Add function to load raster tile maps using contextily #2125

weiji14 · 2022-09-19T16:25:37Z

Description of proposed changes

New dataset function to load XYZ tiles! Uses contextily to retrieve the tiles based on a bounding box.

Preview at https://pygmt-dev--2125.org.readthedocs.build/en/2125/api/generated/pygmt.datasets.load_tile_map.html

This is the first part of #2115 that does the tile loading and georeferencing to an xarray.DataArray. A follow up PR will be needed to take the output xarray.DataArray from pygmt.datasets.load_tile_map and plot it using fig.grdimage (to be discussed in #2115).

Usage:

import contextily
from pygmt.datasets import load_tile_map

raster = load_tile_map(
    region=[103.60, 104.06, 1.22, 1.49],  # West, East, South, North
    source=contextily.providers.Stamen.TerrainBackground,
    lonlat=True,  # bounding box coordinates are longitude/latitude
)

print(raster)
# <xarray.DataArray (band: 3, y: 1024, x: 1536)>
# array([[[208, 208, 205, ..., 210, 213, 213],
#         [208, 208, 205, ..., 213, 213, 213],
#         [205, 205, 204, ..., 213, 213, 213],
#         ...,
#         [153, 153, 153, ..., 153, 153, 153],
#         [153, 153, 153, ..., 153, 153, 153],
#         [153, 153, 153, ..., 153, 153, 153]],
# 
#        [[214, 214, 213, ..., 217, 218, 218],
#         [214, 214, 213, ..., 219, 220, 219],
#         [213, 213, 212, ..., 219, 220, 219],
#         ...,
#         [179, 179, 179, ..., 179, 179, 179],
#         [179, 179, 179, ..., 179, 179, 179],
#         [179, 179, 179, ..., 179, 179, 179]],
# 
#        [[175, 175, 174, ..., 186, 185, 185],
#         [175, 175, 174, ..., 188, 188, 188],
#         [174, 174, 171, ..., 188, 188, 188],
#         ...,
#         [204, 204, 204, ..., 204, 204, 204],
#         [204, 204, 204, ..., 204, 204, 204],
#         [204, 204, 204, ..., 204, 204, 204]]], dtype=uint8)
# Coordinates:
#   * band     (band) int64 0 1 2
#   * y        (y) float64 1.49 1.49 1.489 1.489 1.489 ... 1.221 1.221 1.22 1.22
#   * x        (x) float64 103.6 103.6 103.6 103.6 ... 104.1 104.1 104.1 104.1

Alternatively, they can use a custom Web Map Tile url like so:

raster = load_tile_map(
    region=[103.60, 104.06, 1.22, 1.49],  # West, East, South, North
    source="https://{s}.tile.openstreetmap.org/{z}/{x}/{y}.png",
    lonlat=True,  # bounding box coordinates are longitude/latitude
)
print(raster)
# <xarray.DataArray (band: 3, y: 1024, x: 1536)>
# array([[[174, 174, 174, ..., 200, 200, 200],
#         [181, 188, 181, ..., 200, 200, 200],
#         [201, 196, 197, ..., 200, 200, 200],
#         ...,
#         [170, 170, 170, ..., 170, 170, 170],
#         [170, 170, 170, ..., 170, 170, 170],
#         [170, 170, 170, ..., 170, 170, 170]],
# 
#        [[223, 223, 223, ..., 215, 215, 215],
#         [225, 227, 225, ..., 215, 215, 215],
#         [215, 196, 203, ..., 215, 215, 215],
#         ...,
#         [211, 211, 211, ..., 211, 211, 211],
#         [211, 211, 211, ..., 211, 211, 211],
#         [211, 211, 211, ..., 211, 211, 211]],
# 
#        [[163, 163, 163, ..., 171, 171, 171],
#         [170, 178, 170, ..., 171, 171, 171],
#         [199, 196, 195, ..., 171, 171, 171],
#         ...,
#         [223, 223, 223, ..., 223, 223, 223],
#         [223, 223, 223, ..., 223, 223, 223],
#         [223, 223, 223, ..., 223, 223, 223]]], dtype=uint8)
# Coordinates:
#   * band     (band) int64 0 1 2
#   * y        (y) float64 1.663e+05 1.663e+05 1.663e+05 ... 1.272e+05 1.272e+05
#   * x        (x) float64 1.153e+07 1.153e+07 1.153e+07 ... 1.158e+07 1.158e+07

References:

Addresses first part of #2115

Reminders

Run make format and make check to make sure the code follows the style guide.
Add tests for new features or tests that would have caught the bug that you're fixing.
Add new public functions/methods/classes to doc/api/index.rst.
Write detailed docstrings for all functions/methods.
If wrapping a new module, open a 'Wrap new GMT module' issue and submit reasonably-sized PRs.
If adding new functionality, add an example to docstrings or tutorials.

