Strict the pattern in the JAPE grammar to be in one sentience


Emml Asimadi
 

Yes, Diana's suggestion earlier should be perfect for you then.

Phase:xxxxx
Input: Split xxxx xxxx....
Options: xxxxxx

YourPatterns

Remember to have sentence splitter that generates "Split" annotations before the JAPE.

It will match 'PERSON friend PERSON' in the same sentence only. because the "Spilt" will make YourPattern NOT match accross sentence.

Just for your understanding if wanted to match accross sentence u will do

PERSON friend SPLIT PERSON.
This will match "PERSON friend" in one sentence and PERSON in the next sentence.

Regards,
Emml


On Fri, 20 Jul 2018 at 19:27, Nouf Bindris
<bindris.n@...> wrote:
Thank you for your time.


If I want to extract the following pattern from long text having many sentences and some of the  Person annotations in the boundary.
for example:


Person
Token [0,5]
Token.string == "friend"
Token[0,10]
Person
or any other pattern


I need to find the pattern in one sentence not the first person in one sentence and the other pattern in another sentence.



Thank you,


On 20 July 2018 at 19:07, EMMANUEL ASIMADI <easimadi@...> wrote:
Hi,
If you put split as input to your JAPE it will not match beyond sentence boundary. It shouldn't matter where the tokens occur in the sentence... u can share additional snippets to help understand the exact problem if it doesn't work.


e.g 
Phase: xxxxx
Input: Token Split
Options: xxxx

YourPatterns




On Fri, 20 Jul 2018 at 18:36, bindris.n@...
<bindris.n@...> wrote:
Thank you for your reply.
 
 
What if I have the person name the first word in the sentence.
 
how could I restrict my pattern to be in the sentence boundary?
 
 
thank you,


Nouf Bindris
 

Thank you for your time.


If I want to extract the following pattern from long text having many sentences and some of the  Person annotations in the boundary.
for example:


Person
Token [0,5]
Token.string == "friend"
Token[0,10]
Person
or any other pattern


I need to find the pattern in one sentence not the first person in one sentence and the other pattern in another sentence.



Thank you,


On 20 July 2018 at 19:07, EMMANUEL ASIMADI <easimadi@...> wrote:
Hi,
If you put split as input to your JAPE it will not match beyond sentence boundary. It shouldn't matter where the tokens occur in the sentence... u can share additional snippets to help understand the exact problem if it doesn't work.


e.g 
Phase: xxxxx
Input: Token Split
Options: xxxx

YourPatterns




On Fri, 20 Jul 2018 at 18:36, bindris.n@...
<bindris.n@...> wrote:
Thank you for your reply.
 
 
What if I have the person name the first word in the sentence.
 
how could I restrict my pattern to be in the sentence boundary?
 
 
thank you,


Emml Asimadi
 

Hi,
If you put split as input to your JAPE it will not match beyond sentence boundary. It shouldn't matter where the tokens occur in the sentence... u can share additional snippets to help understand the exact problem if it doesn't work.


e.g 
Phase: xxxxx
Input: Token Split
Options: xxxx

YourPatterns




On Fri, 20 Jul 2018 at 18:36, bindris.n@...
<bindris.n@...> wrote:
Thank you for your reply.
 
 
What if I have the person name the first word in the sentence.
 
how could I restrict my pattern to be in the sentence boundary?
 
 
thank you,


Nouf Bindris
 

Thank you for your reply.
 
 
What if I have the person name the first word in the sentence.
 
how could I restrict my pattern to be in the sentence boundary?
 
 
thank you,


Diana Maynard
 

Assuming your Person annotations don’t overlap a sentence boundary individually, just put Split in the input annotations
Diana

On 18 Jul 2018, at 17:52, bindris.n@... wrote:

Hi

please if I need to strict my pattern to pe extract in one sentence.
For example:

{Person}[0,5] {Token}[0,10] {person}

this pattern needs to be in one sentence boundary.

thank you,


Nouf Bindris
 

Hi

please if I need to strict my pattern to pe extract in one sentence.
For example:

{Person}[0,5]   {Token}[0,10] {person}

this pattern needs to be in one sentence boundary.

thank you,