Details
-
Type: extRequest
-
Status: Closed
-
Priority: Major
-
Resolution: Done
-
Fix Version/s: 2021
-
Component/s: FIWARE-TECH-HELP
-
Labels:None
-
HD-Chapter:Data
-
HD-Enabler:CEP
Description
Hi Avishai,
Glad you are using the CEP.
I'm the contact person for the CEP in FIWARE.
Could you please send me the CEP definition file (using the export to
file)? this way I can see if there is a problem in the definition file.
Please cc Jira (as I did) to document this corresponding in the FIWARE
issue tracking system.
And yes, if you need more help, we can set a phone call.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Date: 04/02/2015 05:56 PM
Subject: fi-ware CEP short quastion
Good evening Tali,
My name is Avishai and i am working at Oviva a start-up company which
located in Switzerland.
We are working on project which use the CEP on our system (cause we are
applying to one of the Fi-ware accelerators) .
I have one question which i can't understand from the documentation,
if i use an EPA type: "sequence" and add two event from the same type with
two different aliases ,
and i just put condition like that : alias1.id == alias2.id
i thought it will check if two event with the same Id happens in the
context,
But at the moment i sent one event i immediately get the derivation event
.
What i am doing wrong ?
Is it possible to talk with someone that have experience with this system
?
(By the way i am an Israeli and i speak Ivrit so please i will be glade to
talk with some one from your team to clear some points about this system )
Best Regards,
Avishai Gurt,
Oviva Team.
-
- hello_window_eventInitiator.json
- 3 kB
- FW External User
-
- hello_window_startup.json
- 2 kB
- FW External User
-
- hello.json
- 2 kB
- FW External User
-
- hello.json
- 2 kB
- FW External User
Activity
Hi Avishai,
I looked at your definition file and all seems to be well defined.
Sequence EPA implements <= on the time of the input events. Hence, in the
special case where you define sequence of two events from the same type
(and no other expression that distinct between them), the first event can
play as the first operand and as the second operand, and the EPA will
detect a sequence even if only one event arrived. A solution for it can be
to add in the condition expression:
message1.userId == message2.userId AND
message1.DetectionTime<message2.DetectionTime
As for your other questions, there is an explanation in the user guide.
Please tell me if the explanation there is not sufficient.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Cc: jira-help-desk@fi-ware.org
Date: 05/02/2015 10:32 AM
Subject: Re: fi-ware CEP short quastion
Hi Tali
I attach the definition file to this mail .
Can you please also explain to me in the EPA definition in the Event
Selection tab
What is the meaning of "instance selection" and "consumption" ?
Best Regards,
Avishai Gurt,
Oviva Team.
On 5 February 2015 at 08:05, Tali Yatzkar-Haham <TALI@il.ibm.com> wrote:
Hi Avishai,
Glad you are using the CEP.
I'm the contact person for the CEP in FIWARE.
Could you please send me the CEP definition file (using the export to
file)? this way I can see if there is a problem in the definition file.
Please cc Jira (as I did) to document this corresponding in the FIWARE
issue tracking system.
And yes, if you need more help, we can set a phone call.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Date: 04/02/2015 05:56 PM
Subject: fi-ware CEP short quastion
Good evening Tali,
My name is Avishai and i am working at Oviva a start-up company which
located in Switzerland.
We are working on project which use the CEP on our system (cause we are
applying to one of the Fi-ware accelerators) .
I have one question which i can't understand from the documentation,
if i use an EPA type: "sequence" and add two event from the same type with
two different aliases ,
and i just put condition like that : alias1.id == alias2.id
i thought it will check if two event with the same Id happens in the
context,
But at the moment i sent one event i immediately get the derivation event
.
What i am doing wrong ?
Is it possible to talk with someone that have experience with this system
?
(By the way i am an Israeli and i speak Ivrit so please i will be glade to
talk with some one from your team to clear some points about this system )
Best Regards,
Avishai Gurt,
Oviva Team.
[attachment "hello.json" deleted by Tali Yatzkar-Haham/Haifa/IBM]
Hi Tali ,
Thank you very much for your answer ,
The point that you raised about sequence EPA and the way the evaluate the
first event that exactly why i was a bit confuse.
