Hi Andrew, Thanks for the bug report; how embarrassing. I have pushed a fix to my repository; it only touches Orion. Casey and I are fixing up some other issues in the Nyx frontend before that gets finished up. I am inclined to leave Maestro and Castro and attempt to unify Boxlib readers rather than salve this wound. This week the workshop is taking a bit of time, but early next week I will take a look at the (admittedly confusing) hierarchy and frontend code and see if I can codify this. The changes to field detection resulted in this. Could you please test it by doing: hg pull -r 736067d72401 http://bitbucket.org/MatthewTurk/yt hg up and then re-running your script? This also includes changes to the FLASH stuff, and I'm waiting on the integration tests to run before I am willing to push it to the main repository. Could you let me know if it works for Orion? -Matt On Wed, Jan 25, 2012 at 12:01 AM, Andrew Myers wrote:

Hello All,

It looks like there is a problem with the current stable version of yt that comes up when creating a PlotCollection using Orion datasets. See the script here and the traceback here.

I think the fix is that there needs to be the line

self.field_info = pf.field_info

somewhere in the constructor for OrionHierarchy (around line 120 of yt/frontends/orion/data_structures.py). It looks like a line like this used to be here, but was removed during some code cleanup. If the other Orion users are anything like me, they haven't updated yt in a while and thus haven't noticed the problem.

Glancing at the Nyx, Maestro, and Castro frontends, those look like they might suffer from the same issue, but I don't have any of their data to confirm or deny this.

-Andrew Myers

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

Hi Andrew, Okay, this looks like there was a non-standard, uncommented line in the inputs file. (This didn't show up with the test data you passed me earlier, which isn't too surprising!) I've pushed a new one to my repo: hg pull -r tip http://bitbucket.org/MatthewTurk/yt/ hg up Should get it. What this will do is simply skip over any unrecognized lines, but send an error about them. -Matt On Wed, Jan 25, 2012 at 10:29 AM, Andrew Myers wrote:

Hi Matt,

Thanks. I'm now getting a different (and earlier) error, on the step where I try to load the parameter file.

-Andrew

On Wed, Jan 25, 2012 at 8:19 AM, Matthew Turk wrote:

...

Hi Andrew,

Thanks for the bug report; how embarrassing.

I have pushed a fix to my repository; it only touches Orion.  Casey and I are fixing up some other issues in the Nyx frontend before that gets finished up.  I am inclined to leave Maestro and Castro and attempt to unify Boxlib readers rather than salve this wound.

This week the workshop is taking a bit of time, but early next week I will take a look at the (admittedly confusing) hierarchy and frontend code and see if I can codify this.  The changes to field detection resulted in this.

Could you please test it by doing:

hg pull -r 736067d72401 http://bitbucket.org/MatthewTurk/yt hg up

and then re-running your script?

This also includes changes to the FLASH stuff, and I'm waiting on the integration tests to run before I am willing to push it to the main repository.  Could you let me know if it works for Orion?

-Matt

On Wed, Jan 25, 2012 at 12:01 AM, Andrew Myers wrote:

...

Hello All,

It looks like there is a problem with the current stable version of yt that comes up when creating a PlotCollection using Orion datasets. See the script here and the traceback here.

I think the fix is that there needs to be the line

self.field_info = pf.field_info

somewhere in the constructor for OrionHierarchy (around line 120 of yt/frontends/orion/data_structures.py). It looks like a line like this used to be here, but was removed during some code cleanup. If the other Orion users are anything like me, they haven't updated yt in a while and thus haven't noticed the problem.

Glancing at the Nyx, Maestro, and Castro frontends, those look like they might suffer from the same issue, but I don't have any of their data to confirm or deny this.

-Andrew Myers

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

Ah, great to hear. I will continue with integration tests and push it when it's been verified. (I am a bit skittish about doing this during the workshop, so being extra careful.) On Wed, Jan 25, 2012 at 11:37 AM, Andrew Myers wrote:

Hi Matt,

Thanks. That works beautifully now.

Andrew

On Wed, Jan 25, 2012 at 8:34 AM, Matthew Turk wrote:

...

Hi Andrew,

Okay, this looks like there was a non-standard, uncommented line in the inputs file.  (This didn't show up with the test data you passed me earlier, which isn't too surprising!)  I've pushed a new one to my repo:

hg pull -r tip http://bitbucket.org/MatthewTurk/yt/ hg up

Should get it.  What this will do is simply skip over any unrecognized lines, but send an error about them.

-Matt

On Wed, Jan 25, 2012 at 10:29 AM, Andrew Myers wrote:

...

Hi Matt,

Thanks. I'm now getting a different (and earlier) error, on the step where I try to load the parameter file.

-Andrew

On Wed, Jan 25, 2012 at 8:19 AM, Matthew Turk wrote:

...

Hi Andrew,

Thanks for the bug report; how embarrassing.

I have pushed a fix to my repository; it only touches Orion.  Casey and I are fixing up some other issues in the Nyx frontend before that gets finished up.  I am inclined to leave Maestro and Castro and attempt to unify Boxlib readers rather than salve this wound.

This week the workshop is taking a bit of time, but early next week I will take a look at the (admittedly confusing) hierarchy and frontend code and see if I can codify this.  The changes to field detection resulted in this.

Could you please test it by doing:

hg pull -r 736067d72401 http://bitbucket.org/MatthewTurk/yt hg up

and then re-running your script?

This also includes changes to the FLASH stuff, and I'm waiting on the integration tests to run before I am willing to push it to the main repository.  Could you let me know if it works for Orion?

-Matt

On Wed, Jan 25, 2012 at 12:01 AM, Andrew Myers wrote:

...

Hello All,

It looks like there is a problem with the current stable version of yt that comes up when creating a PlotCollection using Orion datasets. See the script here and the traceback here.

I think the fix is that there needs to be the line

self.field_info = pf.field_info

somewhere in the constructor for OrionHierarchy (around line 120 of yt/frontends/orion/data_structures.py). It looks like a line like this used to be here, but was removed during some code cleanup. If the other Orion users are anything like me, they haven't updated yt in a while and thus haven't noticed the problem.

Glancing at the Nyx, Maestro, and Castro frontends, those look like they might suffer from the same issue, but I don't have any of their data to confirm or deny this.

-Andrew Myers

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org

_______________________________________________ yt-users mailing list yt-users@lists.spacepope.org http://lists.spacepope.org/listinfo.cgi/yt-users-spacepope.org