You are viewing a plain text version of this content. The canonical link for it is here.
Posted to java-user@lucene.apache.org by Janaka Thilakarathna <bj...@gmail.com> on 2016/08/16 08:29:30 UTC

Using Lucene's Multi Dimensional Space Search for Air traffic handling.

Hi everyone,

I am from University of Moratuwa. I have a quite understanding in Lucene's
Text Search, Geo3DPoint
<https://lucene.apache.org/core/6_0_0/spatial3d/org/apache/lucene/spatial3d/Geo3DPoint.html#newShapeQuery-java.lang.String-org.apache.lucene.spatial3d.geom.GeoShape->,
LatLonPoint
<https://lucene.apache.org/core/6_0_0/sandbox/org/apache/lucene/document/LatLonPoint.html#newPolygonQuery-java.lang.String-double:A-double:A->
but I have never used Lucene for multi dimentional space search.

The idea is to use multi dimensional space search on 4 dimensional space(3
physical dimensions and time as another dimension) to calculate results for
following queries.

   - minimum distance between two air-planes
   - whether air planes goes through and interesting area. (For a example:
   A forbidden air space)

Paths of air-planes can be represented by arrays of 4D points.
(time,x,y,z). In other words I have different x,y,z coordinates for
different time values. My idea is to index these points and query for above
results.

Since there is no much tutorials on this new feature on Lucene 6, I am
quite confused where to start the project. I am really glad if someone can
help me with this.

I just want to know whether I can use Lucene for this use-case. Further if
you can point me out a place to start developing, it will be really help
full.

Thank you!

Regards

-- 
*Janaka Chathuranga Thilakarathna*
Undergraduate at Computer Science and Engineering Department,
UNIVERSITY OF Moratuwa, Sri Lanka*.*

mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
<ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
<bj...@gmail.com>*
skype  : *janaka.chathurangat* |  website : janakact.wordpress.com

my public profiles :  [image: Facebook]
<http://www.facebook.com/bjchathuranga> [image: LinkedIn]
<http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>

Re: Using Lucene's Multi Dimensional Space Search for Air traffic handling.

Posted by Janaka Thilakarathna <bj...@gmail.com>.
Hi Michael,

Got it. :)
Thank you very much for the reply.

Janaka


On Thu, Aug 25, 2016 at 9:22 PM, Michael McCandless <
lucene@mikemccandless.com> wrote:

> Geo3D uses more "straightforward" math since it can do everything on a
> unit (well, slightly squashed) sphere, but the downside is it must store 3
> floats in the doc values.
>
> Whereas LatLonPoint must use Haversine, supposedly more costly, except
> Lucene has a rather optimized implementation, and in practice seems to be
> faster than the 3D math: see our nightly geo benchmarks at
> http://home.apache.org/~mikemccand/geobench.html
>
> That said, having competing choices in open source is very healthy.  E.g.
> we have multiple highlighter implementations, multiple query parsers, etc.
>
> Mike McCandless
>
> http://blog.mikemccandless.com
>
> On Thu, Aug 25, 2016 at 12:09 AM, Janaka Thilakarathna <
> bjchathuranga@gmail.com> wrote:
>
>> Hi Michael,
>>
>> I am kinda confused about the difference between LatLon and Geo3D points.
>> Since, Geo3D does not in fact search 3 dimensions, what is the use of it
>> over LatLonPoint. Why there are two Fields as LatLon and Geo3D if both do
>> the Planet Surface search ?
>>
>> Regards,
>> Janaka
>>
>> On Mon, Aug 22, 2016 at 3:19 PM, Michael McCandless <
>> lucene@mikemccandless.com> wrote:
>>
>>> Hi Janaka,
>>>
>>> Unfortunately, Geo3D does not in fact search three dimensions: it only
>>> handles the same 2D earth surface as lat/lon.  It's just that, internally,
>>> it's using three x,y,z coordinates to do all its math.  But those three
>>> coordinates must be on the earth's surface for the math to be correct.  We
>>> have wanted to rename this package so it's not so confusing, but have not
>>> yet succeeded.  I believe the x,y,z are in the "unit sphere" coordinate
>>> space.
>>>
>>> Mike McCandless
>>>
>>> http://blog.mikemccandless.com
>>>
>>> On Mon, Aug 22, 2016 at 2:13 AM, Janaka Thilakarathna <
>>> bjchathuranga@gmail.com> wrote:
>>>
>>>> Hi Michael,
>>>>
>>>> I started playing with Lucene LatLon and Geo3D points. I have a problem
>>>> about the constructor of Geo3DPoint.
>>>>
>>>> In the other constructor, Geo3DPoint(String name, double x, double y,
>>>> double z)
>>>> <https://lucene.apache.org/core/6_1_0/spatial3d/org/apache/lucene/spatial3d/Geo3DPoint.html#Geo3DPoint-java.lang.String-double-double-double->.
>>>> How can I map*,* Lat, Lon and Altitude into x,y,z. If we use x,y,z
>>>> there should be an Axis system. For an example I have these questions,
>>>>
>>>>    - Where those axises are pointed to?
>>>>    - What are the units (km or m)?
>>>>
>>>> If you can give me an idea on that it will be really helpful. :-)
>>>> Thank you.
>>>>
>>>> Regards,
>>>>
>>>> Janaka.
>>>>
>>>> On Thu, Aug 18, 2016 at 9:55 AM, Janaka Thilakarathna <
>>>> bjchathuranga@gmail.com> wrote:
>>>>
>>>>> Hi Michael,
>>>>>
>>>>> Sorry for the late reply and thank you very much for your quick
>>>>> respond. :-)
>>>>>
>>>>> Yeah, it looks like an interesting data set to play with, but it is
>>>>> really large to start. :D
>>>>> I will try some simple projects and get back to you if I find any
>>>>> trouble.
>>>>>
>>>>> Janaka.
>>>>>
>>>>> On Tue, Aug 16, 2016 at 2:49 PM, Michael McCandless <
>>>>> lucene@mikemccandless.com> wrote:
>>>>>
>>>>>> What a fun use case for dimensional points!  I just saw NASA announce
>>>>>> this data set recently: https://plus.google.
>>>>>> com/+MichaelMcCandless/posts/h8eUtkhizKG
>>>>>>
>>>>>> And I was wondering how to play with it... 36 TB of airplane flight
>>>>>> routes :)
>>>>>>
>>>>>> You can easily index your data (3 spatial dims + 1 time dim) using
>>>>>> e.g. DoublePoint but then the only way to query those points currently is
>>>>>> the PointRangeQuery (4D boxes); maybe you can use that to find the
>>>>>> "interesting area" traversals?
>>>>>>
>>>>>> For "minimum distance between two air-planes", you might be able to
>>>>>> start with LatLonPoint.nearest (KNN search implementation) but generalize
>>>>>> it a bit to N dims not just the 2 (lat, lon) that it supports today?
>>>>>>
>>>>>> Mike McCandless
>>>>>>
>>>>>> http://blog.mikemccandless.com
>>>>>>
>>>>>> On Tue, Aug 16, 2016 at 4:29 AM, Janaka Thilakarathna <
>>>>>> bjchathuranga@gmail.com> wrote:
>>>>>>
>>>>>>> Hi everyone,
>>>>>>>
>>>>>>> I am from University of Moratuwa. I have a quite understanding in
>>>>>>> Lucene's
>>>>>>> Text Search, Geo3DPoint
>>>>>>> <https://lucene.apache.org/core/6_0_0/spatial3d/org/apache/l
>>>>>>> ucene/spatial3d/Geo3DPoint.html#newShapeQuery-java.lang.Stri
>>>>>>> ng-org.apache.lucene.spatial3d.geom.GeoShape->,
>>>>>>> LatLonPoint
>>>>>>> <https://lucene.apache.org/core/6_0_0/sandbox/org/apache/luc
>>>>>>> ene/document/LatLonPoint.html#newPolygonQuery-java.lang.Stri
>>>>>>> ng-double:A-double:A->
>>>>>>> but I have never used Lucene for multi dimentional space search.
>>>>>>>
>>>>>>> The idea is to use multi dimensional space search on 4 dimensional
>>>>>>> space(3
>>>>>>> physical dimensions and time as another dimension) to calculate
>>>>>>> results for
>>>>>>> following queries.
>>>>>>>
>>>>>>>    - minimum distance between two air-planes
>>>>>>>    - whether air planes goes through and interesting area. (For a
>>>>>>> example:
>>>>>>>    A forbidden air space)
>>>>>>>
>>>>>>> Paths of air-planes can be represented by arrays of 4D points.
>>>>>>> (time,x,y,z). In other words I have different x,y,z coordinates for
>>>>>>> different time values. My idea is to index these points and query
>>>>>>> for above
>>>>>>> results.
>>>>>>>
>>>>>>> Since there is no much tutorials on this new feature on Lucene 6, I
>>>>>>> am
>>>>>>> quite confused where to start the project. I am really glad if
>>>>>>> someone can
>>>>>>> help me with this.
>>>>>>>
>>>>>>> I just want to know whether I can use Lucene for this use-case.
>>>>>>> Further if
>>>>>>> you can point me out a place to start developing, it will be really
>>>>>>> help
>>>>>>> full.
>>>>>>>
>>>>>>> Thank you!
>>>>>>>
>>>>>>> Regards
>>>>>>>
>>>>>>> --
>>>>>>> *Janaka Chathuranga Thilakarathna*
>>>>>>> Undergraduate at Computer Science and Engineering Department,
>>>>>>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>>>>>>
>>>>>>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>>>>>>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>>>>>>> <bj...@gmail.com>*
>>>>>>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>>>>>>
>>>>>>> my public profiles :  [image: Facebook]
>>>>>>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>>>>>>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna
>>>>>>> /78/2a8/807/>
>>>>>>>
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> *Janaka Chathuranga Thilakarathna*
>>>>> Undergraduate at Computer Science and Engineering Department,
>>>>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>>>>
>>>>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>>>>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>>>>> <bj...@gmail.com>*
>>>>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>>>>
>>>>> my public profiles :  [image: Facebook]
>>>>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>>>>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> *Janaka Chathuranga Thilakarathna*
>>>> Undergraduate at Computer Science and Engineering Department,
>>>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>>>
>>>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>>>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>>>> <bj...@gmail.com>*
>>>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>>>
>>>> my public profiles :  [image: Facebook]
>>>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>>>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>>>>
>>>
>>>
>>
>>
>> --
>> *Janaka Chathuranga Thilakarathna*
>> Undergraduate at Computer Science and Engineering Department,
>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>
>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>> <bj...@gmail.com>*
>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>
>> my public profiles :  [image: Facebook]
>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>>
>
>