(I even debugged the code and saw that the CEP evaluate in that way, so
that the reason i was thinking i am doing some thing wrong)
about the other question it's OK i found it in the guide but regarding the
consumption definition i didn't get it, in the guide write :
*Consumption – defines the condition for events to be later reused in the
same situation. This*
*operand is applicable only for join operators, aggregation operators, and
selection operators.*
What dose it means "reused" it's happened why should i reused it ?
I have another small question regarding sliding windows in temporal context
:
if i define sliding window : window duration 100 and sliding period 100 is
that mean that every 100 ms i get a new context ?
cause i try that and the EPA didn't detect anything .
I attach to this mail my definition this definition works fine with
temporal interval but at the moment i switch to sliding window it doesn't
detect the event .
Thanks again,
Avishai Gurt,
Oviva Team.
Hi Avishai,
1. Consumption - let's say you have an EPA with sequence of two event
types (A, B).
Now you received two input events: A1, B1.
Then the EPA will detect the situation (based on A1,B1).
Now you receive another input event from type B - B2
If A is set to be with consumption "Reuse" then A1 can be used again, and
the EPA can detect the situation again based on A1, B2.
If A is set to be with consumption "consume" then after the EPA used A1
.to detect the situation based on (A1,B1), A1 can't be used again, and
when B2 arrives, there will be no situation detection. Is it clear?
2. SlidingWindow - your understanding is correct. The file you sent me is
with an error (not related). If you fix the error, and you still see a
problem with the sliding window, please let me know. If this happen, I
will open a defect. In this case please tell me from where you took the
CEP code (from the open source?), and please send me your two definition
files (the one that works with temporal window and the one that doesn't
with the sliding)
Thanks,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Cc: jira-help-desk@fi-ware.org
Date: 09/02/2015 10:57 AM
Subject: Re: fi-ware CEP short quastion
Hi Tali ,
Thank you very much for your answer ,
The point that you raised about sequence EPA and the way the evaluate the
first event that exactly why i was a bit confuse.
(I even debugged the code and saw that the CEP evaluate in that way, so
that the reason i was thinking i am doing some thing wrong)
about the other question it's OK i found it in the guide but regarding the
consumption definition i didn't get it, in the guide write :
Consumption ? defines the condition for events to be later reused in the
same situation. This
operand is applicable only for join operators, aggregation operators, and
selection operators.
What dose it means "reused" it's happened why should i reused it ?
I have another small question regarding sliding windows in temporal
context :
if i define sliding window : window duration 100 and sliding period 100 is
that mean that every 100 ms i get a new context ?
cause i try that and the EPA didn't detect anything .
I attach to this mail my definition this definition works fine with
temporal interval but at the moment i switch to sliding window it doesn't
detect the event .
Thanks again,
Avishai Gurt,
Oviva Team.
On 8 February 2015 at 10:49, Tali Yatzkar-Haham <TALI@il.ibm.com> wrote:
Hi Avishai,
I looked at your definition file and all seems to be well defined.
Sequence EPA implements <= on the time of the input events. Hence, in the
special case where you define sequence of two events from the same type
(and no other expression that distinct between them), the first event can
play as the first operand and as the second operand, and the EPA will
detect a sequence even if only one event arrived. A solution for it can be
to add in the condition expression:
message1.userId == message2.userId AND
message1.DetectionTime<message2.DetectionTime
As for your other questions, there is an explanation in the user guide.
Please tell me if the explanation there is not sufficient.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Cc: jira-help-desk@fi-ware.org
Date: 05/02/2015 10:32 AM
Subject: Re: fi-ware CEP short quastion
Hi Tali
I attach the definition file to this mail .
Can you please also explain to me in the EPA definition in the Event
Selection tab
What is the meaning of "instance selection" and "consumption" ?
Best Regards,
Avishai Gurt,
Oviva Team.
On 5 February 2015 at 08:05, Tali Yatzkar-Haham <TALI@il.ibm.com> wrote:
Hi Avishai,
Glad you are using the CEP.
I'm the contact person for the CEP in FIWARE.
Could you please send me the CEP definition file (using the export to
file)? this way I can see if there is a problem in the definition file.
Please cc Jira (as I did) to document this corresponding in the FIWARE
issue tracking system.
