Hi Tazkera, I think yt should be able to understand your code units and do the unit conversion for you, or easily configured to interpret your code units. See an example here http://yt-project.org/docs/dev/examining/loading_data.html#tipsy-data. This I feel is the better solution. But if you want to the conversion by hand, a quick fix to your code is to define new variables for domain width and domain center: ``` new_domain_width = ds.domain_width*((3.086e18*6e6*0.909090893796)/(0.7)) new_domain_center = ds.domain_center*((3.086e18*6e6*0.909090893796)/(0.7)) yt.ProjectionPlot(ds, 'z', ('gas', 'ElectronPressure'), width= new_domain_width[0], center=new_domain_center) ``` The issue with your code is that ds.domain_width and ds.domain_center are not fields themselves. If you try to add new fields for those constant values, you'll get a long list of those values as you could read from the error message. You could print out dd['gas', 'PhysicalDomainWidth'] or dd['gas', 'PhysicalDomainCenter'] in your code to see what they look like. That might help you understand what was wrong. Best, Bili On Thu, Oct 5, 2017 at 2:59 PM, tazkera haque wrote:

Dear yt people,

I am working on a tipsy output file from a n-body simulation. My file contains some scaled units for fields like density, coordinates, mass etc. To make a physical sense of the data, I had to convert those fields units to some physical units (c.g.s) and I derived new fields having cgs unit. for example:

*dd = ds.all_data()# derive a new domain width with conversion factor (6Mpc*current_time/hubble_constant) def _NewDomainWidth(field, data): return ds.domain_width*((3.086e18*6e6*0.909090893796)/(0.7))ds.add_field(('gas', 'PhysicalDomainWidth'), function=_NewDomainWidth, units="code_length")*

*# derive a new domain center with conversion factor (6Mpc*current_time/hubble_constant) def _NewDomainCenter(field, data): return ds.domain_center*((3.086e18*6e6*0.909090893796)/(0.7))ds.add_field(('gas', 'PhysicalDomainCenter'), function=_NewDomainCenter, units="code_length")*

note that here "code_length" is actually in "cm". but I could not just do dd["Gas",'Coordinates'].in_units('cm'), because this conversion does not take into account some constant corrections and returns the wrong length magnitude.

My goal was to make sliceplot and projectionplot of a newly derived field :

*# derive a new electron pressure fielddef _ElectronPressure(field, data): return data['Gas','Density']*(5.49e-7)*data['Gas','Temperature']ds.add_field(('gas', 'ElectronPressure'), function=_ElectronPressure, units="(code_mass*K)/(code_length**3)")*

*# integration of pressure field along line of sight*

*yt.ProjectionPlot(ds, 'z', ('gas', 'ElectronPressure'), width = dd['gas', 'PhysicalDomainWidth'], center = dd['gas', 'PhysicalDomainCenter'] )*

*and it is returning me *

ValueError Traceback (most recent call last)<ipython-input-8-fa351b6c09d1> in <module>() 1 yt.ProjectionPlot(ds, 'z', ('gas', 'ElectronPressure'), width = dd['gas', 'PhysicalDomainWidth'], ----> 2 center = dd['gas', 'PhysicalDomainCenter'] )

ValueError: operands could not be broadcast together with shapes (39678,) (54608,)

Now I cannot use width =ds.domain_width and center = ds.domain_center for the same unit error. I was wondering how I can make the projectionplot for my newly derived field and newly derived domain width and domain center.

Your suggestion is most welcome

Best

Tazkera

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

On Thu, Oct 5, 2017 at 3:48 PM, Bili Dong - Gmail wrote:

Hi Tazkera,

I think yt should be able to understand your code units and do the unit conversion for you, or easily configured to interpret your code units. See an example here http://yt-project.org/docs/dev/examining/loading_data.html#tipsy-data. This I feel is the better solution.

