[pmwiki-users] timeline integration

Knut Alboldt pmwiki at alboldt.de
Sat May 12 03:42:00 CDT 2007


Hi Jan !

J. Meijer schrieb:
> 
> Hi Knut, long time no see. 
yes, stopped developing "my wiki" the last month and started to use it - 
quite successful. Ok and now I need some more extensions so these 
questions start again ...

>> Has anybody already got some experiences with timeline ?
> 
> I don't even know it. 
well, found out some more: it's a java application. there is also a 
javascript implementation, but that seems to use AJAX and a JSP-server, 
so it seems not so easy to implement it. Well, I'll have to have another 
look to my trials to implement jpgraph.

> Did you look at the new markupexpr.php recipy to see how Pm did it?

that's a good tip ! thanks, I'll study it.
> 
> I think regexes can do it, but I didn't fully understand your date formats, could you give some examples? 

I need a parsing of rather free entered date formats, e.g the user could 
enter in a page in one line:

date: 04.06.2007
  = format: dd.mm.yyyy
or
date: 4.5.2007 to 3.6.2007
or
date: 3.3. - 10.3.
  = format d.m.

and all should be parsed into two vars containing the start and an 
optional end-date in a fixed format (e.g. "year month day" so I can do 
some date calculation o it. I'll have a closer look to the regex's and 
try to parse it.


> Regular markups are processed line by line, so you can store the data with each line. The hard part is to decide when to output. The last record does not have a (:what:). Then again, a single blank line may indicate the end of a record? 
I think I could solve that, cause I'll have an "END"-markup, these 
markups are embendded in a kind of table
> 
> You may consider tagging the record with (:newrecord 1:) .. (:newrecord 2:) .. (:endrecord:), this way it is easier to find the end of a record and you can also specify the record name just once. 
> 
> In order to process the text before it is broken into lines, put in the markup near the beginning of the sequence. (Maybe) use Markup('name','fulltext',etc). 
oh, yes.
> 
> 
>> Or is it better to use pagetextvariables ? 
> 
> You should name them to have different names. Fe:
> 
>   what1: text1
>   date1: dateinfo1
>   who1: person1
> 
>   what2: etc
> 
> It would certainly be easier on the eyes. Probably a good idea. 
yes but very complicated if you add or move some entries, there could be 
up to 50 so you have to check it every time.
> 
> Maybe this would be best for you:
> 
>  (:record1:
>   what: text1
>   date: dateinfo1
>   who: person1
>  :)
> 
> You could then dissect the what/who/date fields with markup expressions from {$:record1}. 
that sounds relay fine, good tip ! I think I'll prefer this solution


Thanks a lot ! I'll try to implement some more and will report.

Knut



More information about the pmwiki-users mailing list