I drew my custom country map over the default map controlling zorder. But my custom country also overlapped my data.
I can't find zorder option of "set_data" and "set_shape" in EOmaps.

histogram bins that extend beyond a colorbar color-split are not properly colored

There is an unnecessary print() statement in the default picker for GeoDataFrames...

At the moment WebMap services that provide tiles at non-standard dpi are not fetched correctly...

The issue originates most probably from somewhere in here (maybe some kind of a tilePixelRatio needs to be implemented
to take care of the different tile-sizes?)

Here's an example for a high-dpi (512 instead of 96) wms service that does not work as expected...

from eomaps import Maps
m = Maps(Maps.CRS.GOOGLE_MERCATOR)
m.set_extent_to_location("wien")
m.add_wms.Austria.AT_basemap.add_layer.bmaphidpi()

# the corresponding normal dpi services works just fine
# m.add_wms.Austria.AT_basemap.add_layer.geolandbasemap()

Any help with this is highly appreciated!

functools.cache was introduced in python 3.9, so using it breaks support for python<=3.8 !

the documentation always needs improvements... suggestions are welcome!

• add (or improve) examples
• make sure all functionalities are well documented
• create nice images (add a showcase page to show nice maps incl. source code?)

• QT code
  • is it QT6 compatible? (if not... make it QT6 compatible)
  • general improvements
• interface to plot from files
  • avoid opening/reading files multiple times
  • create a better overview text for dataset previews
• user interface
  • make it more usable (and improve design)

Sometimes images are not updated fast enough so that artefacts of animated artists remain after blitting...

• this seems to be a general issue with blitting in nbbagg, webagg and ipympl
  check progress of: matplotlib/matplotlib#19116

Hello.
EOmaps has good functions that can be applied to my map data analysis.
However, in my case, the interactive figure output is too slow to use if it is on the cell of the jupyter notebook (%matplotlib notebook).
If I obtain the interactive figure on the new window (%matplotlib qt), the speed is quite good for use.
Actually, I need a just fixed analysis output around the specific country on the jupyter notebook. So can I obtain this kind of noninterative figure output on the jupyter notebook ??

the histogram plot (implemented here) using a custom colorbar should be a dedicated function that can be easily used externally by passing the correct arguments, since it can be quite useful for other (non-map) plots as well

Instead of warning about missing geopandas installation on import, only warn
when relevant functions are called.

Check what's happening with the S1GBM wms service when using python 3.10...

• allow using "eq_hist" as normalization
• dynamically update colorbar based on displayed data

... a feature request that occasionally pops up...

• the idea would be to have a (draggable) arrow pointing north

at the moment cpos="c" is hardcoded ...
(other options need to be checked to work properly)

• for now, we use extend="neither" in the colorbar to allow sharing the axes-limits between ax_cb and ax_cb_plot
• according to the doc 5% of the colorbar range is used as margin for the extension arrows by default

properly incorporate the extend-fraction in the limits of the histogram axes on top of the colorbar

hi Raphael,

trying the example Data-classification and multiple Maps in one figure w/ a pristine 2.1.1 and replacing the line

mg.m_0_0.set_data_specs(data=data, xcoord="lon", ycoord="lat", in_crs=4326)

w/

mg[0, 0].set_data_specs(data=data, xcoord="lon", ycoord="lat", in_crs=4326)

which i assume should work, instead raises...

Traceback (most recent call last):
  File "/opt/ws-py/eomaps2/test_grid.py", line 15, in <module>
    mg[0, 0].set_data_specs(data=data, xcoord="lon", ycoord="lat", in_crs=4326)
  File "/opt/ws-py/eomaps2/.env/lib/python3.10/site-packages/eomaps/eomaps.py", line 2042, in __getitem__
    return getattr(self, f"ax_{key[0]}_{key[1]}")
AttributeError: 'MapsGrid' object has no attribute 'ax_0_0'. Did you mean: 'm_0_0'?

replacing ax w/ m in line 2042 in eomaps.py seems to solve the problem.

also noticed that ax_... is still mentioned in the MapsGrid class documentation.

