Neat Zeppelin Use Cases

classic Classic list List threaded Threaded
16 messages Options
Reply | Threaded
Open this post in threaded view
|

Neat Zeppelin Use Cases

Daniel Haviv
Hi everybody,
We are going to arrange a meetup in Israel where I'm going to dedicate 45 minutes to introduce Zeppelin to the Big Data community in Israel.
I'm going to display general usage of the Spark cli via zeppelin (and it's charting capabilities) and the WordCloud.
 
Please share any cool examples you have of Zeppelin I can use.
After the meetup, I will make sure these examples will be publicly available.


Thanks,
Daniel Haviv
Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

IT CTO
Can I get an invite? which meetup you are presenting in?


On Mon, Jun 29, 2015 at 9:30 AM Daniel Haviv <[hidden email]> wrote:
Hi everybody,
We are going to arrange a meetup in Israel where I'm going to dedicate 45 minutes to introduce Zeppelin to the Big Data community in Israel.
I'm going to display general usage of the Spark cli via zeppelin (and it's charting capabilities) and the WordCloud.
 
Please share any cool examples you have of Zeppelin I can use.
After the meetup, I will make sure these examples will be publicly available.


Thanks,
Daniel Haviv
tog
Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

tog
Hi

Any idea what you would like to showcase ? Any deadline ?

Cheers
Guillaume

On 29 June 2015 at 07:52, IT CTO <[hidden email]> wrote:
Can I get an invite? which meetup you are presenting in?


On Mon, Jun 29, 2015 at 9:30 AM Daniel Haviv <[hidden email]> wrote:
Hi everybody,
We are going to arrange a meetup in Israel where I'm going to dedicate 45 minutes to introduce Zeppelin to the Big Data community in Israel.
I'm going to display general usage of the Spark cli via zeppelin (and it's charting capabilities) and the WordCloud.
 
Please share any cool examples you have of Zeppelin I can use.
After the meetup, I will make sure these examples will be publicly available.


Thanks,
Daniel Haviv



--
PGP KeyID: 2048R/EA31CFC9  subkeys.pgp.net
Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

Daniel Haviv
I'm displaying the basic SQL and charting capabilites but I'd also like to demo some more advanced capabilites (like the WordCloud example).
I need it by the end of the week.

Thanks,
Daniel

On Mon, Jun 29, 2015 at 12:19 PM, tog <[hidden email]> wrote:
Hi

Any idea what you would like to showcase ? Any deadline ?

Cheers
Guillaume

On 29 June 2015 at 07:52, IT CTO <[hidden email]> wrote:
Can I get an invite? which meetup you are presenting in?


On Mon, Jun 29, 2015 at 9:30 AM Daniel Haviv <[hidden email]> wrote:
Hi everybody,
We are going to arrange a meetup in Israel where I'm going to dedicate 45 minutes to introduce Zeppelin to the Big Data community in Israel.
I'm going to display general usage of the Spark cli via zeppelin (and it's charting capabilities) and the WordCloud.
 
Please share any cool examples you have of Zeppelin I can use.
After the meetup, I will make sure these examples will be publicly available.


Thanks,
Daniel Haviv



--
PGP KeyID: 2048R/EA31CFC9  subkeys.pgp.net

tog
Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

tog
As far as the wordcloud is concerned you will find an example in the mailing list. I can help you if you have any issue with it but you need the narrative of your demo to be set ;-)


On 29 June 2015 at 10:25, Daniel Haviv <[hidden email]> wrote:
I'm displaying the basic SQL and charting capabilites but I'd also like to demo some more advanced capabilites (like the WordCloud example).
I need it by the end of the week.

Thanks,
Daniel

On Mon, Jun 29, 2015 at 12:19 PM, tog <[hidden email]> wrote:
Hi

Any idea what you would like to showcase ? Any deadline ?

Cheers
Guillaume

On 29 June 2015 at 07:52, IT CTO <[hidden email]> wrote:
Can I get an invite? which meetup you are presenting in?


On Mon, Jun 29, 2015 at 9:30 AM Daniel Haviv <[hidden email]> wrote:
Hi everybody,
We are going to arrange a meetup in Israel where I'm going to dedicate 45 minutes to introduce Zeppelin to the Big Data community in Israel.
I'm going to display general usage of the Spark cli via zeppelin (and it's charting capabilities) and the WordCloud.
 
Please share any cool examples you have of Zeppelin I can use.
After the meetup, I will make sure these examples will be publicly available.