Slash Commands

You can write slash commands (/command) in the first line of a comment to perform
specific operations. Supported slash commands are:

/format: automatically format and lint the code
/test-gmt-dev: run full tests on the latest GMT development version

New dataset function to load XYZ tiles! Uses contextily to retrieve the tiles based on a bounding box. Included an example doctest that shows how the map tiles can be loaded into an xarray.DataArray. Added a new section in the API docs and intersphinx mappings for contextily and xyzservices.

Can't assume that the input bounding box (which can be in longitude/latitude) is the same as the returned extent (which is always in EPSG:3857).

To fix pylint `C0103: Argument name "ll" doesn't conform to snake_case naming style (invalid-name)`.

Remove the extent and _image variables to prevent `R0914: Too many local variables (17/15) (too-many-locals)`.

Let the Continuous Integration tests run with `contextily`, include it in pyproject.toml and environment.yml, and document it in `doc/install.rst` as an optional dependency.

pygmt/datasets/map_tiles.py

Bounding box coordinates are assumed to be longitude/latitude by default, rather than in Spherical Mercator.

Using the `__doctest_requires__` variable, see https://github.com/astropy/pytest-doctestplus/tree/v0.12.1#doctest-dependencies

weiji14 · 2022-10-12T15:14:16Z

pygmt/datasets/map_tiles.py

+import numpy as np
+import xarray as xr
+
+__doctest_requires__ = {("load_map_tiles"): ["contextily"]}


Found a nice way to skip the doctest when contextily is not installed (e.g. in the Python 3.8 / NumPy 1.21 tests) using https://github.com/astropy/pytest-doctestplus/tree/v0.12.1#doctest-dependencies. Thanks to #1790 for adding pytest-doctestplus!

doc/api/index.rst

pygmt/datasets/map_tiles.py

Co-Authored-By: Dongdong Tian <[email protected]>

Also updated intersphinx link for xarray to new URL at https://docs.xarray.dev/en/stable

yvonnefroehlich · 2022-12-25T17:19:34Z

This PR was bumped back to the 0.9.0 milestone by @seisman two weeks ago. But @weiji14 you added this PR to the list of priority issues/PRs for release v.0.8.0 some days ago. Currently, this PR is still in draft status and there are some conflicts with the main branch. Do you feel you can finish this PR and there is enough time for reviewers to look at it 🙂?

weiji14 · 2022-12-28T01:52:11Z

This PR was bumped back to the 0.9.0 milestone by @seisman two weeks ago. But @weiji14 you added this PR to the list of priority issues/PRs for release v.0.8.0 some days ago. Currently, this PR is still in draft status and there are some conflicts with the main branch. Do you feel you can finish this PR and there is enough time for reviewers to look at it slightly_smiling_face?

Thanks for the ping. I've resolved the conflicts and removed the draft status. No strong opinion on whether to have it in v0.8.0 or v0.9.0, if it takes too long to review, we can merge it next year (2023).

yvonnefroehlich · 2022-12-28T08:29:44Z

This PR was bumped back to the 0.9.0 milestone by @seisman two weeks ago. But @weiji14 you added this PR to the list of priority issues/PRs for release v.0.8.0 some days ago. Currently, this PR is still in draft status and there are some conflicts with the main branch. Do you feel you can finish this PR and there is enough time for reviewers to look at it slightly_smiling_face?

Thanks for the ping. I've resolved the conflicts and removed the draft status. No strong opinion on whether to have it in v0.8.0 or v0.9.0, if it takes too long to review, we can merge it next year (2023).

