough to show a dfg. (I can elaborate on this in the next meeting.)

nils-schmitt commented 7 months ago

In views.py I use a switch-case (or match-case in python). Unfortunately it was integrated in 3.10, which is why the other pylints fail. Should I change it to a less readable if-elif-else structure or do we ignore 3.8 and 3.9 from now on?

tkv29 commented 7 months ago

Regarding activity key First of all gj! I noticed 2 things:

  1. Instead using a selection field, could you please use radio buttons so the design matches with the other filters?
  2. On the results page when I want to refresh the filter my new selection of the activity label doesnt get refreshed
tkv29 commented 7 months ago

@nils-schmitt forget for now the first point, I will write to this later sth

nils-schmitt commented 7 months ago

Weird, does work on both of my machines... Try to pull again?

tkv29 commented 7 months ago

@nils-schmitt Tried it again, works like a charm, also the download and display of the xes! gj!

tkv29 commented 7 months ago

@nils-schmitt based on the pr description, you did #98?

cc: @FR-SON

tkv29 commented 7 months ago

regarding the web scraper, do I need to anything else except installing selenium and the webdriver? I got this error: image

nils-schmitt commented 7 months ago

You might need to install Firefox. As I wrote in the script, it is Firefox exclusive. Could be changed to being chrome exclusive, but to get an idea of the functionality i thought, this is enough

nils-schmitt commented 7 months ago

@nils-schmitt based on the pr description, you did #98?

cc: @FR-SON

Well, kind of, as there is no need for the csv in the code at all. But I dont query the dataframe from the database. I just take it from the orchestrator in the reusltview, as it is the same, that was initialized in the filterview...

tkv29 commented 7 months ago

I also have firefox installed but got the erros above. Maybe it makes sense to specify a path where the firefox file is? Seems like it cant find my firefox

The radio buttons looks good!

nils-schmitt commented 7 months ago

I think, that the errors come from you running the script off ubuntu. As the geckodriver has to be installed you might need a sudo

nils-schmitt commented 7 months ago

@tkv29 Do you want me to implement further changes? If not, please request reviewers.

tkv29 commented 7 months ago

Functionality looks good for me! GJ! Web Scraping will be outsourced in another PR