Thanks,
Daniel Haviv



--
PGP KeyID: 2048R/EA31CFC9  subkeys.pgp.net




--
PGP KeyID: 2048R/EA31CFC9  subkeys.pgp.net
Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

Koji Sekiguchi
In reply to this post by Daniel Haviv
Hi Daniel,

I'd like to share my use case for Zeppelin. NLP4L[1], which is an NLP tool for
Apache Lucene, uses Zeppelin for not only user interface but also charts
in order to visualize word counts in Lucene index.

http://nlp4l.github.io/tutorial.html#withZeppelin

To communicate with Zeppelin, I've implemented NLP4LInterpreter which extends
org.apache.zeppelin.interpreter.Interpreter.

https://github.com/NLP4L/nlp4l/blob/master/src/main/java/org/nlp4l/zeppelin/NLP4LInterpreter.java

Thanks!

Koji

[1] https://github.com/NLP4L/nlp4l


On 2015/06/29 15:30, Daniel Haviv wrote:

> Hi everybody,
> We are going to arrange a meetup in Israel where I'm going to dedicate 45
> minutes to introduce Zeppelin to the Big Data community in Israel.
> I'm going to display general usage of the Spark cli via zeppelin (and it's
> charting capabilities) and the WordCloud.
>
> Please share any cool examples you have of Zeppelin I can use.
> After the meetup, I will make sure these examples will be publicly
> available.
>
>
> Thanks,
> Daniel Haviv
>

Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

Ophir Cohen
In reply to this post by Daniel Haviv
Hi Daniel and good luck.
Can you share the meetup details?
10x

On Mon, Jun 29, 2015 at 9:30 AM, Daniel Haviv <[hidden email]> wrote:
Hi everybody,
We are going to arrange a meetup in Israel where I'm going to dedicate 45 minutes to introduce Zeppelin to the Big Data community in Israel.
I'm going to display general usage of the Spark cli via zeppelin (and it's charting capabilities) and the WordCloud.
 
Please share any cool examples you have of Zeppelin I can use.
After the meetup, I will make sure these examples will be publicly available.


Thanks,
Daniel Haviv

Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

IT CTO
This might be a stupid question but ....
Quick question regarding the NLP4L - why did you need to create an interpreter for NLP4L? what is the deference from just using the %dep and loading from scala and then accessing it from scala paragrap?
Eran

On Tue, Jun 30, 2015 at 10:56 AM Ophir Cohen <[hidden email]> wrote:
Hi Daniel and good luck.
Can you share the meetup details?
10x

On Mon, Jun 29, 2015 at 9:30 AM, Daniel Haviv <[hidden email]> wrote:
Hi everybody,
We are going to arrange a meetup in Israel where I'm going to dedicate 45 minutes to introduce Zeppelin to the Big Data community in Israel.
I'm going to display general usage of the Spark cli via zeppelin (and it's charting capabilities) and the WordCloud.
 
Please share any cool examples you have of Zeppelin I can use.
After the meetup, I will make sure these examples will be publicly available.


Thanks,
Daniel Haviv

Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

DuyHai Doan
Hello Daniel

I had set up some demos demonstrating the added values of Zeppelin for the Spark over Cassandra combo. Source code here: https://github.com/doanduyhai/incubator-zeppelin/tree/Spark_Cassandra_Demo

Video of the meetup I gave in Amsterdam here: https://youtu.be/Y_AjbK4LKB0?t=1118


you have some us unemployment stats from one CSV, some artists & albums from other CSV files to demonstrate Spark join usage and display the results with Zeppelin using PieChart.

You can either take the whole examples or just use the CSV data set and load them in Spark without Cassandra (using the DataFrame API to register an RDD as table). Choices is yours.

Best wishes for your meetup


On Tue, Jun 30, 2015 at 1:25 PM, IT CTO <[hidden email]> wrote:
This might be a stupid question but ....
Quick question regarding the NLP4L - why did you need to create an interpreter for NLP4L? what is the deference from just using the %dep and loading from scala and then accessing it from scala paragrap?
Eran

On Tue, Jun 30, 2015 at 10:56 AM Ophir Cohen <[hidden email]> wrote:
Hi Daniel and good luck.
Can you share the meetup details?
10x

On Mon, Jun 29, 2015 at 9:30 AM, Daniel Haviv <[hidden email]> wrote:
Hi everybody,
We are going to arrange a meetup in Israel where I'm going to dedicate 45 minutes to introduce Zeppelin to the Big Data community in Israel.
I'm going to display general usage of the Spark cli via zeppelin (and it's charting capabilities) and the WordCloud.
 
Please share any cool examples you have of Zeppelin I can use.
After the meetup, I will make sure these examples will be publicly available.