But if you want to the conversion by hand, a quick fix to your code is to define new variables for domain width and domain center:

``` new_domain_width = ds.domain_width*((3.086e18*6e6*0.909090893796)/(0.7)) new_domain_center = ds.domain_center*((3.086e18*6e6*0.909090893796)/(0.7)) yt.ProjectionPlot(ds, 'z', ('gas', 'ElectronPressure'), width= new_domain_width[0], center=new_domain_center) ```

The issue with your code is that ds.domain_width and ds.domain_center are not fields themselves. If you try to add new fields for those constant values, you'll get a long list of those values as you could read from the error message. You could print out dd['gas', 'PhysicalDomainWidth'] or dd['gas', 'PhysicalDomainCenter'] in your code to see what they look like. That might help you understand what was wrong.

Sorry I was wrong. This is not true.

Best, Bili

On Thu, Oct 5, 2017 at 2:59 PM, tazkera haque wrote:

...

Dear yt people,

I am working on a tipsy output file from a n-body simulation. My file contains some scaled units for fields like density, coordinates, mass etc. To make a physical sense of the data, I had to convert those fields units to some physical units (c.g.s) and I derived new fields having cgs unit. for example:

*dd = ds.all_data()# derive a new domain width with conversion factor (6Mpc*current_time/hubble_constant) def _NewDomainWidth(field, data): return ds.domain_width*((3.086e18*6e6*0.909090893796)/(0.7))ds.add_field(('gas', 'PhysicalDomainWidth'), function=_NewDomainWidth, units="code_length")*

*# derive a new domain center with conversion factor (6Mpc*current_time/hubble_constant) def _NewDomainCenter(field, data): return ds.domain_center*((3.086e18*6e6*0.909090893796)/(0.7))ds.add_field(('gas', 'PhysicalDomainCenter'), function=_NewDomainCenter, units="code_length")*

note that here "code_length" is actually in "cm". but I could not just do dd["Gas",'Coordinates'].in_units('cm'), because this conversion does not take into account some constant corrections and returns the wrong length magnitude.

My goal was to make sliceplot and projectionplot of a newly derived field :

*# derive a new electron pressure fielddef _ElectronPressure(field, data): return data['Gas','Density']*(5.49e-7)*data['Gas','Temperature']ds.add_field(('gas', 'ElectronPressure'), function=_ElectronPressure, units="(code_mass*K)/(code_length**3)")*

*# integration of pressure field along line of sight*

*yt.ProjectionPlot(ds, 'z', ('gas', 'ElectronPressure'), width = dd['gas', 'PhysicalDomainWidth'], center = dd['gas', 'PhysicalDomainCenter'] )*

*and it is returning me *

ValueError Traceback (most recent call last)<ipython-input-8-fa351b6c09d1> in <module>() 1 yt.ProjectionPlot(ds, 'z', ('gas', 'ElectronPressure'), width = dd['gas', 'PhysicalDomainWidth'], ----> 2 center = dd['gas', 'PhysicalDomainCenter'] )

ValueError: operands could not be broadcast together with shapes (39678,) (54608,)

Now I cannot use width =ds.domain_width and center = ds.domain_center for the same unit error. I was wondering how I can make the projectionplot for my newly derived field and newly derived domain width and domain center.

Your suggestion is most welcome

Best

Tazkera

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

HI Nathan and Bili, Thanks very much for the help. This makes perfect sense now. I was not aware of the difference of 'gas' and 'Gas' fields. Also the unit override function is very helpful in this case. Best Tazkera On Thu, Oct 5, 2017 at 7:40 PM, Bili Dong - Gmail wrote:

Hi Tazkera,

Now I understand where the error comes from. It's not from the unit system but from the difference between 'Gas' and 'gas' fields. Let me show you the code that works and explain later. The corrected code looks like this:

``` dd = ds.all_data()

# derive a new domain width with conversion factor (6Mpc*current_time/hubble_constant) def _NewDomainWidth(field, data): return ds.domain_width*((3.086e18*6e6*0.909090893796)/(0.7))

ds.add_field(('gas', 'PhysicalDomainWidth'), function=_NewDomainWidth, units="code_length")

# derive a new domain center with conversion factor (6Mpc*current_time/hubble_constant) def _NewDomainCenter(field, data): return ds.domain_center*((3.086e18*6e6*0.909090893796)/(0.7))

ds.add_field(('gas', 'PhysicalDomainCenter'), function=_NewDomainCenter, units="code_length")

# derive a new electron pressure field def _ElectronPressure(field, data): return data['Gas','Density']*(5.49e-7)*data['Gas','Temperature']

ds.add_field(('Gas', 'ElectronPressure'), function=_ElectronPressure, units="(code_mass*K)/(code_length**3)", particle_type=True) f_smooth = ds.add_smoothed_particle_field(('Gas', 'ElectronPressure')) # f_smooth equals ('deposit', 'Gas_smoothed_ElectronPressure')

# integration of pressure field along line of sight yt.ProjectionPlot(ds, 'z', f_smooth, width = dd['gas', 'PhysicalDomainWidth'], center = dd['gas', 'PhysicalDomainCenter'] ) ```

