[developers] Examining profiles produced by art in incr[tsdb()]

goodman.m.w at gmail.com goodman.m.w at gmail.com
Thu Oct 18 01:49:33 CEST 2018


The 'analysis' file in ctn_profile_0.2.tar contains some garbage. It looks
like something wrote compressed data there but it's not the correct content
for the 'analysis' table. Try removing that file (or make it empty) and try
again?

On Wed, Oct 17, 2018 at 4:33 PM Kristen Howell <kphowell at uw.edu> wrote:

> Thanks Woodley. I'm glad to know that they should be compatible. I've
> attached two profiles that I'm having this problem with.
> Best,
> Kristen
>
> On Wed, Oct 17, 2018 at 3:44 PM Woodley Packard <sweaglesw at sweaglesw.org>
> wrote:
>
>> Hi,
>>
>> Profiles from art should load into [incr tsdb()] just fine.  In some
>> cases there may be unpopulated fields.  I would expect the two functions
>> you indicated to both be at least mostly functional though.  Can you
>> provide copies of the profiles in question?
>>
>> Woodley
>>
>> > On Oct 17, 2018, at 1:43 PM, Kristen Howell <kphowell at uw.edu> wrote:
>> >
>> > Hi everyone,
>> >
>> > I've created profiles of various grammars over a test-suite using art
>> v. 0.1.9. I would like to examine these profiles in incr[tsdb()], but not
>> all of  incr[tsdb()]'s utilities work on the profiles. Browse | Results
>> shows the results and Compare | Detail works. But Analyze | Coverage shows
>> 0 items and Compare | Competence says that the profiles are "empty".
>> >
>> > It's not clear to me if the profile produced with art is ill-formed in
>> some way or if there is a mismatch between the schema art outputs and what
>> incr[tsdb()] expects.
>> >
>> > Are profiles created by art expected to load in incr[tsdb()]? If not,
>> is there a way to convert the profile to the scheme incr[tsdb()] expects?
>> >
>> > Thanks in advance,
>> >
>> > Kristen
>>
>>

-- 
-Michael Wayne Goodman
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.delph-in.net/archives/developers/attachments/20181017/37a86314/attachment.html>


More information about the developers mailing list