The location of TriMesh changed between matplotlib v3.5.x and matplotlib v3.6.x
(see matplotlib/matplotlib#24105)

• implement compatibility for TriMesh import for older matplotlib versions

m.plot_map(vmin=..., vmax=...) does not yet convert values to actual data-values if encoded data (e.g. set with m.set_data(... encoding={...})) is plotted!

TODO:

• add conversion for values provided as vmin/vmax
  (vmin_data = encoding["add_offset"] + encoding["scale_factor"] * vmin)

Hi Raphael, I just suddenly came up with an idea. What if you or somebody (perhaps me), try to combine EOmaps with Streamlit?

I guess one benefit will be that more users with little or no Python knowledge can be able to use your EOmaps tools, at least locally.

I understand that a Streamlit APP can be quite big in size. But if just running them locally, then it may not be a big problem? Once a Streamlit App is built by an experienced Python user in a group, he/she can pass it to other users, who will just use the Streamlit APP to carry on some routine analysis?

I am just thinking, not sure about feasibility, challenges, whether it is worth the effort and time.

hi there,

1. trying to run the following code using the module:

from eomaps import Maps
import matplotlib.pyplot as plt
import pandas as pd

df = pd.read_csv('...csv')
m = Maps()
m.set_data(data=df, xcoord='...', ycoord='...', parameter='...', crs=4326)
m.set_shape.geod_circles(10000)
m.plot_map()

i was getting an exception thrown from the 'fiona' module:

Traceback (most recent call last):
  File "/opt/ws-py/eomaps/test.py", line 1, in <module>
    from eomaps import Maps, MapsGrid
  File "/opt/ws-py/eomaps/.env/lib/python3.10/site-packages/eomaps/__init__.py", line 11, in <module>
    from .eomaps import Maps, MapsGrid
  File "/opt/ws-py/eomaps/.env/lib/python3.10/site-packages/eomaps/eomaps.py", line 26, in <module>
    from cartopy import crs as ccrs
  File "/opt/ws-py/eomaps/.env/lib64/python3.10/site-packages/cartopy/__init__.py", line 108, in <module>
    import cartopy.feature  # noqa: F401  (flake8 = unused import)
  File "/opt/ws-py/eomaps/.env/lib64/python3.10/site-packages/cartopy/feature/__init__.py", line 18, in <module>
    import cartopy.io.shapereader as shapereader
  File "/opt/ws-py/eomaps/.env/lib64/python3.10/site-packages/cartopy/io/shapereader.py", line 43, in <module>
    import fiona
  File "/opt/ws-py/eomaps/.env/lib64/python3.10/site-packages/fiona/__init__.py", line 87, in <module>
    with fiona._loading.add_gdal_dll_directories():
AttributeError: partially initialized module 'fiona' has no attribute '_loading' (most likely due to a circular import). Did you mean: 'logging'?

replacing the lines:

import fiona._loading
with fiona._loading.add_gdal_dll_directories():

in .../.env/lib64/python3.10/site-packages/fiona/__init__.py --.env being the directory of a virtual env created inside the project where the code resides-- w/

import _loading
with _loading.add_gdal_dll_directories():

solved this problem.

2. running the code (a) terminates w/ no exception or stack trace, but (b) no map/plot window is shown.

adding

plt.show()

at the end causes the map's figure window to open which seems to solve the problem but i'm unsure of what consequences that work-around may cause later.

my environment is the following:

• Fedora 35, running GNU/Linux 5.15.6-200.fc35.x86_64
• Python 3.10.0

are these known problems, issues related to my environment, or something(s) i missed in installing and using the module?

TIA + cheers;

• improve startup time by loading modules only when they are required
  (this is already done but it can be improved)

• split package into eomaps_base and eomaps

  would be nice to be able to install a "minimal version" and a "full version" of eomaps (similar to matplotlib_base and matplotlib) ... I don't really know how to do this properly... any help is highly appreciated!)

