Hi Etienne,

Exactly ... Use the current api and when we refractor the scraper we just have 
to change all usages equally.

Chris
________________________________
Von: Etienne Robinet <[email protected]>
Gesendet: Freitag, 17. April 2020 20:27
An: [email protected] <[email protected]>
Betreff: Re: Camel Component features

Hi Chris,
So what do tou think? Adapt the old scraper to use in the API and then use this 
in the Camel component?

Etienne

> Le 17 avr. 2020 à 12:10, Christofer Dutz <[email protected]> a écrit :
>
> Hi all,
>
> And we should really rewrite and clean up the scraper internally ;-)
>
> Still got that on my ToDo, but would be happy if someone would beat me to it 
> ;-)
>
> But I also agree that we shouldn't re-invent the wheel and reuse the 
> components we already have.
>
> I think from an API side even a new scraper would be pretty much like the 
> current one and the current one does work fine.
>
> Chris
>
>
> Am 17.04.20, 09:45 schrieb "Julian Feinauer" <[email protected]>:
>
>    Etienne,
>
>    like so often you are absolutely right with your analysis oft he Situation 
> : )
>    Indeed these are very valuable features.
>
>    In plc4x there exists a compontent called the "Scraper" (which I wrote 
> initially) and which was extended to a "TriggeredScraper" by @Tim Mitsch 
> which does exactly that.
>    Watch a bit or a condition and only fetch data, when this condition 
> changes.
>
>    In fact we want (since months) integrate our in-house developed Engine for 
> these Tasks, CRUNCH as a PLC4X Subproject to push this even further (see here 
> for CRUNCH https://github.com/pragmaticminds/crunch).
>
>    I guess it helps you to take a look at both resoures!
>
>    Julian
>
>    Am 17.04.20, 09:21 schrieb "Etienne Robinet" <[email protected]>:
>
>        Hi everyone,
>        As the Camel component is now working correctly, I thought about some 
> features we could add on it, based on work I read from a coworker. I would 
> like to know if these would fit inside this project and if so I would be glad 
> to implement them.
>
>        Trigger & Completion
>        To prevent writing to a PLC that is using data or to fetch a bunch of 
> data that has not changed, we could use Flags (Bits). A transaction will only 
> take place if the Trigger bit is 1, and once it is done, it's set back to 0 
> and the completion bit to 1 (or Error bit). This way, the PLC is the master 
> of the transaction and can tell when it can be written to or read from.
>
>        Polling
>        In order to be aware of the change of state of the trigger, we should 
> do a polling on this bit. This way, we only read 1 bit, at an user specified 
> interval, and once the bit is 1, we do the transaction(read/write)
>
>        Default values
>        If none or not every of these parameters are given, then we could use 
> default values like reading the tags given every 1000ms or so (e.g. for 
> monitoring)
>
>        Hope this is interesting for you and if so I would start next  week.
>
>        Etienne
>
>
>
>

Reply via email to