Thanks,
Daniel Haviv


Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

Koji Sekiguchi
In reply to this post by IT CTO
Hi IT CTO,

Thank you for the reply!

Because I couldn't find any ways to use charts (bar chart, pie chart, etc.) in Zeppelin
without having my own Interpreter... If you know it, please let me know.

And now I see the comment of DepInterpreter, it seems it is used for Spark:

/**
  * DepInterpreter downloads dependencies and pass them when SparkInterpreter initialized.
  * It extends SparkInterpreter but does not create sparkcontext
  *
  */

nevertheless, it can be used for general purpose?

regards,

Koji


On 2015/06/30 20:25, IT CTO wrote:

> This might be a stupid question but ....
> Quick question regarding the NLP4L - why did you need to create an
> interpreter for NLP4L? what is the deference from just using the %dep and
> loading from scala and then accessing it from scala paragrap?
> Eran
>
> On Tue, Jun 30, 2015 at 10:56 AM Ophir Cohen <[hidden email]> wrote:
>
>> Hi Daniel and good luck.
>> Can you share the meetup details?
>> 10x
>>
>> On Mon, Jun 29, 2015 at 9:30 AM, Daniel Haviv <
>> [hidden email]> wrote:
>>
>>> Hi everybody,
>>> We are going to arrange a meetup in Israel where I'm going to dedicate 45
>>> minutes to introduce Zeppelin to the Big Data community in Israel.
>>> I'm going to display general usage of the Spark cli via zeppelin (and
>>> it's charting capabilities) and the WordCloud.
>>>
>>> Please share any cool examples you have of Zeppelin I can use.
>>> After the meetup, I will make sure these examples will be publicly
>>> available.
>>>
>>>
>>> Thanks,
>>> Daniel Haviv
>>>
>>
>>
>

Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

IT CTO
I hope I am not missleading you but as far as I understand the "spark interpreter" is actually a scala interpreter until the time you access one of the context objects, and as much as I understand from you library, you have a stand alone library which should be loaded and accessed locally. 
So - all you need to do is in you scala code use the %dep to load the library and then call it's API.

This is just to save you time next time. 
You need to write an interpreter only if you want to access a back-end computation engine such as SPARK or ignite or any ODBC source. for local api call %dep should be enough.

HTH,
Eran

On Wed, Jul 1, 2015 at 5:24 AM Koji Sekiguchi <[hidden email]> wrote:
Hi IT CTO,

Thank you for the reply!

Because I couldn't find any ways to use charts (bar chart, pie chart, etc.) in Zeppelin
without having my own Interpreter... If you know it, please let me know.

And now I see the comment of DepInterpreter, it seems it is used for Spark:

/**
  * DepInterpreter downloads dependencies and pass them when SparkInterpreter initialized.
  * It extends SparkInterpreter but does not create sparkcontext
  *
  */

nevertheless, it can be used for general purpose?

regards,

Koji


On 2015/06/30 20:25, IT CTO wrote:
> This might be a stupid question but ....
> Quick question regarding the NLP4L - why did you need to create an
> interpreter for NLP4L? what is the deference from just using the %dep and
> loading from scala and then accessing it from scala paragrap?
> Eran
>
> On Tue, Jun 30, 2015 at 10:56 AM Ophir Cohen <[hidden email]> wrote:
>
>> Hi Daniel and good luck.
>> Can you share the meetup details?
>> 10x
>>
>> On Mon, Jun 29, 2015 at 9:30 AM, Daniel Haviv <
>> [hidden email]> wrote:
>>
>>> Hi everybody,
>>> We are going to arrange a meetup in Israel where I'm going to dedicate 45
>>> minutes to introduce Zeppelin to the Big Data community in Israel.
>>> I'm going to display general usage of the Spark cli via zeppelin (and
>>> it's charting capabilities) and the WordCloud.
>>>
>>> Please share any cool examples you have of Zeppelin I can use.
>>> After the meetup, I will make sure these examples will be publicly
>>> available.
>>>
>>>
>>> Thanks,
>>> Daniel Haviv
>>>
>>
>>
>

Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

Koji Sekiguchi
Hi IT CTO,

Thank you for the reply!