And yes, if you need more help, we can set a phone call.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Date: 04/02/2015 05:56 PM
Subject: fi-ware CEP short quastion
Good evening Tali,
My name is Avishai and i am working at Oviva a start-up company which
located in Switzerland.
We are working on project which use the CEP on our system (cause we are
applying to one of the Fi-ware accelerators) .
I have one question which i can't understand from the documentation,
if i use an EPA type: "sequence" and add two event from the same type with
two different aliases ,
and i just put condition like that : alias1.id == alias2.id
i thought it will check if two event with the same Id happens in the
context,
But at the moment i sent one event i immediately get the derivation event
.
What i am doing wrong ?
Is it possible to talk with someone that have experience with this system
?
(By the way i am an Israeli and i speak Ivrit so please i will be glade to
talk with some one from your team to clear some points about this system )
Best Regards,
Avishai Gurt,
Oviva Team.
[attachment "hello.json" deleted by Tali Yatzkar-Haham/Haifa/IBM]
[attachment "hello.json" deleted by Tali Yatzkar-Haham/Haifa/IBM]
Hi Tali
Thank you again for your answer .
1. Regarding the sliding window i have noticed that if i set my condition
evaluation policy to : deferred, and my temporal context is sliding window
.
if the initiator of the temporal context is event initiators everything
works fine the EPA detects correctly in the end of each winnow
but if the initiator set to absolute time or at startup the sliding
windows working (cause i see in the log "sliding temporal partition"
message) but the EPA doesn't detect anything.
I attach the two examples .
and another thing, regarding temporal context definition, the "Repeating
Interval" in the absolute time initiator what does it means i couldn't find
anything about it in the user guide ?
2. about the code i just add the .war files to eclipse workspace and
connect debugger to the tomcat (i didn't build but it was enough for
debugging) .
3. after some work on the proton engine i found out that at the moment i
set a new definition to the proton server all the context is gone i mean
the information is reset
For us it will be very important to persist the information some how. Do
you have any idea how to persist the information between deployments ?
Best Regards,
Avishai Gurt,
Oviva Team.
Hi Avishai,
1. So just to be sure - you see problem with sliding window when the
context initiator is either 'absolute time' or 'startup'. Things work fine
with 'event initiator'. I forward it to our development lead.
2. Repeating Interval in Absolute time initiator - I'm not sure about it,
I forward your question to our development lead.
3. Yes, currently in Proton, when you update the definition file, the
internal state is reset. I don't think we are going to support 'hot
updates' in the near future. I forward this question to our development
lead to see if this task is scheduled. We did implement it in a previous
asset we had and it wasn't a simple task, so I don't have ideas how to
implement it in a simple way.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Cc: jira-help-desk@fi-ware.org
Date: 10/02/2015 11:32 AM
Subject: Re: fi-ware CEP short quastion
Hi Tali
Thank you again for your answer .
1. Regarding the sliding window i have noticed that if i set my condition
evaluation policy to : deferred, and my temporal context is sliding
window .
if the initiator of the temporal context is event initiators everything
works fine the EPA detects correctly in the end of each winnow
but if the initiator set to absolute time or at startup the sliding
windows working (cause i see in the log "sliding temporal partition"
message) but the EPA doesn't detect anything.
I attach the two examples .
and another thing, regarding temporal context definition, the "Repeating
Interval" in the absolute time initiator what does it means i couldn't
find anything about it in the user guide ?
2. about the code i just add the .war files to eclipse workspace and
connect debugger to the tomcat (i didn't build but it was enough for
debugging) .
3. after some work on the proton engine i found out that at the moment i
set a new definition to the proton server all the context is gone i mean
the information is reset
For us it will be very important to persist the information some how. Do
you have any idea how to persist the information between deployments ?
Best Regards,
Avishai Gurt,
Oviva Team.
On 10 February 2015 at 08:53, Tali Yatzkar-Haham <TALI@il.ibm.com> wrote:
Hi Avishai,
1. Consumption - let's say you have an EPA with sequence of two event
types (A, B).
Now you received two input events: A1, B1.
Then the EPA will detect the situation (based on A1,B1).
Now you receive another input event from type B - B2
If A is set to be with consumption "Reuse" then A1 can be used again, and
the EPA can detect the situation again based on A1, B2.