  • eomaps_base: only the most basic needs (e.g. matplotlib, cartopy, numpy etc.)
  • eomaps: eomaps_base and all optional dependencies ( e.g. pyqt, datashader, xarray, rasterio etc.)

• add more shapes for drawing
• add support to draw a shape with text inside
• make polygons editable
  • pick polygons to select them
  • drag vertices to edit drawn polygons
  • move/rotate whole polygon
  • change color, linewidth etc.

• inset maps should allow setting the layer on initialization!
• inset maps should only be visible on the assigned layer

While running a script with >90 million datapoints following error occured, even tough i installed the datashader module explicitly during the installation of eomaps:

`AssertionError: EOmaps: Missing dependency: 'datashader' 
please install (conda install -c conda-forge datashader) to use 'shade_raster'

Output from spyder call 'get_namespace_view':
EOmaps-Warning: you attempt to plot a large dataset(90280000 datapoints) but the 'datashader' library could not be imported!
The plot might take long to finish!... defaulting to 'ellipses' as plot-shape.`

A Check with conda list showed that it truly is installed:
datashader 0.14.1 pyh6c4a22f_1 conda-forge

Trying an import in the beginning of my script (import datashader as ds) escalates in:

Traceback (most recent call last):

  File "C:\Users\KFalkner\AppData\Local\Temp\ipykernel_6124\3866946782.py", line 1, in <cell line: 1>
    import datashader as ds

  File "C:\Users\KFalkner\Anaconda3\envs\eomaps\lib\site-packages\datashader\__init__.py", line 8, in <module>
    from .core import Canvas                                 # noqa (API import)

  File "C:\Users\KFalkner\Anaconda3\envs\eomaps\lib\site-packages\datashader\core.py", line 13, in <module>
    from .utils import Dispatcher, ngjit, calc_res, calc_bbox, orient_array, \

  File "C:\Users\KFalkner\Anaconda3\envs\eomaps\lib\site-packages\datashader\utils.py", line 8, in <module>
    import numba as nb

  File "C:\Users\KFalkner\Anaconda3\envs\eomaps\lib\site-packages\numba\__init__.py", line 200, in <module>
    _ensure_critical_deps()

  File "C:\Users\KFalkner\Anaconda3\envs\eomaps\lib\site-packages\numba\__init__.py", line 140, in _ensure_critical_deps
    raise ImportError("Numba needs NumPy 1.21 or less")

ImportError: Numba needs NumPy 1.21 or less

This are the versions installed:

numba                     0.55.0          py310h4ed8f06_0
numpy                     1.23.1          py310h8a5b91a_0    conda-forge

I have very little knowledge about modules/packages, but somehow it seems that when installing eomaps a newer (or the newest?) version of numpy is used, which is not compatible with numba yet.

Solution:

I downgraded numpy and everything works fine now:
pip install numpy==1.21.6

I had the same issue on 2 Computers so i guess this could be a common problem why i wanted to let you know.

Receive a warning when importing Maps from eomaps: Cannot import name 'TriMesh' from matplotlib.tri.

Hello again.
Thank you very much for your work developing EOmaps. I wish for the prosperity of EOmaps.

I have a question for the options of "m.set_shape.". Currently, "m.set_shape." has no option for geodesic rectangular. Is geodesic rectangular possible for a option of "m.set_shape." ?

In my field, unit of collecting data is a 1km or 2km rectangular region of the country. So I have data of which columns are (lon, lat) of the rectangular center and measurement value on the rectangular. Currently, I can only rely on "m.set_shape.geod_circles(radius)". This gives me a enough good visualization, but there is a regret that I want it to be a rectangle.

something like this:

• if the box is checked, switch to the newly created layer
• if the box is unchecked stay on the currently visible layer

The position of the extension-arrows is not properly updated when using m.subplots_adjust(...) after the colorbar has been added

m = Maps()
m.set_data([1,2,3], [1,2,3], [1,2,3])
m.plot_map(vmin=1.2, vmax=2.5)
m.add_colorbar()
m.subplots_adjust(left=.3)

crs-specific webmap services in the companion-widget are not properly cached!