Ok, so now my question is only visualization of charts.

http://nlp4l.github.io/tutorial.html#withZeppelin_visualize

How can I use charts to display word count appearing Lucene index
without implementing Interpreter?

regards,

Koji

On 2015/07/02 15:14, IT CTO wrote:

> I hope I am not missleading you but as far as I understand the "spark
> interpreter" is actually a scala interpreter until the time you access one
> of the context objects, and as much as I understand from you library, you
> have a stand alone library which should be loaded and accessed locally.
> So - all you need to do is in you scala code use the %dep to load the
> library and then call it's API.
>
> This is just to save you time next time.
> You need to write an interpreter only if you want to access a back-end
> computation engine such as SPARK or ignite or any ODBC source. for local
> api call %dep should be enough.
>
> HTH,
> Eran
>
> On Wed, Jul 1, 2015 at 5:24 AM Koji Sekiguchi <[hidden email]>
> wrote:
>
>> Hi IT CTO,
>>
>> Thank you for the reply!
>>
>> Because I couldn't find any ways to use charts (bar chart, pie chart,
>> etc.) in Zeppelin
>> without having my own Interpreter... If you know it, please let me know.
>>
>> And now I see the comment of DepInterpreter, it seems it is used for Spark:
>>
>> /**
>>    * DepInterpreter downloads dependencies and pass them when
>> SparkInterpreter initialized.
>>    * It extends SparkInterpreter but does not create sparkcontext
>>    *
>>    */
>>
>> nevertheless, it can be used for general purpose?
>>
>> regards,
>>
>> Koji
>>
>>
>> On 2015/06/30 20:25, IT CTO wrote:
>>> This might be a stupid question but ....
>>> Quick question regarding the NLP4L - why did you need to create an
>>> interpreter for NLP4L? what is the deference from just using the %dep and
>>> loading from scala and then accessing it from scala paragrap?
>>> Eran
>>>
>>> On Tue, Jun 30, 2015 at 10:56 AM Ophir Cohen <[hidden email]> wrote:
>>>
>>>> Hi Daniel and good luck.
>>>> Can you share the meetup details?
>>>> 10x
>>>>
>>>> On Mon, Jun 29, 2015 at 9:30 AM, Daniel Haviv <
>>>> [hidden email]> wrote:
>>>>
>>>>> Hi everybody,
>>>>> We are going to arrange a meetup in Israel where I'm going to dedicate
>> 45
>>>>> minutes to introduce Zeppelin to the Big Data community in Israel.
>>>>> I'm going to display general usage of the Spark cli via zeppelin (and
>>>>> it's charting capabilities) and the WordCloud.
>>>>>
>>>>> Please share any cool examples you have of Zeppelin I can use.
>>>>> After the meetup, I will make sure these examples will be publicly
>>>>> available.
>>>>>
>>>>>
>>>>> Thanks,
>>>>> Daniel Haviv
>>>>>
>>>>
>>>>
>>>
>>
>>
>

Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

IT CTO
Based on the documentation (http://zeppelin.incubator.apache.org/docs/display.html) you can use the following:
Display as Table, Chart If output starts with %table, it is interpreted as a table. Table can be seen as chart. Output's format should be, row separated by '\n' (newline) and column separated by '\t' (tab). First row is header. (http://zeppelin.incubator.apache.org/assets/themes/zeppelin/img/screenshots/display_table.png

Eran

On Thu, Jul 2, 2015 at 9:26 AM Koji Sekiguchi <[hidden email]> wrote:
Hi IT CTO,

Thank you for the reply!

Ok, so now my question is only visualization of charts.

http://nlp4l.github.io/tutorial.html#withZeppelin_visualize

How can I use charts to display word count appearing Lucene index
without implementing Interpreter?