Hm. As this PR was in draft status, I thought you plan more changes here. But if it is already ready for review let's see, if people
can have a look at it.

seisman · 2023-02-28T02:20:48Z

pygmt/datasets/tile_map.py

+    source : xyzservices.TileProvider or str
+        Optional. The tile source: web tile provider or path to a local file.
+        The web tile provider can be in the form of a
+        :class:`xyzservices.TileProvider` object or a URL. The placeholders for
+        the XYZ in the URL need to be {x}, {y}, {z}, respectively. For local
+        file paths, the file is read with :doc:`rasterio <rasterio:index>` and
+        all bands are loaded into the basemap. IMPORTANT: tiles are assumed to
+        be in the Spherical Mercator projection (EPSG:3857). [Default is
+        ``xyzservices.providers.Stamen.Terrain``, i.e. Stamen Terrain web
+        tiles].


Currently, xyzservices.TileProvider is linked to its API documentation (https://xyzservices.readthedocs.io/en/stable/api.html#xyzservices.TileProvider) which is very technical.

Perhaps we should add a link to https://contextily.readthedocs.io/en/latest/intro_guide.html#Providers so that users can quickly see the list of all available providers?

Ok, I've linked to https://contextily.readthedocs.io/en/stable/providers_deepdive.html which seems a bit better (actually mentioned before in #2125 (comment)) and split out the three possible source options as bullet points. Done in commit aa2d5fb. Take a look to see if it's ok.

Split out the three possible source options into bullet points. Link to https://contextily.readthedocs.io/en/latest/providers_deepdive.html, give an example OpenStreetMap URL, and link to https://contextily.readthedocs.io/en/stable/working_with_local_files.html.

weiji14 · 2023-03-02T05:42:35Z

Perhaps we should convert the returned image to longitude/latitude because it's more commonly used than the Spherical Mercator coordinate system, following https://contextily.readthedocs.io/en/latest/warping_guide.html.

Most tiles are originally served as Web Mercator (EPSG:3857), and reprojecting to latlon (EPSG:4326) would result in distortion. I'd prefer the reprojection to be a user controlled step (e.g. using rioxarray's .rio.to_crs), because if latlon isn't what the user wants, there would be double reprojection (EPSG:3857 -＞EPSG:4326 -＞EPSG:????) which is less accurate than single reprojection (EPSG:3857 -＞EPSG:????).

What about having an option so that users can decide on the desired projection and don't have to learn the syntax of the rioxarray package?

Maybe in a separate PR?

This PR looks good to me, but I'm still unsure about the default coordinate system of the returned grid. Currently it's EPSG:3857. Are we going to change the default coordinate system to EPSG:4326 if the option mentioned above is implemented?

Even with the option to reproject to EPSG:4326, I still think the default should be EPSG:3857 since most web tiles are served using EPSG:3857 as mentioned in #2125 (comment). We shouldn't be reprojecting (and causing a distortion in the pixels) by default to EPSG:4326.

pygmt/datasets/tile_map.py

michaelgrund · 2023-03-02T08:21:41Z

Will be a great improvement!

Co-authored-by: Michael Grund <[email protected]>

weiji14 · 2023-03-02T08:30:55Z

Thanks everyone! I'll leave this open for another day in case there's anything else.

yvonnefroehlich · 2023-03-02T08:56:45Z

pygmt/datasets/tile_map.py

+    zoom : int
+        Optional. Level of detail. [Default is 'auto'].


Hm, from the default it seems like that there is also string input possible.
Currently, I am wondering what "level of detail" means. Do users have to pass a percentage value to zoom? Or is zoom something similar we have for rivers of pygmt.Figure.coast? Maybe we should also state what "auto" actually means?

Suggested change

zoom : int

Optional. Level of detail. [Default is 'auto'].

zoom : int or str

Optional. Level of detail [Default is ``"auto"``].

Good catch. Yes, I did wonder if people actually knew what to use for zoom! The zoom here refers to an integer from 0 to 22, where 0 is the most zoomed out (i.e. the whole globe), and 22 is the most zoomed in (i.e. you can see buildings. See https://docs.mapbox.com/help/glossary/zoom-level or https://leafletjs.com/examples/zoom-levels for a good definition.

There is a small complication in that not all tile sources supports zoom up to level 22, so it's a bit hard to just document something like "Use an integer between 0 and 22 (inclusive)". But let me try adding a better description of this zoom parameter.

Ok, more detail added in 9f14d31. Let me know if that reads ok!

