View Single Post
Old 04-21-2020, 12:06 PM   #510
NoSmint
Junior Member
NoSmint began at the beginning.
 
Posts: 6
Karma: 10
Join Date: Jan 2020
Location: Germany
Device: Kindle PW3
I did some extensive testing today and although there were a bunch of errors at frist, I was able to figure most of it out. Assuming I already have working AuthorProfile, Start/End-Profiles, my workflow for X-RAYs only was as follows:
  • Take an existing Original Amazon XRAY.entities.ASIN.asc (Jurassic Park in my case) and extract it into an XML File (5th Button to the right)
  • Edit the XML File to my liking, set aliases, rename, add or delete stuff
  • Use File as Source and locate the saved XML file
  • Build the X-Ray File, no aliases edited when prompted

Following these steps I ended up with a huge amount of "An error occurred while searching for start of highlight."-errors

Quote:
Originally Posted by Ephemerality View Post
This one happens when it thinks a paragraph contains the term (or its aliases) but it's not able to find where it is within the text. If you send me a copy of the book & the error log, I can have a look (in a PM via any upload site, Google Drive, or I can PM you my email if necessary, whatever works for you).
If it was in the German content you mentioned, it might be a UTF8/encoding issue, which has always been a pain for this stuff, but definitely fixable.
After some testing I was able to figure it out. It's all about Case matching. Jurassic Park contains the character "Tim Murphy", where I put "Tim" as an alias. Without the "Match Case" option checked, your script will find any words that contain the letters "tim" in that order, and believe me there are a lot in German . It does however not recognize the highlight since it's not the begenning of the word as it is not encapsulated in spaces. Hence the error as it finds the occurence but not the starting point. My solution was, to turn Case Sensitivity on for every Entry in the List. As a result all corresponding errors were gone instantly. Your searching script now ignored - as intended - all the occurences within other words. So my suggestion would be to make case sensitivity the standard option.

Quote:
Originally Posted by Ephemerality View Post
I've uploaded a new version (2.1.108) that should fix a bunch of the issues, the "create terms" UI is still kind of bad though, sorry about that.
Secondly - and this is a very special error that is very unlikely to ever occur because there's probably nobody using this, but - there's a problem with aliases for Settings (not Persons). I came across this because I have the German Version of the Book but there only is an Englisch Version of the X-Ray File. So I needed to rename all the places or give them aliases.

And now the bug, aliases for Settings won't get saved. A Person's alias is saved in the .aliases file oder "ext", a Setting's is not. Neither in that file nor in the XML. So it's basically ignored alltogether. I had to rename the Places instead of using aliases.

Once I figured those two things out I was able to create a flawless X-Ray with the positioning on the spot, every occurence found and no errors. Even in my special case where I used an English source for a German book.

Bottom Line: You might wanna look into saving aliases for Settings and Match Case vs Match Only. It was no encoding or UTF8 Issue after all. Also, if you could make the Terms Creator window and the terms lsit just beeing rezisable it would already make things a lot easier.
NoSmint is offline   Reply With Quote