regards,

Koji

On 2015/07/02 15:14, IT CTO wrote:
> I hope I am not missleading you but as far as I understand the "spark
> interpreter" is actually a scala interpreter until the time you access one
> of the context objects, and as much as I understand from you library, you
> have a stand alone library which should be loaded and accessed locally.
> So - all you need to do is in you scala code use the %dep to load the
> library and then call it's API.
>
> This is just to save you time next time.
> You need to write an interpreter only if you want to access a back-end
> computation engine such as SPARK or ignite or any ODBC source. for local
> api call %dep should be enough.
>
> HTH,
> Eran
>
> On Wed, Jul 1, 2015 at 5:24 AM Koji Sekiguchi <[hidden email]>
> wrote:
>
>> Hi IT CTO,
>>
>> Thank you for the reply!
>>
>> Because I couldn't find any ways to use charts (bar chart, pie chart,
>> etc.) in Zeppelin
>> without having my own Interpreter... If you know it, please let me know.
>>
>> And now I see the comment of DepInterpreter, it seems it is used for Spark:
>>
>> /**
>>    * DepInterpreter downloads dependencies and pass them when
>> SparkInterpreter initialized.
>>    * It extends SparkInterpreter but does not create sparkcontext
>>    *
>>    */
>>
>> nevertheless, it can be used for general purpose?
>>
>> regards,
>>
>> Koji
>>
>>
>> On 2015/06/30 20:25, IT CTO wrote:
>>> This might be a stupid question but ....
>>> Quick question regarding the NLP4L - why did you need to create an
>>> interpreter for NLP4L? what is the deference from just using the %dep and
>>> loading from scala and then accessing it from scala paragrap?
>>> Eran
>>>
>>> On Tue, Jun 30, 2015 at 10:56 AM Ophir Cohen <[hidden email]> wrote:
>>>
>>>> Hi Daniel and good luck.
>>>> Can you share the meetup details?
>>>> 10x
>>>>
>>>> On Mon, Jun 29, 2015 at 9:30 AM, Daniel Haviv <
>>>> [hidden email]> wrote:
>>>>
>>>>> Hi everybody,
>>>>> We are going to arrange a meetup in Israel where I'm going to dedicate
>> 45
>>>>> minutes to introduce Zeppelin to the Big Data community in Israel.
>>>>> I'm going to display general usage of the Spark cli via zeppelin (and
>>>>> it's charting capabilities) and the WordCloud.
>>>>>
>>>>> Please share any cool examples you have of Zeppelin I can use.
>>>>> After the meetup, I will make sure these examples will be publicly
>>>>> available.
>>>>>
>>>>>
>>>>> Thanks,
>>>>> Daniel Haviv
>>>>>
>>>>
>>>>
>>>
>>
>>
>

Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

Koji Sekiguchi
I know it.

My question is how I can do it without implementing Interpreter.

$ find . -name \*.java|xargs grep "%table"
./hive/src/main/java/org/apache/zeppelin/hive/HiveInterpreter.java:        msg = new
StringBuilder("%table ");
./ignite/src/main/java/org/apache/zeppelin/ignite/IgniteSqlInterpreter.java:    StringBuilder msg =
new StringBuilder("%table ");
./lens/src/main/java/org/apache/zeppelin/lens/LensInterpreter.java:        sb.append("%table " +
entry.getKey() + " \n");
./lens/src/main/java/org/apache/zeppelin/lens/LensInterpreter.java:      sb.append("%table ");
./spark/src/main/java/org/apache/zeppelin/spark/ZeppelinContext.java:    return "%table " + msg;
./tajo/src/main/java/org/apache/zeppelin/tajo/TajoInterpreter.java:        msg = new
StringBuilder("%table ");

DepInterpreter you recommended doesn't implement it?


On 2015/07/02 15:36, IT CTO wrote:

> Based on the documentation (
> http://zeppelin.incubator.apache.org/docs/display.html) you can use the
> following:
> Display as Table, Chart If output starts with %table, it is interpreted as
> a table. Table can be seen as chart. Output's format should be, row
> separated by '\n' (newline) and column separated by '\t' (tab). First row
> is header. (
> http://zeppelin.incubator.apache.org/assets/themes/zeppelin/img/screenshots/display_table.png
> )
>
> Eran
>
> On Thu, Jul 2, 2015 at 9:26 AM Koji Sekiguchi <[hidden email]>
> wrote:
>
>> Hi IT CTO,
>>
>> Thank you for the reply!
>>
>> Ok, so now my question is only visualization of charts.
>>
>> http://nlp4l.github.io/tutorial.html#withZeppelin_visualize
>>
>> How can I use charts to display word count appearing Lucene index
>> without implementing Interpreter?
>>
>> regards,
>>
>> Koji
>>
>> On 2015/07/02 15:14, IT CTO wrote:
>>> I hope I am not missleading you but as far as I understand the "spark
>>> interpreter" is actually a scala interpreter until the time you access
>> one
>>> of the context objects, and as much as I understand from you library, you
>>> have a stand alone library which should be loaded and accessed locally.
>>> So - all you need to do is in you scala code use the %dep to load the
>>> library and then call it's API.
>>>
>>> This is just to save you time next time.
>>> You need to write an interpreter only if you want to access a back-end
>>> computation engine such as SPARK or ignite or any ODBC source. for local
>>> api call %dep should be enough.
>>>
>>> HTH,
>>> Eran
>>>
>>> On Wed, Jul 1, 2015 at 5:24 AM Koji Sekiguchi <
>> [hidden email]>
>>> wrote:
>>>
>>>> Hi IT CTO,
>>>>
>>>> Thank you for the reply!
>>>>
>>>> Because I couldn't find any ways to use charts (bar chart, pie chart,
>>>> etc.) in Zeppelin
>>>> without having my own Interpreter... If you know it, please let me know.
>>>>
>>>> And now I see the comment of DepInterpreter, it seems it is used for
>> Spark:
>>>>
>>>> /**
>>>>     * DepInterpreter downloads dependencies and pass them when
>>>> SparkInterpreter initialized.
>>>>     * It extends SparkInterpreter but does not create sparkcontext
>>>>     *
>>>>     */
>>>>
>>>> nevertheless, it can be used for general purpose?
>>>>
>>>> regards,
>>>>
>>>> Koji
>>>>
>>>>
>>>> On 2015/06/30 20:25, IT CTO wrote:
>>>>> This might be a stupid question but ....
>>>>> Quick question regarding the NLP4L - why did you need to create an
>>>>> interpreter for NLP4L? what is the deference from just using the %dep
>> and
>>>>> loading from scala and then accessing it from scala paragrap?
>>>>> Eran
>>>>>
>>>>> On Tue, Jun 30, 2015 at 10:56 AM Ophir Cohen <[hidden email]> wrote:
>>>>>
>>>>>> Hi Daniel and good luck.
>>>>>> Can you share the meetup details?
>>>>>> 10x
>>>>>>
>>>>>> On Mon, Jun 29, 2015 at 9:30 AM, Daniel Haviv <
>>>>>> [hidden email]> wrote:
>>>>>>
>>>>>>> Hi everybody,
>>>>>>> We are going to arrange a meetup in Israel where I'm going to
>> dedicate
>>>> 45
>>>>>>> minutes to introduce Zeppelin to the Big Data community in Israel.
>>>>>>> I'm going to display general usage of the Spark cli via zeppelin (and
>>>>>>> it's charting capabilities) and the WordCloud.
>>>>>>>
>>>>>>> Please share any cool examples you have of Zeppelin I can use.
>>>>>>> After the meetup, I will make sure these examples will be publicly
>>>>>>> available.
>>>>>>>
>>>>>>>
>>>>>>> Thanks,
>>>>>>> Daniel Haviv
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>>
>

tog
Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

tog
Hi Koji,

Can you distribute a jar file of your package - then it is easy (I believe) to show you how to do it without interpreter ?

Cheers
Guillaume

On 2 July 2015 at 07:55, Koji Sekiguchi <[hidden email]> wrote:
I know it.

My question is how I can do it without implementing Interpreter.

$ find . -name \*.java|xargs grep "%table"
./hive/src/main/java/org/apache/zeppelin/hive/HiveInterpreter.java:        msg = new StringBuilder("%table ");
./ignite/src/main/java/org/apache/zeppelin/ignite/IgniteSqlInterpreter.java:    StringBuilder msg = new StringBuilder("%table ");
./lens/src/main/java/org/apache/zeppelin/lens/LensInterpreter.java:        sb.append("%table " + entry.getKey() + " \n");
./lens/src/main/java/org/apache/zeppelin/lens/LensInterpreter.java:      sb.append("%table ");
./spark/src/main/java/org/apache/zeppelin/spark/ZeppelinContext.java:    return "%table " + msg;
./tajo/src/main/java/org/apache/zeppelin/tajo/TajoInterpreter.java:        msg = new StringBuilder("%table ");

DepInterpreter you recommended doesn't implement it?


On 2015/07/02 15:36, IT CTO wrote:
Based on the documentation (

http://zeppelin.incubator.apache.org/docs/display.html) you can use the
following:
Display as Table, Chart If output starts with %table, it is interpreted as
a table. Table can be seen as chart. Output's format should be, row
separated by '\n' (newline) and column separated by '\t' (tab). First row
is header. (
http://zeppelin.incubator.apache.org/assets/themes/zeppelin/img/screenshots/display_table.png
)

Eran

On Thu, Jul 2, 2015 at 9:26 AM Koji Sekiguchi <[hidden email]>
wrote:

Hi IT CTO,

Thank you for the reply!

Ok, so now my question is only visualization of charts.

http://nlp4l.github.io/tutorial.html#withZeppelin_visualize

How can I use charts to display word count appearing Lucene index
without implementing Interpreter?

regards,

Koji

On 2015/07/02 15:14, IT CTO wrote:
I hope I am not missleading you but as far as I understand the "spark
interpreter" is actually a scala interpreter until the time you access
one
of the context objects, and as much as I understand from you library, you
have a stand alone library which should be loaded and accessed locally.
So - all you need to do is in you scala code use the %dep to load the
library and then call it's API.

