Which affine format to use in Rasterstats zonal_stats?

Question

I am working with the Rasterstats package to run zonal statistics on two raster datasets. The documentation calls for the following:

import rasterio
from rasterstats import zonal_stats

with rasterio.open('tests/data/slope.tif') as src:
    affine = src.affine
    array = src.read(1)
zs = zonal_stats('tests/data/polygons.shp', array, affine=affine)

However, src.affine is no longer supported in Rasterio. Instead, the documentation recommends using src.transform, which yields something like:

Affine(1.0, 0.0, 670625.0016573452,
       0.0, -1.0, 5282947.795216726)

When I do something like the following, I get an error:

from rasterstats import zonal_stats

dataset = rasterio.open('/path/to/image.tif')
affine = dataset.transform * (0,0) # upper left corner
zs = zonal_stats(segments, ndvi, affine=affine, stats="mean")

---

IndexError: only integers, slices (`:`), ellipsis (`...`), numpy.newaxis (`None`) and integer or boolean arrays are valid indices

---

What affine format does the Rasterstats zonal_stats function require?

Answer 1

The results of dataset.transform and dataset.affine are equals (see Migrating to Rasterio 1.0: affine.Affine() vs. GDAL-style geotransforms)

With rasterio version < 1 (affine = affine transform, transform = GDAL-style geotransforms,

import rasterio 
rasterio.__version__
'0.36.0'
with rasterio.open('slope.tif') as src:
   affine = src.affine
   transform = src.transform
affine
Affine(25.52514657450613, 0.0, 244300.61494985913,
   0.0, -25.52514657450613, 1000868.7876863468)
print(affine)
| 25.53, 0.00, 244300.61|
| 0.00,-25.53, 1000868.79|
| 0.00, 0.00, 1.00|
transform
[244300.61494985913, 25.52514657450613, 0.0, 1000868.7876863468, 0.0, -25.52514657450613]

And the conversion

from affine import Affine
# or
from rasterio import Affine
aff = Affine.from_gdal(*transform)
aff
Affine(25.52514657450613, 0.0, 244300.61494985913,
   0.0, -25.52514657450613, 1000868.7876863468)
gda = aff.to_gdal()
gda
(244300.61494985913, 25.52514657450613, 0.0, 1000868.7876863468, 0.0, -25.52514657450613)

With rasterio version > 1.0b (transform = affine transform)

import rasterio
rasterio.__version__
'1.0.13'
with rasterio.open('slope.tif') as src:
    transform = src.transform
transform
Affine(25.52514657450613, 0.0, 244300.61494985913,
   0.0, -25.52514657450613, 1000868.7876863468)
print(transform)
| 25.53, 0.00, 244300.61|
| 0.00,-25.53, 1000868.79|
| 0.00, 0.00, 1.00|
transform.to_gdal()
(244300.61494985913, 25.52514657450613, 0.0, 1000868.7876863468, 0.0, -25.52514657450613)