-- 
*Janaka Chathuranga Thilakarathna*
Undergraduate at Computer Science and Engineering Department,
UNIVERSITY OF Moratuwa, Sri Lanka*.*

mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
<ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
<bj...@gmail.com>*
skype  : *janaka.chathurangat* |  website : janakact.wordpress.com

my public profiles :  [image: Facebook]
<http://www.facebook.com/bjchathuranga> [image: LinkedIn]
<http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>

Re: Using Lucene's Multi Dimensional Space Search for Air traffic handling.

Posted by Michael McCandless <lu...@mikemccandless.com>.
Geo3D uses more "straightforward" math since it can do everything on a unit
(well, slightly squashed) sphere, but the downside is it must store 3
floats in the doc values.

Whereas LatLonPoint must use Haversine, supposedly more costly, except
Lucene has a rather optimized implementation, and in practice seems to be
faster than the 3D math: see our nightly geo benchmarks at
http://home.apache.org/~mikemccand/geobench.html

That said, having competing choices in open source is very healthy.  E.g.
we have multiple highlighter implementations, multiple query parsers, etc.

Mike McCandless

http://blog.mikemccandless.com

On Thu, Aug 25, 2016 at 12:09 AM, Janaka Thilakarathna <
bjchathuranga@gmail.com> wrote:

> Hi Michael,
>
> I am kinda confused about the difference between LatLon and Geo3D points.
> Since, Geo3D does not in fact search 3 dimensions, what is the use of it
> over LatLonPoint. Why there are two Fields as LatLon and Geo3D if both do
> the Planet Surface search ?
>
> Regards,
> Janaka
>
> On Mon, Aug 22, 2016 at 3:19 PM, Michael McCandless <
> lucene@mikemccandless.com> wrote:
>
>> Hi Janaka,
>>
>> Unfortunately, Geo3D does not in fact search three dimensions: it only
>> handles the same 2D earth surface as lat/lon.  It's just that, internally,
>> it's using three x,y,z coordinates to do all its math.  But those three
>> coordinates must be on the earth's surface for the math to be correct.  We
>> have wanted to rename this package so it's not so confusing, but have not
>> yet succeeded.  I believe the x,y,z are in the "unit sphere" coordinate
>> space.
>>
>> Mike McCandless
>>
>> http://blog.mikemccandless.com
>>
>> On Mon, Aug 22, 2016 at 2:13 AM, Janaka Thilakarathna <
>> bjchathuranga@gmail.com> wrote:
>>
>>> Hi Michael,
>>>
>>> I started playing with Lucene LatLon and Geo3D points. I have a problem
>>> about the constructor of Geo3DPoint.
>>>
>>> In the other constructor, Geo3DPoint(String name, double x, double y,
>>> double z)
>>> <https://lucene.apache.org/core/6_1_0/spatial3d/org/apache/lucene/spatial3d/Geo3DPoint.html#Geo3DPoint-java.lang.String-double-double-double->.
>>> How can I map*,* Lat, Lon and Altitude into x,y,z. If we use x,y,z
>>> there should be an Axis system. For an example I have these questions,
>>>
>>>    - Where those axises are pointed to?
>>>    - What are the units (km or m)?
>>>
>>> If you can give me an idea on that it will be really helpful. :-)
>>> Thank you.
>>>
>>> Regards,
>>>
>>> Janaka.
>>>
>>> On Thu, Aug 18, 2016 at 9:55 AM, Janaka Thilakarathna <
>>> bjchathuranga@gmail.com> wrote:
>>>
>>>> Hi Michael,
>>>>
>>>> Sorry for the late reply and thank you very much for your quick
>>>> respond. :-)
>>>>
>>>> Yeah, it looks like an interesting data set to play with, but it is
>>>> really large to start. :D
>>>> I will try some simple projects and get back to you if I find any
>>>> trouble.
>>>>
>>>> Janaka.
>>>>
>>>> On Tue, Aug 16, 2016 at 2:49 PM, Michael McCandless <
>>>> lucene@mikemccandless.com> wrote:
>>>>
>>>>> What a fun use case for dimensional points!  I just saw NASA announce
>>>>> this data set recently: https://plus.google.
>>>>> com/+MichaelMcCandless/posts/h8eUtkhizKG
>>>>>
>>>>> And I was wondering how to play with it... 36 TB of airplane flight
>>>>> routes :)
>>>>>
>>>>> You can easily index your data (3 spatial dims + 1 time dim) using
>>>>> e.g. DoublePoint but then the only way to query those points currently is
>>>>> the PointRangeQuery (4D boxes); maybe you can use that to find the
>>>>> "interesting area" traversals?
>>>>>
>>>>> For "minimum distance between two air-planes", you might be able to
>>>>> start with LatLonPoint.nearest (KNN search implementation) but generalize
>>>>> it a bit to N dims not just the 2 (lat, lon) that it supports today?
>>>>>
>>>>> Mike McCandless
>>>>>
>>>>> http://blog.mikemccandless.com
>>>>>
>>>>> On Tue, Aug 16, 2016 at 4:29 AM, Janaka Thilakarathna <
>>>>> bjchathuranga@gmail.com> wrote:
>>>>>
>>>>>> Hi everyone,
>>>>>>
>>>>>> I am from University of Moratuwa. I have a quite understanding in
>>>>>> Lucene's
>>>>>> Text Search, Geo3DPoint
>>>>>> <https://lucene.apache.org/core/6_0_0/spatial3d/org/apache/l
>>>>>> ucene/spatial3d/Geo3DPoint.html#newShapeQuery-java.lang.Stri
>>>>>> ng-org.apache.lucene.spatial3d.geom.GeoShape->,
>>>>>> LatLonPoint
>>>>>> <https://lucene.apache.org/core/6_0_0/sandbox/org/apache/luc
>>>>>> ene/document/LatLonPoint.html#newPolygonQuery-java.lang.Stri
>>>>>> ng-double:A-double:A->
>>>>>> but I have never used Lucene for multi dimentional space search.
>>>>>>
>>>>>> The idea is to use multi dimensional space search on 4 dimensional
>>>>>> space(3
>>>>>> physical dimensions and time as another dimension) to calculate
>>>>>> results for
>>>>>> following queries.
>>>>>>
>>>>>>    - minimum distance between two air-planes
>>>>>>    - whether air planes goes through and interesting area. (For a
>>>>>> example:
>>>>>>    A forbidden air space)
>>>>>>
>>>>>> Paths of air-planes can be represented by arrays of 4D points.
>>>>>> (time,x,y,z). In other words I have different x,y,z coordinates for
>>>>>> different time values. My idea is to index these points and query for
>>>>>> above
>>>>>> results.
>>>>>>
>>>>>> Since there is no much tutorials on this new feature on Lucene 6, I am
>>>>>> quite confused where to start the project. I am really glad if
>>>>>> someone can
>>>>>> help me with this.
>>>>>>
>>>>>> I just want to know whether I can use Lucene for this use-case.
>>>>>> Further if
>>>>>> you can point me out a place to start developing, it will be really
>>>>>> help
>>>>>> full.
>>>>>>
>>>>>> Thank you!
>>>>>>
>>>>>> Regards
>>>>>>
>>>>>> --
>>>>>> *Janaka Chathuranga Thilakarathna*
>>>>>> Undergraduate at Computer Science and Engineering Department,
>>>>>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>>>>>
>>>>>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>>>>>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>>>>>> <bj...@gmail.com>*
>>>>>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>>>>>
>>>>>> my public profiles :  [image: Facebook]
>>>>>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>>>>>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna
>>>>>> /78/2a8/807/>
>>>>>>
>>>>>
>>>>>
>>>>
>>>>
>>>> --
>>>> *Janaka Chathuranga Thilakarathna*
>>>> Undergraduate at Computer Science and Engineering Department,
>>>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>>>
>>>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>>>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>>>> <bj...@gmail.com>*
>>>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>>>
>>>> my public profiles :  [image: Facebook]
>>>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>>>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>>>>
>>>
>>>
>>>
>>> --
>>> *Janaka Chathuranga Thilakarathna*
>>> Undergraduate at Computer Science and Engineering Department,
>>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>>
>>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>>> <bj...@gmail.com>*
>>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>>
>>> my public profiles :  [image: Facebook]
>>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>>>
>>
>>
>
>
> --
> *Janaka Chathuranga Thilakarathna*
> Undergraduate at Computer Science and Engineering Department,
> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>
> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
> <bj...@gmail.com>*
> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>
> my public profiles :  [image: Facebook]
> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>

Re: Using Lucene's Multi Dimensional Space Search for Air traffic handling.

Posted by Janaka Thilakarathna <bj...@gmail.com>.
Hi Michael,

I am kinda confused about the difference between LatLon and Geo3D points.
Since, Geo3D does not in fact search 3 dimensions, what is the use of it
over LatLonPoint. Why there are two Fields as LatLon and Geo3D if both do
the Planet Surface search ?