This is just to save you time next time.
You need to write an interpreter only if you want to access a back-end
computation engine such as SPARK or ignite or any ODBC source. for local
api call %dep should be enough.

HTH,
Eran

On Wed, Jul 1, 2015 at 5:24 AM Koji Sekiguchi <
[hidden email]>
wrote:

Hi IT CTO,

Thank you for the reply!

Because I couldn't find any ways to use charts (bar chart, pie chart,
etc.) in Zeppelin
without having my own Interpreter... If you know it, please let me know.

And now I see the comment of DepInterpreter, it seems it is used for
Spark:

/**
    * DepInterpreter downloads dependencies and pass them when
SparkInterpreter initialized.
    * It extends SparkInterpreter but does not create sparkcontext
    *
    */

nevertheless, it can be used for general purpose?

regards,

Koji


On 2015/06/30 20:25, IT CTO wrote:
This might be a stupid question but ....
Quick question regarding the NLP4L - why did you need to create an
interpreter for NLP4L? what is the deference from just using the %dep
and
loading from scala and then accessing it from scala paragrap?
Eran

On Tue, Jun 30, 2015 at 10:56 AM Ophir Cohen <[hidden email]> wrote:

Hi Daniel and good luck.
Can you share the meetup details?
10x

On Mon, Jun 29, 2015 at 9:30 AM, Daniel Haviv <
[hidden email]> wrote:

Hi everybody,
We are going to arrange a meetup in Israel where I'm going to
dedicate
45
minutes to introduce Zeppelin to the Big Data community in Israel.
I'm going to display general usage of the Spark cli via zeppelin (and
it's charting capabilities) and the WordCloud.

Please share any cool examples you have of Zeppelin I can use.
After the meetup, I will make sure these examples will be publicly
available.


Thanks,
Daniel Haviv














--
PGP KeyID: 2048R/EA31CFC9  subkeys.pgp.net
Reply | Threaded
Open this post in threaded view
|

Re: Neat Zeppelin Use Cases

Koji Sekiguchi
Hi Guillaume,

Thank you for your follow-up reply!

After the following discussion, I got a private mail from him and he said
"I see you are familiar with all the zeppelin options so I am sure you made
the right choices." in his mail. So now I'm pretty sure I need to implement
my own Interpreter to display charts.

Thanks again to IT CTO and Guillaume!

Koji

On 2015/07/02 17:33, tog wrote:

