17280Re: [Clip] Re: Concept-based Search
- Jan 15, 2008
On Jan 15, 2008 9:23 AM, Flo <flo.gehrke@...> wrote:
--- In firstname.lastname@example.org, "ebbtidalflats" <ebbtidalflats@...>
A specific problem is how to deal with basic words (lemmata) and word
forms (flexions), also with compounds etc. So far, the clip searches
for whole words only (using \b in the Replace Command). Consequently,
the word list must contain the search term in any word form that it
is searched. If we make the clip search for substrings, we possibly
get some nonsense comments because the search term as a whole could
have a different meaning than a substring of that term.
"Build" an alternation regex? (on the fly, build a regex) Thus some
customization of the resultant regex can happen at each time that do a
Have/store the forms/compounds in their line with delimiter -- when needed,
put into an array.
(some of my syntax is not exactly correct but is hopefully close enough so
as to portray the meaning intended from or based upon my example).
example storage line:
To use, just find that doc for myword and select line/get line into array
then check, what is it, ^%forms0% (the array index) for how many forms or
compounds there are. Then,
well, I guess you'd need to loop and find one at a time and INC the array
index at each loop iteration
^!Find "^%forms%^%indx%" ris
Instead, if you were replacing:
^!Replace "^%forms1%|^%forms2%|^%forms3%" >> "whatever" riswa
| meaning "or" (any of those 3 get replaced)
Just an idea.
[Non-text portions of this message have been removed]
- << Previous post in topic Next post in topic >>