find value closest to certain mesh location

[Noxical]

Hello,

I have created a temperature profile over a mesh2D grid. Now I want to get at certain location (point) in the mesh the exact temperature, say X = 4.0 and Y = 2.0:
I know how to do by creating pandas dataframe from the cellValue, see example:

x, y = mesh.cellCenters
df = pd.DataFrame({'x': x, 'y': y, 'T': phi.value})
target_x = 4.8
target_y = 2.2
diff_x = np.abs(df['x'] - target_x)
diff_y = np.abs(df['y'] - target_y)
combined_distance = diff_x + diff_y
closest_row_index = np.argmin(combined_distance)
closest_t_value = df.loc[closest_row_index, 'T']
print(closest_t_value)

works like a charm...
However, how would this work in pure fipy? Is it possible or should I just stick to pandas as shown above.

My prefered option would be using fipy. I was trying something like

T_line = phi.faceValue[mesh.physicalFaces['interface']]

However this gives me all the temperatures over certain line (or area if i define cells) but I was not able to extract specific nearest point data from physicalFaces line data.

Ruud

[guyer]

phi((target_x, target_y)) will give you the value in the cell closest to the point of interest.

phi((target_x, target_y), order=1) will give a first-order interpolation of the value at the point of interest.

See documentation for the __call__() method

[Noxical]

Thanks for your feedback. According the documentation this should work, however I am using a 2Dmesh created with gsmh without uniform grid and I think that is way I get the following error:

ValueError: shape mismatch: value array of shape (5042,) could not be broadcast to indexing result of shape (2,)

So the cell positions are not ordered according a structured grid and thus interpolation is not straighforward?

[guyer]

Sorry, I had the position vector misoriented. Grid2D is more tolerant than Gmsh2D. Try

phi([[target_x], [target_y]]