> Hi Koji,
>
> Can you distribute a jar file of your package - then it is easy (I believe)
> to show you how to do it without interpreter ?
>
> Cheers
> Guillaume
>
> On 2 July 2015 at 07:55, Koji Sekiguchi <[hidden email]> wrote:
>
>> I know it.
>>
>> My question is how I can do it without implementing Interpreter.
>>
>> $ find . -name \*.java|xargs grep "%table"
>> ./hive/src/main/java/org/apache/zeppelin/hive/HiveInterpreter.java:
>> msg = new StringBuilder("%table ");
>> ./ignite/src/main/java/org/apache/zeppelin/ignite/IgniteSqlInterpreter.java:
>>    StringBuilder msg = new StringBuilder("%table ");
>> ./lens/src/main/java/org/apache/zeppelin/lens/LensInterpreter.java:
>> sb.append("%table " + entry.getKey() + " \n");
>> ./lens/src/main/java/org/apache/zeppelin/lens/LensInterpreter.java:
>> sb.append("%table ");
>> ./spark/src/main/java/org/apache/zeppelin/spark/ZeppelinContext.java:
>> return "%table " + msg;
>> ./tajo/src/main/java/org/apache/zeppelin/tajo/TajoInterpreter.java:
>> msg = new StringBuilder("%table ");
>>
>> DepInterpreter you recommended doesn't implement it?
>>
>>
>> On 2015/07/02 15:36, IT CTO wrote:
>>
>>> Based on the documentation (
>>>
>>> http://zeppelin.incubator.apache.org/docs/display.html) you can use the
>>> following:
>>> Display as Table, Chart If output starts with %table, it is interpreted as
>>> a table. Table can be seen as chart. Output's format should be, row
>>> separated by '\n' (newline) and column separated by '\t' (tab). First row
>>> is header. (
>>>
>>> http://zeppelin.incubator.apache.org/assets/themes/zeppelin/img/screenshots/display_table.png
>>> )
>>>
>>> Eran
>>>
>>> On Thu, Jul 2, 2015 at 9:26 AM Koji Sekiguchi <
>>> [hidden email]>
>>> wrote:
>>>
>>>   Hi IT CTO,
>>>>
>>>> Thank you for the reply!
>>>>
>>>> Ok, so now my question is only visualization of charts.
>>>>
>>>> http://nlp4l.github.io/tutorial.html#withZeppelin_visualize
>>>>
>>>> How can I use charts to display word count appearing Lucene index
>>>> without implementing Interpreter?
>>>>
>>>> regards,
>>>>
>>>> Koji
>>>>
>>>> On 2015/07/02 15:14, IT CTO wrote:
>>>>
>>>>> I hope I am not missleading you but as far as I understand the "spark
>>>>> interpreter" is actually a scala interpreter until the time you access
>>>>>
>>>> one
>>>>
>>>>> of the context objects, and as much as I understand from you library,
>>>>> you
>>>>> have a stand alone library which should be loaded and accessed locally.
>>>>> So - all you need to do is in you scala code use the %dep to load the
>>>>> library and then call it's API.
>>>>>
>>>>> This is just to save you time next time.
>>>>> You need to write an interpreter only if you want to access a back-end
>>>>> computation engine such as SPARK or ignite or any ODBC source. for local
>>>>> api call %dep should be enough.
>>>>>
>>>>> HTH,
>>>>> Eran
>>>>>
>>>>> On Wed, Jul 1, 2015 at 5:24 AM Koji Sekiguchi <
>>>>>
>>>> [hidden email]>
>>>>
>>>>> wrote:
>>>>>
>>>>>   Hi IT CTO,
>>>>>>
>>>>>> Thank you for the reply!
>>>>>>
>>>>>> Because I couldn't find any ways to use charts (bar chart, pie chart,
>>>>>> etc.) in Zeppelin
>>>>>> without having my own Interpreter... If you know it, please let me
>>>>>> know.
>>>>>>
>>>>>> And now I see the comment of DepInterpreter, it seems it is used for
>>>>>>
>>>>> Spark:
>>>>
>>>>>
>>>>>> /**
>>>>>>      * DepInterpreter downloads dependencies and pass them when
>>>>>> SparkInterpreter initialized.
>>>>>>      * It extends SparkInterpreter but does not create sparkcontext
>>>>>>      *
>>>>>>      */
>>>>>>
>>>>>> nevertheless, it can be used for general purpose?
>>>>>>
>>>>>> regards,
>>>>>>
>>>>>> Koji
>>>>>>
>>>>>>
>>>>>> On 2015/06/30 20:25, IT CTO wrote:
>>>>>>
>>>>>>> This might be a stupid question but ....
>>>>>>> Quick question regarding the NLP4L - why did you need to create an
>>>>>>> interpreter for NLP4L? what is the deference from just using the %dep
>>>>>>>
>>>>>> and
>>>>
>>>>> loading from scala and then accessing it from scala paragrap?
>>>>>>> Eran
>>>>>>>
>>>>>>> On Tue, Jun 30, 2015 at 10:56 AM Ophir Cohen <[hidden email]>
>>>>>>> wrote:
>>>>>>>
>>>>>>>   Hi Daniel and good luck.
>>>>>>>> Can you share the meetup details?
>>>>>>>> 10x
>>>>>>>>
>>>>>>>> On Mon, Jun 29, 2015 at 9:30 AM, Daniel Haviv <
>>>>>>>> [hidden email]> wrote:
>>>>>>>>
>>>>>>>>   Hi everybody,
>>>>>>>>> We are going to arrange a meetup in Israel where I'm going to
>>>>>>>>>
>>>>>>>> dedicate
>>>>
>>>>> 45
>>>>>>
>>>>>>> minutes to introduce Zeppelin to the Big Data community in Israel.
>>>>>>>>> I'm going to display general usage of the Spark cli via zeppelin
>>>>>>>>> (and
>>>>>>>>> it's charting capabilities) and the WordCloud.
>>>>>>>>>
>>>>>>>>> Please share any cool examples you have of Zeppelin I can use.
>>>>>>>>> After the meetup, I will make sure these examples will be publicly
>>>>>>>>> available.
>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Thanks,
>>>>>>>>> Daniel Haviv
>>>>>>>>>
>>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>
>>>>
>>>
>>
>
>


--
最新ブログ記事〜LuceneとMahoutの文書分類機能比較
http://soleami.com/blog_ja/lucene-mahout-classifier_ja.html
==========================================
株式会社 ロンウイット
関口宏司
105-0003 東京都港区西新橋1-18-6
クロスオフィス内幸町 11階
TEL 03-5288-5927
FAX 03-5288-5928
http://www.rondhuit.com/
ブログ http://lucene.jugem.jp/