Regards,
Janaka

On Mon, Aug 22, 2016 at 3:19 PM, Michael McCandless <
lucene@mikemccandless.com> wrote:

> Hi Janaka,
>
> Unfortunately, Geo3D does not in fact search three dimensions: it only
> handles the same 2D earth surface as lat/lon.  It's just that, internally,
> it's using three x,y,z coordinates to do all its math.  But those three
> coordinates must be on the earth's surface for the math to be correct.  We
> have wanted to rename this package so it's not so confusing, but have not
> yet succeeded.  I believe the x,y,z are in the "unit sphere" coordinate
> space.
>
> Mike McCandless
>
> http://blog.mikemccandless.com
>
> On Mon, Aug 22, 2016 at 2:13 AM, Janaka Thilakarathna <
> bjchathuranga@gmail.com> wrote:
>
>> Hi Michael,
>>
>> I started playing with Lucene LatLon and Geo3D points. I have a problem
>> about the constructor of Geo3DPoint.
>>
>> In the other constructor, Geo3DPoint(String name, double x, double y,
>> double z)
>> <https://lucene.apache.org/core/6_1_0/spatial3d/org/apache/lucene/spatial3d/Geo3DPoint.html#Geo3DPoint-java.lang.String-double-double-double->.
>> How can I map*,* Lat, Lon and Altitude into x,y,z. If we use x,y,z there
>> should be an Axis system. For an example I have these questions,
>>
>>    - Where those axises are pointed to?
>>    - What are the units (km or m)?
>>
>> If you can give me an idea on that it will be really helpful. :-)
>> Thank you.
>>
>> Regards,
>>
>> Janaka.
>>
>> On Thu, Aug 18, 2016 at 9:55 AM, Janaka Thilakarathna <
>> bjchathuranga@gmail.com> wrote:
>>
>>> Hi Michael,
>>>
>>> Sorry for the late reply and thank you very much for your quick respond.
>>> :-)
>>>
>>> Yeah, it looks like an interesting data set to play with, but it is
>>> really large to start. :D
>>> I will try some simple projects and get back to you if I find any
>>> trouble.
>>>
>>> Janaka.
>>>
>>> On Tue, Aug 16, 2016 at 2:49 PM, Michael McCandless <
>>> lucene@mikemccandless.com> wrote:
>>>
>>>> What a fun use case for dimensional points!  I just saw NASA announce
>>>> this data set recently: https://plus.google.
>>>> com/+MichaelMcCandless/posts/h8eUtkhizKG
>>>>
>>>> And I was wondering how to play with it... 36 TB of airplane flight
>>>> routes :)
>>>>
>>>> You can easily index your data (3 spatial dims + 1 time dim) using e.g.
>>>> DoublePoint but then the only way to query those points currently is the
>>>> PointRangeQuery (4D boxes); maybe you can use that to find the "interesting
>>>> area" traversals?
>>>>
>>>> For "minimum distance between two air-planes", you might be able to
>>>> start with LatLonPoint.nearest (KNN search implementation) but generalize
>>>> it a bit to N dims not just the 2 (lat, lon) that it supports today?
>>>>
>>>> Mike McCandless
>>>>
>>>> http://blog.mikemccandless.com
>>>>
>>>> On Tue, Aug 16, 2016 at 4:29 AM, Janaka Thilakarathna <
>>>> bjchathuranga@gmail.com> wrote:
>>>>
>>>>> Hi everyone,
>>>>>
>>>>> I am from University of Moratuwa. I have a quite understanding in
>>>>> Lucene's
>>>>> Text Search, Geo3DPoint
>>>>> <https://lucene.apache.org/core/6_0_0/spatial3d/org/apache/l
>>>>> ucene/spatial3d/Geo3DPoint.html#newShapeQuery-java.lang.Stri
>>>>> ng-org.apache.lucene.spatial3d.geom.GeoShape->,
>>>>> LatLonPoint
>>>>> <https://lucene.apache.org/core/6_0_0/sandbox/org/apache/luc
>>>>> ene/document/LatLonPoint.html#newPolygonQuery-java.lang.Stri
>>>>> ng-double:A-double:A->
>>>>> but I have never used Lucene for multi dimentional space search.
>>>>>
>>>>> The idea is to use multi dimensional space search on 4 dimensional
>>>>> space(3
>>>>> physical dimensions and time as another dimension) to calculate
>>>>> results for
>>>>> following queries.
>>>>>
>>>>>    - minimum distance between two air-planes
>>>>>    - whether air planes goes through and interesting area. (For a
>>>>> example:
>>>>>    A forbidden air space)
>>>>>
>>>>> Paths of air-planes can be represented by arrays of 4D points.
>>>>> (time,x,y,z). In other words I have different x,y,z coordinates for
>>>>> different time values. My idea is to index these points and query for
>>>>> above
>>>>> results.
>>>>>
>>>>> Since there is no much tutorials on this new feature on Lucene 6, I am
>>>>> quite confused where to start the project. I am really glad if someone
>>>>> can
>>>>> help me with this.
>>>>>
>>>>> I just want to know whether I can use Lucene for this use-case.
>>>>> Further if
>>>>> you can point me out a place to start developing, it will be really
>>>>> help
>>>>> full.
>>>>>
>>>>> Thank you!
>>>>>
>>>>> Regards
>>>>>
>>>>> --
>>>>> *Janaka Chathuranga Thilakarathna*
>>>>> Undergraduate at Computer Science and Engineering Department,
>>>>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>>>>
>>>>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>>>>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>>>>> <bj...@gmail.com>*
>>>>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>>>>
>>>>> my public profiles :  [image: Facebook]
>>>>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>>>>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna
>>>>> /78/2a8/807/>
>>>>>
>>>>
>>>>
>>>
>>>
>>> --
>>> *Janaka Chathuranga Thilakarathna*
>>> Undergraduate at Computer Science and Engineering Department,
>>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>>
>>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>>> <bj...@gmail.com>*
>>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>>
>>> my public profiles :  [image: Facebook]
>>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>>>
>>
>>
>>
>> --
>> *Janaka Chathuranga Thilakarathna*
>> Undergraduate at Computer Science and Engineering Department,
>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>
>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>> <bj...@gmail.com>*
>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>
>> my public profiles :  [image: Facebook]
>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>>
>
>


