maping TIFF-format raster

[lwq-star]

I need to create maps for 6 raster datasets. How can I read TIFF-format raster data and then use the fig.colorbar from the ultraplot library to draw a colorbar on the leftmost side? I tried using ax.imshow, contourf, and pcolor but none worked.

Similar to replacing the basemap with my own raster data

[cvanelteren]

Like this:

import ultraplot as uplt, numpy as np

a = np.random.rand(10, 10) * 2 - 1
fig, ax = uplt.subplots(ncols=2, proj="cyl")
for axi in ax:
    h = axi.pcolormesh(a)


fig.colorbar(h, loc="l")
ax.format(
    land=True,
    lonlim=(-10, 10),
    latlim=(-10, 10),
)
uplt.show(block=True)

[cvanelteren]

Not sure I understand the question tbh. The colorbar can be plotted using either the functions themselves (e.g. pcolormesh or pcolor with the colorbar keyword) or with a mappable as in the example.

[lwq-star]

This is my code

import ultraplot as uplt
import rasterio
import numpy as np
import geopandas as gpd
from ultraplot import rc

# fig = uplt.figure(share=0)
fig, axs = uplt.subplots(ncols=2, nrows=3, span=False, proj="cyl", journal='nat2')
axs.format(abc=True, abcloc="ul")

gdf = gpd.read_file(r"C:\Users\A1827\Desktop\area.shp")

variables = ['ndvi', 'kndvi', 'evi', 'nirv', 'gpp', 'lai']
for idx, name in enumerate(variables):
    ax = axs[idx]
    
    file = f"C:/Users/A1827/Desktop/{name}_lambda_xt.tif"
    with rasterio.open(file) as src:
        data = src.read(1)
        data[data == src.nodata] = np.nan
        # m = show(src, ax=ax, cmap="viridis")
        
    m = ax.pcolormesh(data, cmap="viridis")
    gdf.plot(ax=ax, edgecolor='gray', facecolor='none', linewidth=1)
    
    ax.format(lonlim=(73, 135.5), latlim=(18, 54))
    
fig.colorbar(m, label="ac1", ticks=4, loc="r")

But the grid is only partially displayed

[cvanelteren]

Could it be due to your nan setter, i.e. this line data[data == src.nodata] = np.nan?
A minimal verifiable example would help for me to assist you.

[lwq-star]

Could it be due to your nan setter, i.e. this line data[data == src.nodata] = np.nan?这可能是由于您的 nan setter，即这一行 data[data == src.nodata] = np.nan ？ A minimal verifiable example would help for me to assist you.一个最小的可验证示例将有助于我帮助您。

I don't think that's the issue. When I use rasterio.plot for plotting, everything works fine, but I can't use fig.colorbar.

from rasterio.plot import show
m = show(src, ax=ax, cmap="viridis")

[cvanelteren]

Dunno what's in the data but our docs state:

*args (z or x, y, z) – The data passed as positional or keyword arguments. Interpreted as follows: If only z coordinates are passed, try to infer the x and y coordinates from the DataFrame indices and columns or the DataArray coordinates. Otherwise, the y coordinates are np.arange(0, y.shape[0]) and the x coordinates are np.arange(0, y.shape[1]).

[lwq-star]

this is my data, coordinate system is wgs84.

this is code

import ultraplot as uplt
import rasterio
from rasterio.plot import show
import numpy as np

# fig = uplt.figure(share=0)
fig, ax = uplt.subplots(proj="cyl", journal='nat2')
  
file = r"C:/Users/A1827/Desktop/ndvi_lambda_xt.tif"
with rasterio.open(file) as src:
    data = src.read(1)
    # data[data == src.nodata] = np.nan
    # show(src, ax=ax, cmap="viridis")
    
m = ax.pcolor(data, cmap="viridis")

# no limit
# ax.format(lonlim=(73, 135.5), latlim=(18, 54))
    
fig.colorbar(m, label="ac1", ticks=4, loc="r")
ax.format(land=True)

I suspect there is an issue with the coordinate system.

[lwq-star]

When using the rasterio library to read a TIFF file and convert it into an array for plotting, the positional alignment has shifted.

[cvanelteren]

We support xarray support not directly TIFF images. If you can convert the TIFF to an xarray it would work fine I suspect.

[cvanelteren]

If you want direct TIFF support then PR is most definitely welcome but I don't have time to work on this for some time.

[cvanelteren]

For example this works:

import ultraplot as uplt, numpy as np, pandas as pd, xarray as xr
import cartopy.crs as ccrs

# Mock data over China
lon_min, lon_max = 73, 135
lat_min, lat_max = 18, 54

# Grid resolution
nx, ny = 100, 80
# Define Grid
lons = np.linspace(lon_min, lon_max, nx)
lats = np.linspace(lat_min, lat_max, ny)
lon2d, lat2d = np.meshgrid(lons, lats)

data = np.exp(-((lon2d - 105) ** 2 + (lat2d - 35) ** 2) / (2 * 10**2))
# Create xarray DataArray
da = xr.DataArray(
    data,
    coords={"lat": lats, "lon": lons},
    dims=["lat", "lon"],
    name="mock_data",
    attrs={"units": "arbitrary", "description": "Mock data over China"},
)

fig, ax = uplt.subplots(proj="cyl")
ax.pcolormesh(da)
ax.format(borders=True, coast=True, oceancolor="blue")
uplt.show()

[cvanelteren]

Also raterio would fill this gap and we are compatible I suspect with it -- I reckon you can use the mappable directly from them.

[lwq-star]

Using the rioxarray library to read a raster file (TIFF format), everything works correctly.

import rioxarray as rxr
# fig = uplt.figure(share=0)
fig, ax = uplt.subplots(proj="cyl", journal='nat2')
  
file = r"C:/Users/A1827/Desktop/ndvi_lambda_xt.tif"
with rxr.open_rasterio(file, masked=True) as data:
    data = data.sel(band=1)
    
m = ax.pcolormesh(data, cmap="viridis")

# no limit
ax.format(lonlim=(73, 135.5), latlim=(18, 54))
    
fig.colorbar(m, label="ac1", ticks=4, loc="r")