If A is set to be with consumption "consume" then after the EPA used A1
.to detect the situation based on (A1,B1), A1 can't be used again, and
when B2 arrives, there will be no situation detection. Is it clear?
2. SlidingWindow - your understanding is correct. The file you sent me is
with an error (not related). If you fix the error, and you still see a
problem with the sliding window, please let me know. If this happen, I
will open a defect. In this case please tell me from where you took the
CEP code (from the open source?), and please send me your two definition
files (the one that works with temporal window and the one that doesn't
with the sliding)
Thanks,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Cc: jira-help-desk@fi-ware.org
Date: 09/02/2015 10:57 AM
Subject: Re: fi-ware CEP short quastion
Hi Tali ,
Thank you very much for your answer ,
The point that you raised about sequence EPA and the way the evaluate the
first event that exactly why i was a bit confuse.
(I even debugged the code and saw that the CEP evaluate in that way, so
that the reason i was thinking i am doing some thing wrong)
about the other question it's OK i found it in the guide but regarding the
consumption definition i didn't get it, in the guide write :
Consumption ? defines the condition for events to be later reused in the
same situation. This
operand is applicable only for join operators, aggregation operators, and
selection operators.
What dose it means "reused" it's happened why should i reused it ?
I have another small question regarding sliding windows in temporal
context :
if i define sliding window : window duration 100 and sliding period 100 is
that mean that every 100 ms i get a new context ?
cause i try that and the EPA didn't detect anything .
I attach to this mail my definition this definition works fine with
temporal interval but at the moment i switch to sliding window it doesn't
detect the event .
Thanks again,
Avishai Gurt,
Oviva Team.
On 8 February 2015 at 10:49, Tali Yatzkar-Haham <TALI@il.ibm.com> wrote:
Hi Avishai,
I looked at your definition file and all seems to be well defined.
Sequence EPA implements <= on the time of the input events. Hence, in the
special case where you define sequence of two events from the same type
(and no other expression that distinct between them), the first event can
play as the first operand and as the second operand, and the EPA will
detect a sequence even if only one event arrived. A solution for it can be
to add in the condition expression:
message1.userId == message2.userId AND
message1.DetectionTime<message2.DetectionTime
As for your other questions, there is an explanation in the user guide.
Please tell me if the explanation there is not sufficient.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Cc: jira-help-desk@fi-ware.org
Date: 05/02/2015 10:32 AM
Subject: Re: fi-ware CEP short quastion
Hi Tali
I attach the definition file to this mail .
Can you please also explain to me in the EPA definition in the Event
Selection tab
What is the meaning of "instance selection" and "consumption" ?
Best Regards,
Avishai Gurt,
Oviva Team.
On 5 February 2015 at 08:05, Tali Yatzkar-Haham <TALI@il.ibm.com> wrote:
Hi Avishai,
Glad you are using the CEP.
I'm the contact person for the CEP in FIWARE.
Could you please send me the CEP definition file (using the export to
file)? this way I can see if there is a problem in the definition file.
Please cc Jira (as I did) to document this corresponding in the FIWARE
issue tracking system.
And yes, if you need more help, we can set a phone call.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Date: 04/02/2015 05:56 PM
Subject: fi-ware CEP short quastion
Good evening Tali,
My name is Avishai and i am working at Oviva a start-up company which
located in Switzerland.
We are working on project which use the CEP on our system (cause we are
applying to one of the Fi-ware accelerators) .
I have one question which i can't understand from the documentation,
if i use an EPA type: "sequence" and add two event from the same type with
two different aliases ,
and i just put condition like that : alias1.id == alias2.id
i thought it will check if two event with the same Id happens in the
context,
But at the moment i sent one event i immediately get the derivation event
.
What i am doing wrong ?
Is it possible to talk with someone that have experience with this system
?
(By the way i am an Israeli and i speak Ivrit so please i will be glade to
talk with some one from your team to clear some points about this system )
Best Regards,
Avishai Gurt,
Oviva Team.