-- 
*Janaka Chathuranga Thilakarathna*
Undergraduate at Computer Science and Engineering Department,
UNIVERSITY OF Moratuwa, Sri Lanka*.*

mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
<ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
<bj...@gmail.com>*
skype  : *janaka.chathurangat* |  website : janakact.wordpress.com

my public profiles :  [image: Facebook]
<http://www.facebook.com/bjchathuranga> [image: LinkedIn]
<http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>

Re: Using Lucene's Multi Dimensional Space Search for Air traffic handling.

Posted by Michael McCandless <lu...@mikemccandless.com>.
Hi Janaka,

Unfortunately, Geo3D does not in fact search three dimensions: it only
handles the same 2D earth surface as lat/lon.  It's just that, internally,
it's using three x,y,z coordinates to do all its math.  But those three
coordinates must be on the earth's surface for the math to be correct.  We
have wanted to rename this package so it's not so confusing, but have not
yet succeeded.  I believe the x,y,z are in the "unit sphere" coordinate
space.

Mike McCandless

http://blog.mikemccandless.com

On Mon, Aug 22, 2016 at 2:13 AM, Janaka Thilakarathna <
bjchathuranga@gmail.com> wrote:

> Hi Michael,
>
> I started playing with Lucene LatLon and Geo3D points. I have a problem
> about the constructor of Geo3DPoint.
>
> In the other constructor, Geo3DPoint(String name, double x, double y,
> double z)
> <https://lucene.apache.org/core/6_1_0/spatial3d/org/apache/lucene/spatial3d/Geo3DPoint.html#Geo3DPoint-java.lang.String-double-double-double->.
> How can I map*,* Lat, Lon and Altitude into x,y,z. If we use x,y,z there
> should be an Axis system. For an example I have these questions,
>
>    - Where those axises are pointed to?
>    - What are the units (km or m)?
>
> If you can give me an idea on that it will be really helpful. :-)
> Thank you.
>
> Regards,
>
> Janaka.
>
> On Thu, Aug 18, 2016 at 9:55 AM, Janaka Thilakarathna <
> bjchathuranga@gmail.com> wrote:
>
>> Hi Michael,
>>
>> Sorry for the late reply and thank you very much for your quick respond.
>> :-)
>>
>> Yeah, it looks like an interesting data set to play with, but it is
>> really large to start. :D
>> I will try some simple projects and get back to you if I find any
>> trouble.
>>
>> Janaka.
>>
>> On Tue, Aug 16, 2016 at 2:49 PM, Michael McCandless <
>> lucene@mikemccandless.com> wrote:
>>
>>> What a fun use case for dimensional points!  I just saw NASA announce
>>> this data set recently: https://plus.google.
>>> com/+MichaelMcCandless/posts/h8eUtkhizKG
>>>
>>> And I was wondering how to play with it... 36 TB of airplane flight
>>> routes :)
>>>
>>> You can easily index your data (3 spatial dims + 1 time dim) using e.g.
>>> DoublePoint but then the only way to query those points currently is the
>>> PointRangeQuery (4D boxes); maybe you can use that to find the "interesting
>>> area" traversals?
>>>
>>> For "minimum distance between two air-planes", you might be able to
>>> start with LatLonPoint.nearest (KNN search implementation) but generalize
>>> it a bit to N dims not just the 2 (lat, lon) that it supports today?
>>>
>>> Mike McCandless
>>>
>>> http://blog.mikemccandless.com
>>>
>>> On Tue, Aug 16, 2016 at 4:29 AM, Janaka Thilakarathna <
>>> bjchathuranga@gmail.com> wrote:
>>>
>>>> Hi everyone,
>>>>
>>>> I am from University of Moratuwa. I have a quite understanding in
>>>> Lucene's
>>>> Text Search, Geo3DPoint
>>>> <https://lucene.apache.org/core/6_0_0/spatial3d/org/apache/l
>>>> ucene/spatial3d/Geo3DPoint.html#newShapeQuery-java.lang.Stri
>>>> ng-org.apache.lucene.spatial3d.geom.GeoShape->,
>>>> LatLonPoint
>>>> <https://lucene.apache.org/core/6_0_0/sandbox/org/apache/luc
>>>> ene/document/LatLonPoint.html#newPolygonQuery-java.lang.Stri
>>>> ng-double:A-double:A->
>>>> but I have never used Lucene for multi dimentional space search.
>>>>
>>>> The idea is to use multi dimensional space search on 4 dimensional
>>>> space(3
>>>> physical dimensions and time as another dimension) to calculate results
>>>> for
>>>> following queries.
>>>>
>>>>    - minimum distance between two air-planes
>>>>    - whether air planes goes through and interesting area. (For a
>>>> example:
>>>>    A forbidden air space)
>>>>
>>>> Paths of air-planes can be represented by arrays of 4D points.
>>>> (time,x,y,z). In other words I have different x,y,z coordinates for
>>>> different time values. My idea is to index these points and query for
>>>> above
>>>> results.
>>>>
>>>> Since there is no much tutorials on this new feature on Lucene 6, I am
>>>> quite confused where to start the project. I am really glad if someone
>>>> can
>>>> help me with this.
>>>>
>>>> I just want to know whether I can use Lucene for this use-case. Further
>>>> if
>>>> you can point me out a place to start developing, it will be really help
>>>> full.
>>>>
>>>> Thank you!
>>>>
>>>> Regards
>>>>
>>>> --
>>>> *Janaka Chathuranga Thilakarathna*
>>>> Undergraduate at Computer Science and Engineering Department,
>>>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>>>
>>>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>>>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>>>> <bj...@gmail.com>*
>>>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>>>
>>>> my public profiles :  [image: Facebook]
>>>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>>>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna
>>>> /78/2a8/807/>
>>>>
>>>
>>>
>>
>>
>> --
>> *Janaka Chathuranga Thilakarathna*
>> Undergraduate at Computer Science and Engineering Department,
>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>
>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>> <bj...@gmail.com>*
>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>
>> my public profiles :  [image: Facebook]
>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>>
>
>
>
> --
> *Janaka Chathuranga Thilakarathna*
> Undergraduate at Computer Science and Engineering Department,
> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>
> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
> <bj...@gmail.com>*
> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>
> my public profiles :  [image: Facebook]
> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>

