Matching \w marked words with glossary entries

Question

Is there a checking tool that will run through all the \w ... \w* marked words and check if they are (or more importantly, aren't) present in the glossary?

--------------------------------------

Less important, is there a tool which will reverse this process? I.e., find words in the text which probably ought to be marked with \w based on their presence in the glossary?

Answer 1

The Paratext markers check in PT8 should verify that a term marked with \w ...\w* is in the Glossary (this no longer works in PT9, but has been reported and a solution is being examined). If the word is a form of the glossary entry then the mark up should include the alternate glossary entry - \w word|entry\w*

There is no tool that I'm aware of that identifies words that probably should be marked since any given glossary entry could have literally hundereds of occurences or forms in the text.

Here is a quick tip video on linking terms to the glossary through the Biblical Terms tool

Comments

The Paratext Markers Check is not working for me. For example, I typed \w bob\w* into the text of an otherwise Arabic script project, thinking there's no way that "bob" is in our glossary. But when I ran the markers check, it didn't say anything about that particular verse.

Some background might help, though I don't believe it's the main problem.

We can't use PT's "Bib Terms tool to Glossary linking" because PT can't handle all our glossary entries being in \c 1 of the Glossary. Quite a long time ago we found that PT was becoming completely unresponsive as it tried to display the (currently 253) glossary entries, so we split it up into chapters. I understand that the linking won't work when the glossary entries are in different chapters, and can't think of a viable way that it could be designed to work, so I'm happy to not complain about that loss of functionality.

But I don't see how or why that would affect connecting \w with \k in the glossary. The glossary is essentially a large text document, so the fact that some lines have \c 2 or \c 3 on them shouldn't affect that kind of check. But, as I said, the markers check isn't telling me that \w entries aren't in the glossary. (In fact, if PT is simplistically searching only \c 1 for some reason, I'd expect it to warn me about 100% of the entries since none of the \k entries are in \c 1.)

Answer 2

I miswrote in saying that Paratext will verify if a word marked with \w...\w* is in the GLO book. This was true in version 8, but no longer works in PT9. One option to identify the words that are marked with \w...\w* would be to use a regular expression either in the Find dialog or better in the RegExPal tool. One expression that will find the words is:

regex:\\w .*?\\w\*

If you use the expression in RegExPal you can use the Count/Extract tool to sort the items and then compare to the GLO.

Comments

Thank you. Using regex seems like a good way to do it. Were you thinking I'd do a similar search on \k in the glossary, and then compare the two lists? Would that comparison be done outside of PT, by hand or using some other tool? (If so, I would probably use WinMerge or python.) I can't see any way in RegexPal to add/remove items from a list like you can within the regular PT search.

Using

(?<=\\w\s).*?(?=\\w\*)

gives me a list without the \w markers, which isn't a big deal either way. But ideally I'd like the list only with the terms after any | dividers. Does anyone with better regex skills than me know how to do that?

---

(?<=\\w\s.*?\|).*?(?=\\w\*) should find all of the terms that follow the |

(?<=\\w\s)[^|]+?(?=\\w\*) can be used to find terms that don't have any |

Yes, you could also search for the words in the GLO that are marked with \k...\k* and compare the two lists.