• make sure all fetched layers are written to disc (not only the ones relevant for the current crs)

• at the moment m.read_data, m.from_file and m.new_layer_from_file convert values to float...
  • this is problematic for extremely large datasets stored as integers!
• handle encoding of GeoTIFFS and NetCDF's accordingly
• allow providing custom encodings

This line gives an error after installation

Somehow bullet-lists don't render properly at the moment when using docutils > 0.16
(docs use test_env.yml environment)

Any help to make the docs render properly with the latest docutils version is highly appreciated!

relevant sources:

• readthedocs/sphinx_rtd_theme#1300
• question on stackoverflow

Hello.

I have reconstructed my anaconda env.
Before reconstructing the env, I used code "m.ax.add_geometries~" with eomaps, which is pure cartopy code except "m.", to draw modified map. But now that code does not work. is this only my problem?

code example:

import cartopy.crs as ccrs
import cartopy.feature as cf
from eomaps import Maps, MapsGrid
import cartopy.io.shapereader as shpreader

m = Maps(crs=5179)
m.ax.set_extent([124.598822,131.87222222,38.70000000,33.10000000], crs=ccrs.PlateCarree()) ## Important
m.ax.axis('off')
m.ax.add_feature(cf.OCEAN.with_scale("50m")) 
shpfilename = shpreader.natural_earth(resolution='50m', category='cultural', name='admin_0_countries')
reader = shpreader.Reader(shpfilename)
countries = reader.records()
for country in countries:
    if country.attributes['ADM0_A3'] == 'KOR':
        m.ax.add_geometries(country.geometry, ccrs.PlateCarree(), facecolor=cf.COLORS['water'], edgecolor=cf.COLORS['water'])
m.ax.add_feature(cf.BORDERS, linewidth=0.7, linestyle=':', zorder=5) # country borders in cartopy.

All keys for keypress-callbacks are expected to be entered as strings!
m.cb.keypress.attach(key=0) does not throw a warning... but only m.cb.keypress.attach(key="0") works!

Somehow cartopy's reprojection does not work here

from eomaps import Maps
m = Maps(crs = Maps.CRS.Robinson())
gdf = m.add_feature.preset.ocean()

...but geopandas-reprojection does!:

from eomaps import Maps
m = Maps(crs = Maps.CRS.Robinson())
gdf = m.add_feature.preset.ocean(reproject="gpd")

or

from eomaps import Maps
m = Maps(crs = Maps.CRS.Robinson())

gdf = m.add_feature.preset.ocean.feature.get_gdf()
m.add_gdf(gdf, fc='lightsteelblue')

At the moment, legends added by m.add_gdf are not properly handled.

• colorbars are not layer-sensitive

from eomaps import Maps
m = Maps()
gdf = m.add_feature.cultural.admin_0_countries.get_gdf()
m.add_gdf(gdf, legend=True, column="scalerank")

• legends do not show up at all

from eomaps import Maps
m = Maps()
gdf = m.add_feature.cultural.admin_0_countries.get_gdf()
m.add_gdf(gdf, legend=True, column="scalerank", scheme="Quantiles")

Temporary fix:

# find the added legend object and add it as an artist
leg = next(i for i in m.ax.get_children() if i.__class__.__name__=="Legend")
m.BM.add_artist(leg, layer=m.layer)

Just a question. As EOmaps is based on matplotlib for plotting, is it possible to use it for embedding maps into wxpython based app? Matplotlib plot can be embedded so I imagine EOmaps could be embedded too.

Cheers.

currently the layer-selector shows only layers with features...
but if it just inherits features from the "all" layer it should be shown (e.g. callback-only layers)

1. when clicking on "save" in the widget, the window to select the path appears for a short moment and then disappears
2. widget and open maps/figures are in suspense until console is closed (spyder editor, linux )

currently, if a dataset with "nan" in the coordinates is attempted to be plotted, the kernel is interrupted!
("nan" in the data with valid coordinates is fine)

TODO

• check if all coordinates are valid
• mask for invalid coordinates prior to plotting (or at least issue a warning)

check

• check what part of the code can't deal with nan's as coordinates and how to deal with this efficiently
• avoid that this becomes a performance-issue for very large datasets!

