Both sides previous revision
Previous revision
Next revision
|
Previous revision
Next revision
Both sides next revision
|
user:zeman:interset:how-to-write-a-driver [2008/03/06 16:28] zeman Correction: Actually this has already been implemented. |
user:zeman:interset:how-to-write-a-driver [2008/03/08 11:06] zeman Testing all drivers. |
| |
See [[user:zeman:interset:Common Problems]] for a list of suggestions for phenomena difficult to match between tagsets and the Interset. | See [[user:zeman:interset:Common Problems]] for a list of suggestions for phenomena difficult to match between tagsets and the Interset. |
| |
| |
| |
| |
When you have written a driver for a new tagset, you should test it. The driver package contains a test script called ''driver-test.pl''. When running it, give the driver name as argument, without the ''tagset::'' prefix. You can also use the ''-d'' option to turn on debug messages (list of tags being tested). | When you have written a driver for a new tagset, you should test it. The driver package contains a test script called ''driver-test.pl''. When running it, give the driver name as argument, without the ''tagset::'' prefix. You can also use the ''-d'' option to turn on debug messages (list of tags being tested). |
| |
<code>driver-test.pl ar::conll</code> | <code>driver-test.pl ar::conll |
| driver-test.pl -a</code> |
| |
Running ''driver-test.pl'' without arguments will list the drivers available on the system. | Running ''driver-test.pl'' without arguments will list the drivers available on the system. Running it with the ''-a'' option will test all the drivers. |
| |
Note that only drivers implementing the ''list()'' function can be tested. | Note that only drivers implementing the ''list()'' function can be tested. Most testing involves generating the list of all possible tags and testing the driver on each tag separately. |
| |
The following tests will be performed: | The following tests will be performed: |
| |
* Get list of possible tags by calling ''list()''. Go through the list and check for each tag that ''encode(decode($tag)) eq $tag''. While sometimes it can be annoying to try to preserve some obscure information hidden in the tags, this test can also reveal many unwanted bugs. Besides, you should preserve information of your own tagset because people may want to use your driver merely to //access// the tags, instead of //converting// them. | * Decode each tag and check that only known features and values are set. In addition to a built-in list, every feature can have an empty value, and the features "tagset" and "other" can have any value. |
| * Check for each tag that ''encode(decode($tag)) eq $tag''. While sometimes it can be annoying to try to preserve some obscure information hidden in the tags, this test can also reveal many unwanted bugs. Besides, you should preserve information of your own tagset because people may want to use your driver merely to //access// the tags, instead of //converting// them. |
| |