Agree that the description of zoom should be expanded a little bit since in my opinion this is a great (GMT-) improvement for people doing work on small scale levels such as cities or communities and want to have geographic content (streets, buidldings, parks etc.) in the background as reference. I missed such things in the past and used other libraries or created customized stuff. Definitely a feature we should highlight in the next release! Great work @weiji14 😉

Thanks @weiji14 for expanding the docstring for zoom! Now it is clearer how to use this parameter 🙂.

Co-Authored-By: Yvonne Fröhlich <[email protected]>

michaelgrund

The description of zoom should be sufficient to understand what the parameter does.

pygmt/datasets/tile_map.py

Co-authored-by: Michael Grund <[email protected]>

pygmt/datasets/tile_map.py

Co-authored-by: Yvonne Fröhlich <[email protected]>

seisman · 2023-03-03T05:43:42Z

The doctest fails for the CI run with Windows + Python 3.11 + NumPy 1.24 (see https://github.com/GenericMappingTools/pygmt/actions/runs/4318832687/jobs/7537596112), but passes on all other CI runs.

_______________ [doctest] pygmt.datasets.tile_map.load_tile_map _______________
094     >>> import contextily
095     >>> from pygmt.datasets import load_tile_map
096     >>> raster = load_tile_map(
097     ...     region=[103.60, 104.06, 1.22, 1.49],  # West, East, South, North
098     ...     source=contextily.providers.Stamen.TerrainBackground,
099     ...     lonlat=True,  # bounding box coordinates are longitude/latitude
100     ... )
101     >>> raster.sizes
102     Frozen({'band': 3, 'y': 1024, 'x': 1536})
103     >>> raster.coords
Differences (unified diff with -expected +actual):
    @@ -1,4 +1,4 @@
     Coordinates:
    -  * band     (band) int64 0 1 2
    -  * y        (y) float64 1.663e+05 1.663e+05 1.663e+05 ... 1.272e+05 ...
    -  * x        (x) float64 1.153e+07 1.153e+07 1.153e+07 ... 1.158e+07 ...
    +  * band     (band) int32 0 1 2
    +  * y        (y) float64 1.663e+05 1.663e+05 1.663e+05 ... 1.272e+05 1.272e+05
    +  * x        (x) float64 1.153e+07 1.153e+07 1.153e+07 ... 1.158e+07 1.158e+07

pygmt/datasets/tile_map.py

weiji14 · 2023-03-03T22:02:00Z

The doctest fails for the CI run with Windows + Python 3.11 + NumPy 1.24 (see https://github.com/GenericMappingTools/pygmt/actions/runs/4318832687/jobs/7537596112), but passes on all other CI runs.

_______________ [doctest] pygmt.datasets.tile_map.load_tile_map _______________
094     >>> import contextily
095     >>> from pygmt.datasets import load_tile_map
096     >>> raster = load_tile_map(
097     ...     region=[103.60, 104.06, 1.22, 1.49],  # West, East, South, North
098     ...     source=contextily.providers.Stamen.TerrainBackground,
099     ...     lonlat=True,  # bounding box coordinates are longitude/latitude
100     ... )
101     >>> raster.sizes
102     Frozen({'band': 3, 'y': 1024, 'x': 1536})
103     >>> raster.coords
Differences (unified diff with -expected +actual):
    @@ -1,4 +1,4 @@
     Coordinates:
    -  * band     (band) int64 0 1 2
    -  * y        (y) float64 1.663e+05 1.663e+05 1.663e+05 ... 1.272e+05 ...
    -  * x        (x) float64 1.153e+07 1.153e+07 1.153e+07 ... 1.158e+07 ...
    +  * band     (band) int32 0 1 2
    +  * y        (y) float64 1.663e+05 1.663e+05 1.663e+05 ... 1.272e+05 1.272e+05
    +  * x        (x) float64 1.153e+07 1.153e+07 1.153e+07 ... 1.158e+07 1.158e+07

Hmm, I don't see why the band coordinate would be int32 on Windows but int64 on Linux. Also the fact that it's only on Windows+NumPy 1.24? To be honest though, that dtype should really just be uint8.

