ATT format - Revision history

Francis Tyers at 21:24, 13 March 2017

2017-03-13T21:24:43Z

Rcrowther at 11:49, 26 September 2016

2016-09-26T11:49:20Z

TommiPirinen: weight example

2016-03-07T17:26:14Z

weight example

Unhammer at 14:18, 10 March 2014

2014-03-10T14:18:41Z

Francis Tyers at 22:22, 28 March 2013

2013-03-28T22:22:36Z

Francis Tyers at 21:38, 25 March 2013

2013-03-25T21:38:46Z

Francis Tyers: Created page with ''''ATT format''' is a transducer format based on a four-column layout. It is a tab separated four-column format. ==Example output==

  $ cat test.dix     
2013-03-25T21:38:17Z
Created page with ''''ATT format''' is a transducer format based on a four-column layout. It is a tab separated four-column format.    ==Example output==  <pre>  $ cat test.dix  <dictionary>   <alp…'
New page
'''ATT format''' is a transducer format based on a four-column layout. It is a tab separated four-column format.







==Example output==



<pre>



$ cat test.dix 

<dictionary>

  <alphabet>abcdefghijklmnopqrstuvwxyz</alphabet>

  <sdefs>

    <sdef n="n"/>

  </sdefs>

  <section id="main" type="standard">

    <e><p><l>test</l><r>foo</r></p></e>

  </section>

</dictionary>





$ lt-comp lr test.dix test.bin

main@standard 5 4





$ lt-print test.bin 

0	1	t	f	

1	2	e	o	

2	3	s	o	

3	4	t	ε	

4



</pre>







[[Category:Documentation]]

@@ Line 1: / Line 1: @@
 '''ATT format''' is a transducer format based on a four-column layout. It is a tab separated four-column format.
-Both lttoolbox and HFST can read ATT format as input to compile dictionaries (lt-comp, hfst-txt2fst), and print compiled dictionaries to ATT format (lt-print, hfst-fst2txt).
+Both lttoolbox and HFST can read ATT format as input to compile dictionaries (<code>lt-comp</code>, <code>hfst-txt2fst</code>), and print compiled dictionaries to ATT format (<code>lt-print</code>, <code>hfst-fst2txt</code>).
 ==Example==
 Say we want to represent the following transducer:
-[[File:Test att.png|thumb|center|400px]]
+[[File:Test att.png|thumb|center|500px]]
 We can do it thusly:

← Older revision		Revision as of 11:49, 26 September 2016
Line 62:		Line 62:

	[[Category:Documentation]]		[[Category:Documentation]]
			[[Category:Documentation in English]]

@@ Line 33: / Line 33: @@
 </pre>
+==Weights==
+AT&T format supports "weights", for example to estimate likelihoods. The default interpretation is bigger the weight (heavier) the worse it is (aka penalties). E.g.:
+<pre>
+	1	c	c	1.000000
+	2	d	d	2.000000
+	3	a	a	0.000000
+	4	o	o	0.000000
+	5	t	t	0.000000
+	5	g	g	0.000000
+	6	s	s	10.000000
+	0.000000
+	0.000000
+</pre>
+would be appropriate to have weights 1 for cat, 2 for dog, and additional 10 pounds for beign a plural.
+Commonly weights are estimated e.g. from probabilities using -log().
 ==See also==

@@ Line 1: / Line 1: @@
 '''ATT format''' is a transducer format based on a four-column layout. It is a tab separated four-column format.
+Both lttoolbox and HFST can read ATT format as input to compile dictionaries (lt-comp, hfst-txt2fst), and print compiled dictionaries to ATT format (lt-print, hfst-fst2txt).
-==Example output==
+==Example==
 Say we want to represent the following transducer:
@@ Line 8: / Line 9: @@
 We can do it thusly:
 <pre>
 $ cat test.dix
 <dictionary>
@@ Line 33: / Line 32: @@
 	4	t	ε
 </pre>

@@ Line 2: / Line 2: @@
 ==Example output==
+Say we want to represent the following transducer:
+[[File:Test att.png|thumb|center|400px]]
+We can do it thusly:
 <pre>