navigating-stories / orange-story-navigator

Add-on to the Orange3 data mining toolkit with text processing widgets from the project Navigating Stories
https://research-software-directory.org/projects/navigating-stories
Other
2 stars 2 forks source link

text of the stories changes when running the analysis #43

Closed f-hafner closed 5 months ago

f-hafner commented 6 months ago
Story Navigator version

0.0.18

Orange version

3.36.2  orange-network version: 1.9.0 orange-text version: 1.15.0

Actual behavior

When I open the actors widget, the story I am using is displayed complete. As soon as I check the boxes for highlighting the subjects, other potential actors etc. parts of my story disappear.

Expected behavior

I expect the text to be displayed complete

Steps to reproduce the behavior
  1. import file
  2. select elements widget
  3. connect file with ‘elements’
  4. connect ‘elements’ with data table
  5. select actors widget
  6. connect elements to actors
  7. connect story to actors
  8. open actors widget --> story is complete
  9. have subjects highlighted --> parts of the text disappear
  10. Additional info (worksheets, data, screenshots, ...)

General screenshots

image image
kodymoodley commented 6 months ago

Need more info about which sentences are missing / removed and in which stories / data.

kodymoodley commented 6 months ago

Saartje will send sentences for "eenzaamheid" / "jalouzie".

f-hafner commented 5 months ago

@kodymoodley , I have trouble reproducing this. I tried with the "eenzamheid" and the "overleven" stories. I copy-pasted the displayed text from the Actors widget into a text file and ran a diff between them. For "eenzamheid", I find 0 difference. For "overleven", it's hard to see full words/sentences missing. There are sometimes additional letters at the start of a line or additional white spaces around commas.

f-hafner commented 5 months ago

double-checked again with Thijs. when I open story 1 in docx format, the actors widget already does not show the full text when I open it in (before doing any highlighting).

unclear what is going on. needs more investigation.

kodymoodley commented 5 months ago

fixed by #58