The y and x coordinates are of different lengths, but the numbers look the same. Not sure what we can do about that, but let me open up a patch.

weiji14 added the feature Brand new feature label Sep 19, 2022

weiji14 added this to the 0.8.0 milestone Sep 19, 2022

weiji14 self-assigned this Sep 19, 2022

Format to remove triple dots

917604d

weiji14 mentioned this pull request Sep 19, 2022

Integration with contextily to plot xyz basemaps #2115

Closed

weiji14 added 5 commits September 19, 2022 13:42

Use correct Spherical Mercator coordinates

66f66cc

Can't assume that the input bounding box (which can be in longitude/latitude) is the same as the returned extent (which is always in EPSG:3857).

Fix coordinates for doctest

e99ed4d

Change ll parameter to lonlat

3c437c0

To fix pylint `C0103: Argument name "ll" doesn't conform to snake_case naming style (invalid-name)`.

Reduce number of local variables to fix pylint R0914

c738375

Remove the extent and _image variables to prevent `R0914: Too many local variables (17/15) (too-many-locals)`.

Add contextily to CI build matrix and include it as optional dependency

30fd9a2

Let the Continuous Integration tests run with `contextily`, include it in pyproject.toml and environment.yml, and document it in `doc/install.rst` as an optional dependency.

seisman reviewed Sep 20, 2022

View reviewed changes

pygmt/datasets/map_tiles.py Outdated Show resolved Hide resolved

weiji14 added 3 commits September 20, 2022 16:37

Set default bounding box coordinates to be lonlat

f913222

Bounding box coordinates are assumed to be longitude/latitude by default, rather than in Spherical Mercator.

Merge branch 'main' into contextily/load_map_tiles

1270c3f

Skip doctest when contextily is not installed

0455bb4

Using the `__doctest_requires__` variable, see https://github.com/astropy/pytest-doctestplus/tree/v0.12.1#doctest-dependencies

weiji14 commented Oct 12, 2022

View reviewed changes

seisman reviewed Oct 14, 2022

View reviewed changes

doc/api/index.rst Outdated Show resolved Hide resolved

pygmt/datasets/map_tiles.py Outdated Show resolved Hide resolved

pygmt/datasets/map_tiles.py Outdated Show resolved Hide resolved

weiji14 and others added 4 commits November 4, 2022 15:18

Merge branch 'main' into contextily/load_map_tiles

8b5317f

Fix typo

6142b40

Co-Authored-By: Dongdong Tian <[email protected]>

Add intersphinx link for rasterio

5400710

Also updated intersphinx link for xarray to new URL at https://docs.xarray.dev/en/stable

Merge branch 'main' into contextily/load_map_tiles

8aa1c7a

seisman mentioned this pull request Dec 6, 2022

List deprecated load_* functions in a separate table #2228

Merged

6 tasks

seisman modified the milestones: 0.8.0, 0.9.0 Dec 11, 2022

weiji14 mentioned this pull request Dec 20, 2022

Release PyGMT v0.8.0 #2244

Closed

65 tasks

Merge branch 'main' into contextily/load_map_tiles

bf1c007

weiji14 marked this pull request as ready for review December 28, 2022 01:44

seisman reviewed Feb 28, 2023

View reviewed changes

seisman approved these changes Mar 2, 2023

View reviewed changes

seisman added final review call This PR requires final review and approval from a second reviewer and removed needs review This PR has higher priority and needs review. labels Mar 2, 2023