Re: Using Lucene's Multi Dimensional Space Search for Air traffic handling.

Posted by Janaka Thilakarathna <bj...@gmail.com>.
Hi Michael,

I started playing with Lucene LatLon and Geo3D points. I have a problem
about the constructor of Geo3DPoint.

In the other constructor, Geo3DPoint(String name, double x, double y,
double z)
<https://lucene.apache.org/core/6_1_0/spatial3d/org/apache/lucene/spatial3d/Geo3DPoint.html#Geo3DPoint-java.lang.String-double-double-double->.
How can I map*,* Lat, Lon and Altitude into x,y,z. If we use x,y,z there
should be an Axis system. For an example I have these questions,

   - Where those axises are pointed to?
   - What are the units (km or m)?

If you can give me an idea on that it will be really helpful. :-)
Thank you.

Regards,

Janaka.

On Thu, Aug 18, 2016 at 9:55 AM, Janaka Thilakarathna <
bjchathuranga@gmail.com> wrote:

> Hi Michael,
>
> Sorry for the late reply and thank you very much for your quick respond.
> :-)
>
> Yeah, it looks like an interesting data set to play with, but it is really
> large to start. :D
> I will try some simple projects and get back to you if I find any trouble.
>
> Janaka.
>
> On Tue, Aug 16, 2016 at 2:49 PM, Michael McCandless <
> lucene@mikemccandless.com> wrote:
>
>> What a fun use case for dimensional points!  I just saw NASA announce
>> this data set recently: https://plus.google.
>> com/+MichaelMcCandless/posts/h8eUtkhizKG
>>
>> And I was wondering how to play with it... 36 TB of airplane flight
>> routes :)
>>
>> You can easily index your data (3 spatial dims + 1 time dim) using e.g.
>> DoublePoint but then the only way to query those points currently is the
>> PointRangeQuery (4D boxes); maybe you can use that to find the "interesting
>> area" traversals?
>>
>> For "minimum distance between two air-planes", you might be able to start
>> with LatLonPoint.nearest (KNN search implementation) but generalize it a
>> bit to N dims not just the 2 (lat, lon) that it supports today?
>>
>> Mike McCandless
>>
>> http://blog.mikemccandless.com
>>
>> On Tue, Aug 16, 2016 at 4:29 AM, Janaka Thilakarathna <
>> bjchathuranga@gmail.com> wrote:
>>
>>> Hi everyone,
>>>
>>> I am from University of Moratuwa. I have a quite understanding in
>>> Lucene's
>>> Text Search, Geo3DPoint
>>> <https://lucene.apache.org/core/6_0_0/spatial3d/org/apache/l
>>> ucene/spatial3d/Geo3DPoint.html#newShapeQuery-java.lang.Stri
>>> ng-org.apache.lucene.spatial3d.geom.GeoShape->,
>>> LatLonPoint
>>> <https://lucene.apache.org/core/6_0_0/sandbox/org/apache/luc
>>> ene/document/LatLonPoint.html#newPolygonQuery-java.lang.Stri
>>> ng-double:A-double:A->
>>> but I have never used Lucene for multi dimentional space search.
>>>
>>> The idea is to use multi dimensional space search on 4 dimensional
>>> space(3
>>> physical dimensions and time as another dimension) to calculate results
>>> for
>>> following queries.
>>>
>>>    - minimum distance between two air-planes
>>>    - whether air planes goes through and interesting area. (For a
>>> example:
>>>    A forbidden air space)
>>>
>>> Paths of air-planes can be represented by arrays of 4D points.
>>> (time,x,y,z). In other words I have different x,y,z coordinates for
>>> different time values. My idea is to index these points and query for
>>> above
>>> results.
>>>
>>> Since there is no much tutorials on this new feature on Lucene 6, I am
>>> quite confused where to start the project. I am really glad if someone
>>> can
>>> help me with this.
>>>
>>> I just want to know whether I can use Lucene for this use-case. Further
>>> if
>>> you can point me out a place to start developing, it will be really help
>>> full.
>>>
>>> Thank you!
>>>
>>> Regards
>>>
>>> --
>>> *Janaka Chathuranga Thilakarathna*
>>> Undergraduate at Computer Science and Engineering Department,
>>> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>>>
>>> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
>>> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
>>> <bj...@gmail.com>*
>>> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>>>
>>> my public profiles :  [image: Facebook]
>>> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
>>> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/
>>> >
>>>
>>
>>
>
>
> --
> *Janaka Chathuranga Thilakarathna*
> Undergraduate at Computer Science and Engineering Department,
> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>
> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
> <bj...@gmail.com>*
> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>
> my public profiles :  [image: Facebook]
> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>