[attachment "hello.json" deleted by Tali Yatzkar-Haham/Haifa/IBM]
[attachment "hello.json" deleted by Tali Yatzkar-Haham/Haifa/IBM]
[attachment "hello_window_startup.json" deleted by Tali
Yatzkar-Haham/Haifa/IBM] [attachment "hello_window_eventInitiator.json"
deleted by Tali Yatzkar-Haham/Haifa/IBM]
Hi Avishai,
I got answers from the development lead.
1. Sliding window - there is currently no support for using startup or
absolute time with Sliding Window.
2. Absolute time - there is currently on-going work to improve the support
in absolute time. The improved support would allow you to define some
partial Absolute time (e.g., time with no date which implies Daily), and
it will support Repeating Interval as well. The current implementation of
Absolute time Repeating interval is not yet tested, so please don't use it
till we improve the support of Absolute time.
3. The task of "hot-updates" (replacing the definition file with no reset
to the internal state) was requested by another project as well, but it is
not scheduled or designed yet. So I can't tell if/when/how we support it.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Date: 13/02/2015 08:10 AM
Subject: Re: fi-ware CEP short quastion
Good morning Tali,
Thank you again for your answers .
regarding 1 , exactly that is the problem and in the files i attache you
can see it .
So if you have any updates from the dev team ill be happy to hear them .
Best Regards,
Avishai Gurt,
Oviva Team.
On 13 February 2015 at 07:31, Tali Yatzkar-Haham <TALI@il.ibm.com> wrote:
Hi Avishai,
1. So just to be sure - you see problem with sliding window when the
context initiator is either 'absolute time' or 'startup'. Things work fine
with 'event initiator'. I forward it to our development lead.
2. Repeating Interval in Absolute time initiator - I'm not sure about it,
I forward your question to our development lead.
3. Yes, currently in Proton, when you update the definition file, the
internal state is reset. I don't think we are going to support 'hot
updates' in the near future. I forward this question to our development
lead to see if this task is scheduled. We did implement it in a previous
asset we had and it wasn't a simple task, so I don't have ideas how to
implement it in a simple way.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Cc: jira-help-desk@fi-ware.org
Date: 10/02/2015 11:32 AM
Subject: Re: fi-ware CEP short quastion
Hi Tali
Thank you again for your answer .
1. Regarding the sliding window i have noticed that if i set my condition
evaluation policy to : deferred, and my temporal context is sliding
window .
if the initiator of the temporal context is event initiators everything
works fine the EPA detects correctly in the end of each winnow
but if the initiator set to absolute time or at startup the sliding
windows working (cause i see in the log "sliding temporal partition"
message) but the EPA doesn't detect anything.
I attach the two examples .
and another thing, regarding temporal context definition, the "Repeating
Interval" in the absolute time initiator what does it means i couldn't
find anything about it in the user guide ?
2. about the code i just add the .war files to eclipse workspace and
connect debugger to the tomcat (i didn't build but it was enough for
debugging) .
3. after some work on the proton engine i found out that at the moment i
set a new definition to the proton server all the context is gone i mean
the information is reset
For us it will be very important to persist the information some how. Do
you have any idea how to persist the information between deployments ?
Best Regards,
Avishai Gurt,
Oviva Team.
On 10 February 2015 at 08:53, Tali Yatzkar-Haham <TALI@il.ibm.com> wrote:
Hi Avishai,
1. Consumption - let's say you have an EPA with sequence of two event
types (A, B).
Now you received two input events: A1, B1.
Then the EPA will detect the situation (based on A1,B1).
Now you receive another input event from type B - B2
If A is set to be with consumption "Reuse" then A1 can be used again, and
the EPA can detect the situation again based on A1, B2.
If A is set to be with consumption "consume" then after the EPA used A1
.to detect the situation based on (A1,B1), A1 can't be used again, and
when B2 arrives, there will be no situation detection. Is it clear?
2. SlidingWindow - your understanding is correct. The file you sent me is
with an error (not related). If you fix the error, and you still see a
problem with the sliding window, please let me know. If this happen, I
will open a defect. In this case please tell me from where you took the
CEP code (from the open source?), and please send me your two definition
files (the one that works with temporal window and the one that doesn't
with the sliding)
Thanks,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Cc: jira-help-desk@fi-ware.org
Date: 09/02/2015 10:57 AM
Subject: Re: fi-ware CEP short quastion
Hi Tali ,
Thank you very much for your answer ,
The point that you raised about sequence EPA and the way the evaluate the
first event that exactly why i was a bit confuse.