First, let's go back to the error message. Where do the magic numbers (39678,) (54608,) come from? Print out dd['Gas', 'Density'].shape and dd['gas', 'density'].shape and you'll see. The issue with your original code is that you tried to create a 'gas' field from a 'Gas' field but their dimension doesn't match. Here 'Gas' field is a particle field retrieved from your snapshot directly, but 'gas' field is an SPH field calculated from the particle field. They have different dimensions. More detailed documentation could be found here:

http://yt-project.org/docs/dev/analyzing/fields.html#sph-fields

That's for solving the error. As a side note, I do not recommend defining fields like your dd['gas', 'PhysicalDomainCenter'], though it works (to my surprise). Because usually we assume the same field type means the same dimension (as in the first number of shape). Your case is more suitable to define a new variable.

Hope it helps.

Best, Bili

On Thu, Oct 5, 2017 at 3:54 PM, Bili Dong - Gmail wrote:

...

On Thu, Oct 5, 2017 at 3:48 PM, Bili Dong - Gmail wrote:

...

Hi Tazkera,

I think yt should be able to understand your code units and do the unit conversion for you, or easily configured to interpret your code units. See an example here http://yt-project.org/docs/dev/examining/loading_data.html#tipsy-data. This I feel is the better solution.

But if you want to the conversion by hand, a quick fix to your code is to define new variables for domain width and domain center:

``` new_domain_width = ds.domain_width*((3.086e18*6e6*0.909090893796)/(0.7)) new_domain_center = ds.domain_center*((3.086e18* 6e6*0.909090893796)/(0.7)) yt.ProjectionPlot(ds, 'z', ('gas', 'ElectronPressure'), width= new_domain_width[0], center=new_domain_center) ```

The issue with your code is that ds.domain_width and ds.domain_center are not fields themselves. If you try to add new fields for those constant values, you'll get a long list of those values as you could read from the error message. You could print out dd['gas', 'PhysicalDomainWidth'] or dd['gas', 'PhysicalDomainCenter'] in your code to see what they look like. That might help you understand what was wrong.

Sorry I was wrong. This is not true.

...

Best, Bili

On Thu, Oct 5, 2017 at 2:59 PM, tazkera haque wrote:

...

Dear yt people,

I am working on a tipsy output file from a n-body simulation. My file contains some scaled units for fields like density, coordinates, mass etc. To make a physical sense of the data, I had to convert those fields units to some physical units (c.g.s) and I derived new fields having cgs unit. for example:

*dd = ds.all_data()# derive a new domain width with conversion factor (6Mpc*current_time/hubble_constant) def _NewDomainWidth(field, data): return ds.domain_width*((3.086e18*6e6*0.909090893796)/(0.7))ds.add_field(('gas', 'PhysicalDomainWidth'), function=_NewDomainWidth, units="code_length")*

*# derive a new domain center with conversion factor (6Mpc*current_time/hubble_constant) def _NewDomainCenter(field, data): return ds.domain_center*((3.086e18*6e6*0.909090893796)/(0.7))ds.add_field(('gas', 'PhysicalDomainCenter'), function=_NewDomainCenter, units="code_length")*

note that here "code_length" is actually in "cm". but I could not just do dd["Gas",'Coordinates'].in_units('cm'), because this conversion does not take into account some constant corrections and returns the wrong length magnitude.

My goal was to make sliceplot and projectionplot of a newly derived field :

*# derive a new electron pressure fielddef _ElectronPressure(field, data): return data['Gas','Density']*(5.49e-7)*data['Gas','Temperature']ds.add_field(('gas', 'ElectronPressure'), function=_ElectronPressure, units="(code_mass*K)/(code_length**3)")*

*# integration of pressure field along line of sight*

*yt.ProjectionPlot(ds, 'z', ('gas', 'ElectronPressure'), width = dd['gas', 'PhysicalDomainWidth'], center = dd['gas', 'PhysicalDomainCenter'] )*

*and it is returning me *

ValueError Traceback (most recent call last)<ipython-input-8-fa351b6c09d1> in <module>() 1 yt.ProjectionPlot(ds, 'z', ('gas', 'ElectronPressure'), width = dd['gas', 'PhysicalDomainWidth'], ----> 2 center = dd['gas', 'PhysicalDomainCenter'] )

ValueError: operands could not be broadcast together with shapes (39678,) (54608,)

Now I cannot use width =ds.domain_width and center = ds.domain_center for the same unit error. I was wondering how I can make the projectionplot for my newly derived field and newly derived domain width and domain center.

Your suggestion is most welcome

Best

Tazkera

_______________________________________________ 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