michaelgrund approved these changes Mar 2, 2023

View reviewed changes

pygmt/datasets/tile_map.py Outdated Show resolved Hide resolved

pygmt/datasets/tile_map.py Outdated Show resolved Hide resolved

pygmt/datasets/tile_map.py Outdated Show resolved Hide resolved

pygmt/datasets/tile_map.py Outdated Show resolved Hide resolved

Remove extra fullstops and fix typos

d74f878

Co-authored-by: Michael Grund <[email protected]>

yvonnefroehlich reviewed Mar 2, 2023

View reviewed changes

Add more detail about the zoom level of detail

9f14d31

Co-Authored-By: Yvonne Fröhlich <[email protected]>

michaelgrund approved these changes Mar 2, 2023

View reviewed changes

pygmt/datasets/tile_map.py Outdated Show resolved Hide resolved

Remove another fullstop

b02b50c

Co-authored-by: Michael Grund <[email protected]>

yvonnefroehlich reviewed Mar 2, 2023

View reviewed changes

pygmt/datasets/tile_map.py Outdated Show resolved Hide resolved

pygmt/datasets/tile_map.py Outdated Show resolved Hide resolved

yvonnefroehlich reviewed Mar 2, 2023

View reviewed changes

pygmt/datasets/tile_map.py Outdated Show resolved Hide resolved

pygmt/datasets/tile_map.py Outdated Show resolved Hide resolved

pygmt/datasets/tile_map.py Outdated Show resolved Hide resolved

pygmt/datasets/tile_map.py Outdated Show resolved Hide resolved

Wrap integers in backticks and fix some typos

10af781

Co-authored-by: Yvonne Fröhlich <[email protected]>

weiji14 merged commit f379164 into main Mar 2, 2023

weiji14 deleted the contextily/load_map_tiles branch March 2, 2023 22:31

weiji14 removed the final review call This PR requires final review and approval from a second reviewer label Mar 2, 2023

yvonnefroehlich reviewed Mar 3, 2023

View reviewed changes

pygmt/datasets/tile_map.py Show resolved Hide resolved

This was referenced Mar 3, 2023

Fix bug with different band dtype from load_tile_map by casting to uint8 #2393

Merged

Add Figure.tilemap to plot XYZ tile maps #2394

Merged

weiji14 mentioned this pull request May 2, 2023

CI: Install geopandas using pip in the "GMT Dev Tests" workflow #2525

Merged

7 tasks

weiji14 mentioned this pull request Jan 15, 2024

Update failing doctests with upstream dataset changes #2988

Merged

weiji14 mentioned this pull request Jul 8, 2024

**BREAKING** pygmt.grdcut: Refactor to store output in virtualfiles for grids #3115

Open

seisman mentioned this pull request Oct 4, 2024

load_map_tiles: Allow a new parameter to control the projection of the returned DataArray #3484

Closed

seisman mentioned this pull request Nov 26, 2024

load_tile_map: Add the new parameter 'crs' to set the CRS of the returned dataarray #3554

Merged

7 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add function to load raster tile maps using contextily #2125

Add function to load raster tile maps using contextily #2125

weiji14 commented Sep 19, 2022 •

edited

Loading

weiji14 Oct 12, 2022

yvonnefroehlich commented Dec 25, 2022 •

edited

Loading

weiji14 commented Dec 28, 2022

yvonnefroehlich commented Dec 28, 2022

seisman Feb 28, 2023

weiji14 Mar 2, 2023

weiji14 commented Mar 2, 2023

michaelgrund commented Mar 2, 2023

weiji14 commented Mar 2, 2023

yvonnefroehlich Mar 2, 2023

weiji14 Mar 2, 2023 •

edited

Loading

weiji14 Mar 2, 2023

michaelgrund Mar 2, 2023

yvonnefroehlich Mar 2, 2023

michaelgrund left a comment

seisman commented Mar 3, 2023

weiji14 commented Mar 3, 2023

Add function to load raster tile maps using contextily #2125

Add function to load raster tile maps using contextily #2125

Conversation

weiji14 commented Sep 19, 2022 • edited Loading

weiji14 Oct 12, 2022

Choose a reason for hiding this comment

yvonnefroehlich commented Dec 25, 2022 • edited Loading

weiji14 commented Dec 28, 2022

yvonnefroehlich commented Dec 28, 2022

seisman Feb 28, 2023

Choose a reason for hiding this comment

weiji14 Mar 2, 2023

Choose a reason for hiding this comment

weiji14 commented Mar 2, 2023

michaelgrund commented Mar 2, 2023

weiji14 commented Mar 2, 2023

yvonnefroehlich Mar 2, 2023

Choose a reason for hiding this comment

weiji14 Mar 2, 2023 • edited Loading

Choose a reason for hiding this comment

weiji14 Mar 2, 2023

Choose a reason for hiding this comment

michaelgrund Mar 2, 2023

Choose a reason for hiding this comment

yvonnefroehlich Mar 2, 2023

Choose a reason for hiding this comment

michaelgrund left a comment

Choose a reason for hiding this comment

seisman commented Mar 3, 2023

weiji14 commented Mar 3, 2023

weiji14 commented Sep 19, 2022 •

edited

Loading

yvonnefroehlich commented Dec 25, 2022 •

edited

Loading

weiji14 Mar 2, 2023 •

edited

Loading