Jane Austen's peculiarity

[Richmond]

On 09/08/15 00:26, Peter M. Brigham wrote:
> On Aug 8, 2015, at 4:51 PM, Richmond wrote:
>
>> On 08/08/15 23:33, Peter M. Brigham wrote:
>>
>> <snip>
>>> Missing an "of" in the two lines above:
>>>     put line textLine *of* $TEKST into line cookedLine of fld "COOKED" etc
>>> Don't know if that's the problem.
>> <snip>
>>
>>> Your script logic seems unnecessarily complex. Since it looks as if only the last occurrence is ending up in the output field, instead of using a counter to keep track of the next line in the field, you could just
>>>     put cr & line textLine of $TEKST after fld "COOKED"
>>> But once again, loading a line into a field repeatedly will be much slower than putting it into a variable in the repeat loop and then  putting the variable into the field just once when the repeat is done. Getting or putting something from or into a field is much slower than doing the same in a variable, so just do it once.
>>>
>>> Also, I can see no reason to be loading your data into system variables, which is what "$WERBS" etc is defining. The only reason to put something into a variable beginning with "$" is if you want some other system process besides LC to be able to access the data.
>>>
>>> -- Peter
>>>
>>>
>> Well, as per your suggestion I did this:
>>
>> on mouseUp
>>    put 1 into textLine
>>    put fld "WERBS" into WERBS
>>    put fld "TEKST" into TEKST
>>    repeat until line textLine of TEKST contains "finalSolution666"
>>       put textLine into fld "KOUNT"
>>       put 1 into verbLine
>>       repeat until line verbLine of WERBS is empty
>>          put line textLine of TEKST into fld "LYNE"
>>          put line verbLine of WERBS into WERB
>>          put "were" && WERB into FRAZE
>>          put FRAZE into fld "FRAZE"
>>          if line textLine of TEKST contains FRAZE then
>>             if fld "COOKED" is empty then
>>                put line textLine of TEKST after fld "COOKED"
>>                 -- this is here so that line 1 of fld "COOKED" does not end up empty
>>                else
>>                   put cr & line textLine of TEKST after fld "COOKED"
>>                   end if
>>          end if
>>          add 1 to verbLine
>>       end repeat
>>       add 1 to textLine
>>    end repeat
>> end mouseUp
>>
>> but still get only the last value.
> Well, your logic still makes my head hurt, too many counters.

My logic is based on the belief that one has to keep count of the lines 
in the Ur-text, the verb list, and the output field.

It is also derived from a program I wrote to make concordances in PASCAL 
in 1985.

The idea of using a carriage-return is very useful (haven't actually 
thought about those since my BA, as typed all my work on an Olivetti
portable which my Mum was given by her Mum and Dad for her 21st birthday).

>   Here's what I'd do, using a variant of my original function since it appears that you want to list the lines the relevant phrases occur in, not just the isolated phrases.

Isolated phrases are not much use for subsequent analysis; i.e. to see 
which collocations they occur in, context, and so on.

>
> function findWere pText
>    -- returns a comma-delim list of all the line offsets matching "were *ed"
>    --    or "were" && <a word in your preterite list>.
>    put fld "WERBS" into pretList
>    put wordOffsets("were", pText, true) into offList
>    repeat for each item w in offList
>       put word w+1 of pText into testWord
>       if testWord ends with "ed" or testWord is among the words of pretList then
>          put (the number of lines of word 1 to w of pText) & comma after outList
>       end if
>    end repeat
>    return char 1 to -2 of outList
> end if
>
> then:
>
> on mouseup
>    put fld "TEKXT" into tText
>    put findWere(tText) into linesList
>    repeat for each item i in linesList
>      put line i of tText & cr after relevantLines
>    end repeat
>    put char 1 to -2 of relevantLines into fld "COOKED"
> end mouseup
>
> Untested, but you get the idea.

I do.

Thanks.

I shall put your code in another button and see what happens.

>
> -- Peter
>
>

Richmond.