-- 
*Janaka Chathuranga Thilakarathna*
Undergraduate at Computer Science and Engineering Department,
UNIVERSITY OF Moratuwa, Sri Lanka*.*

mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
<ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
<bj...@gmail.com>*
skype  : *janaka.chathurangat* |  website : janakact.wordpress.com

my public profiles :  [image: Facebook]
<http://www.facebook.com/bjchathuranga> [image: LinkedIn]
<http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>

Re: Using Lucene's Multi Dimensional Space Search for Air traffic handling.

Posted by Michael McCandless <lu...@mikemccandless.com>.
What a fun use case for dimensional points!  I just saw NASA announce this
data set recently: https://plus.google.com/+MichaelMcCandless/posts/
h8eUtkhizKG

And I was wondering how to play with it... 36 TB of airplane flight routes
:)

You can easily index your data (3 spatial dims + 1 time dim) using e.g.
DoublePoint but then the only way to query those points currently is the
PointRangeQuery (4D boxes); maybe you can use that to find the "interesting
area" traversals?

For "minimum distance between two air-planes", you might be able to start
with LatLonPoint.nearest (KNN search implementation) but generalize it a
bit to N dims not just the 2 (lat, lon) that it supports today?

Mike McCandless

http://blog.mikemccandless.com

On Tue, Aug 16, 2016 at 4:29 AM, Janaka Thilakarathna <
bjchathuranga@gmail.com> wrote:

> Hi everyone,
>
> I am from University of Moratuwa. I have a quite understanding in Lucene's
> Text Search, Geo3DPoint
> <https://lucene.apache.org/core/6_0_0/spatial3d/org/apache/
> lucene/spatial3d/Geo3DPoint.html#newShapeQuery-java.lang.
> String-org.apache.lucene.spatial3d.geom.GeoShape->,
> LatLonPoint
> <https://lucene.apache.org/core/6_0_0/sandbox/org/apache/luc
> ene/document/LatLonPoint.html#newPolygonQuery-java.lang.
> String-double:A-double:A->
> but I have never used Lucene for multi dimentional space search.
>
> The idea is to use multi dimensional space search on 4 dimensional space(3
> physical dimensions and time as another dimension) to calculate results for
> following queries.
>
>    - minimum distance between two air-planes
>    - whether air planes goes through and interesting area. (For a example:
>    A forbidden air space)
>
> Paths of air-planes can be represented by arrays of 4D points.
> (time,x,y,z). In other words I have different x,y,z coordinates for
> different time values. My idea is to index these points and query for above
> results.
>
> Since there is no much tutorials on this new feature on Lucene 6, I am
> quite confused where to start the project. I am really glad if someone can
> help me with this.
>
> I just want to know whether I can use Lucene for this use-case. Further if
> you can point me out a place to start developing, it will be really help
> full.
>
> Thank you!
>
> Regards
>
> --
> *Janaka Chathuranga Thilakarathna*
> Undergraduate at Computer Science and Engineering Department,
> UNIVERSITY OF Moratuwa, Sri Lanka*.*
>
> mobile :(+94)* 713315725 **| *email :  *janaka.13@cse.mrt.ac.lk*
> <ja...@cse.mrt.ac.lk>*, bjchathuranga@gmail.com
> <bj...@gmail.com>*
> skype  : *janaka.chathurangat* |  website : janakact.wordpress.com
>
> my public profiles :  [image: Facebook]
> <http://www.facebook.com/bjchathuranga> [image: LinkedIn]
> <http://lk.linkedin.com/pub/janaka-chathuranga-thilakarathna/78/2a8/807/>
>