(I even debugged the code and saw that the CEP evaluate in that way, so
that the reason i was thinking i am doing some thing wrong)
about the other question it's OK i found it in the guide but regarding the
consumption definition i didn't get it, in the guide write :
Consumption ? defines the condition for events to be later reused in the
same situation. This
operand is applicable only for join operators, aggregation operators, and
selection operators.
What dose it means "reused" it's happened why should i reused it ?
I have another small question regarding sliding windows in temporal
context :
if i define sliding window : window duration 100 and sliding period 100 is
that mean that every 100 ms i get a new context ?
cause i try that and the EPA didn't detect anything .
I attach to this mail my definition this definition works fine with
temporal interval but at the moment i switch to sliding window it doesn't
detect the event .
Thanks again,
Avishai Gurt,
Oviva Team.
On 8 February 2015 at 10:49, Tali Yatzkar-Haham <TALI@il.ibm.com> wrote:
Hi Avishai,
I looked at your definition file and all seems to be well defined.
Sequence EPA implements <= on the time of the input events. Hence, in the
special case where you define sequence of two events from the same type
(and no other expression that distinct between them), the first event can
play as the first operand and as the second operand, and the EPA will
detect a sequence even if only one event arrived. A solution for it can be
to add in the condition expression:
message1.userId == message2.userId AND
message1.DetectionTime<message2.DetectionTime
As for your other questions, there is an explanation in the user guide.
Please tell me if the explanation there is not sufficient.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Cc: jira-help-desk@fi-ware.org
Date: 05/02/2015 10:32 AM
Subject: Re: fi-ware CEP short quastion
Hi Tali
I attach the definition file to this mail .
Can you please also explain to me in the EPA definition in the Event
Selection tab
What is the meaning of "instance selection" and "consumption" ?
Best Regards,
Avishai Gurt,
Oviva Team.
On 5 February 2015 at 08:05, Tali Yatzkar-Haham <TALI@il.ibm.com> wrote:
Hi Avishai,
Glad you are using the CEP.
I'm the contact person for the CEP in FIWARE.
Could you please send me the CEP definition file (using the export to
file)? this way I can see if there is a problem in the definition file.
Please cc Jira (as I did) to document this corresponding in the FIWARE
issue tracking system.
And yes, if you need more help, we can set a phone call.
Regards,
Tali
Tali Yatzkar Haham
Event-based Middleware & Solutions
IBM Haifa Research Lab, Israel
tali@il.ibm.com 972-4-8296320
From: Avishai Gurt <avishaigurt@oviva.ch>
To: Tali Yatzkar-Haham/Haifa/IBM@IBMIL
Date: 04/02/2015 05:56 PM
Subject: fi-ware CEP short quastion
Good evening Tali,
My name is Avishai and i am working at Oviva a start-up company which
located in Switzerland.
We are working on project which use the CEP on our system (cause we are
applying to one of the Fi-ware accelerators) .
I have one question which i can't understand from the documentation,
if i use an EPA type: "sequence" and add two event from the same type with
two different aliases ,
and i just put condition like that : alias1.id == alias2.id
i thought it will check if two event with the same Id happens in the
context,
But at the moment i sent one event i immediately get the derivation event
.
What i am doing wrong ?
Is it possible to talk with someone that have experience with this system
?
(By the way i am an Israeli and i speak Ivrit so please i will be glade to
talk with some one from your team to clear some points about this system )
Best Regards,
Avishai Gurt,
Oviva Team.
[attachment "hello.json" deleted by Tali Yatzkar-Haham/Haifa/IBM]
[attachment "hello.json" deleted by Tali Yatzkar-Haham/Haifa/IBM]
[attachment "hello_window_startup.json" deleted by Tali
Yatzkar-Haham/Haifa/IBM] [attachment "hello_window_eventInitiator.json"
deleted by Tali Yatzkar-Haham/Haifa/IBM]
Hi Tali
I attach the definition file to this mail .
Can you please also explain to me in the EPA definition in the Event
Selection tab
What is the meaning of "instance selection" and "consumption" ?
Best Regards,
Avishai Gurt,
Oviva Team.