• unify syntax for
  • m.new_layer_from_file.GeoTIFF
  • m.new_layer_from_file.NetCDF
  • m.new_layer_from_file.CSV
  • ... and similarly for Maps.from_file....
• [maybe] detect file-type by file-ending and avoid explicit functions for each filetype?
  • is this feasible/useful or does it cause problems/confusion ?
• add more filetypes (from_hdf, from_zarr, etc.?)

Add support to draw geometries obtained from WebFeatureServices

• use cartopy.feature.WFSFeature

I don't have any experience with WFS yet but it would be nice to have...

• collect a list of open-access WFS services
• add a m.add_wfs function to draw WFS service layers
• add dot-access functionality of pre-defined services (similar to m.add_wms)
• [maybe] add direct support to make geometries pickable?

Irregularly sampled data causes issues with the raster shape if a zoomed-in region is selected.

m = Maps()
sx, sy = 100, 100
lon = np.sort(np.random.choice(np.linspace(-150,150, sx*10), size=sx, replace=False))
lat = np.sort(np.random.choice(np.linspace(-75, 68, sy*10), size=sy, replace=False))
data = np.random.randint(0, 100, (lon.size, lat.size))
m.set_data(data, lon, lat)
m.set_shape.raster()
m.plot_map()

Temporary workaround:

Always use the full dataset instead of pre-selecting the data based on the visible extent

m = Maps()
# set the margin-factors BEFORE setting the data to disable 
# pre-selection of data with respect to the visible extent
m._data_manager.set_margin_factors(np.inf, np.inf)
...

When I import EOmaps, I obtain the following import error

ImportError Traceback (most recent call last)
Input In [9], in <cell line: 1>()
----> 1 import eomaps

File ~\anaconda3\envs\specavail1\lib\site-packages\eomaps_init_.py:1, in
----> 1 from .eomaps import Maps, MapsGrid
2 from ._version import version
4 author = "Raphael Quast"

File ~\anaconda3\envs\specavail1\lib\site-packages\eomaps\eomaps.py:112, in
102 from cartopy import crs as ccrs
104 from .helpers import (
105 pairwise,
106 cmap_alpha,
(...)
110 searchtree,
111 )
--> 112 from ._shapes import shapes
114 from ._containers import (
115 data_specs,
116 map_objects,
(...)
120 NaturalEarth_features,
121 )
123 from ._cb_container import cb_container

File ~\anaconda3\envs\specavail1\lib\site-packages\eomaps_shapes.py:2, in
1 from matplotlib.collections import PolyCollection, QuadMesh
----> 2 from matplotlib.tri import TriMesh, Triangulation
3 import numpy as np
5 from pyproj import CRS, Transformer

ImportError: cannot import name 'TriMesh' from 'matplotlib.tri' (C:\Users\user\anaconda3\envs\specavail1\lib\site-packages\matplotlib\tri_init_.py)

At the moment the "frameon" kwarg (forwarded to plt.figure() does not work as expected since spines/frames are
handled differently in EOmaps v6.x )

from eomaps import Maps
m = Maps(frameon=False)
m.add_feature.preset.coastline()

Expected behaviour: no frames should be drawn for the axes of this Maps object with frameon=False

As temporary quick-fix one can use:

m = Maps()
m.ax.spines["geo"].set_edgecolor("none")

After update, savefig gives me black paper.

my code:

m = Maps(crs=5179, frameon=False)
~
m.snapshot(clear=True)
m.savefig('name.png', dpi=300, bbox_inches='tight', pad_inches=0.1, facecolor='auto', edgecolor='auto')
plt.close()   

snapshot works well but savefig gives me black paper

At the moment the companion-widget is untested!

• test functionalities
  • add/remove drawn shapes and annotations
  • add/remove artists and features
  • add/remove layers
  • open/plot files
• add tests to make sure buttons etc. do what they are supposed to do