From 8a59ec0b622b478257d54e2aedf522124ae75f41 Mon Sep 17 00:00:00 2001
From: Cresson Remi <remi.cresson@irstea.fr>
Date: Wed, 6 Mar 2024 20:16:59 +0100
Subject: [PATCH 01/26] Update 2 files
- /simplestac/extents.py
- /pyproject.toml
---
pyproject.toml | 1 +
simplestac/extents.py | 113 +++++++++++++++---------------------------
2 files changed, 40 insertions(+), 74 deletions(-)
diff --git a/pyproject.toml b/pyproject.toml
index 7dd4f3d..405ec2f 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -26,6 +26,7 @@ dependencies = [
"rioxarray",
"stac_static@git+https://github.com/jsignell/stac-static",
"stackstac",
+ "rtree",
]
[project.optional-dependencies]
diff --git a/simplestac/extents.py b/simplestac/extents.py
index 1979f67..379898d 100644
--- a/simplestac/extents.py
+++ b/simplestac/extents.py
@@ -1,12 +1,12 @@
"""
-Module to deal with STAC Extents.
+Deal with STAC Extents.
"""
import pystac
from dataclasses import dataclass
from datetime import datetime
from typing import Union
-
-Coords = list[Union[int, float]]
+from stacflow.common_types import Bbox
+from rtree import index
@dataclass
@@ -14,7 +14,7 @@ class SmartBbox:
"""
Small class to work with a single 2D bounding box.
"""
- coords: Coords = None # [xmin, ymin, xmax, ymax]
+ coords: Bbox = None # [xmin, ymin, xmax, ymax]
def touches(self, other: "SmartBbox") -> bool:
"""
@@ -53,7 +53,7 @@ class SmartBbox:
]
-def clusterize_bboxes(bboxes: list[Coords]) -> list[Coords]:
+def clusterize_bboxes(bboxes: list[Bbox]) -> list[Bbox]:
"""
Computes a list of bounding boxes regrouping all overlapping ones.
@@ -64,74 +64,38 @@ def clusterize_bboxes(bboxes: list[Coords]) -> list[Coords]:
list of 2D bounding boxes (list of int of float)
"""
- # Regroup bboxes into clusters of bboxes
- smart_bboxes = [SmartBbox(bbox) for bbox in bboxes]
- clusters = bboxes_to_bboxes_clusters(smart_bboxes)
-
- # Compute clusters extents
- clusters_unions = [SmartBbox() for _ in clusters]
- for i, cluster in enumerate(clusters):
- for smart_bbox in cluster:
- clusters_unions[i].update(smart_bbox)
-
- return [smart_bbox.coords.copy() for smart_bbox in clusters_unions]
-
-
-def bboxes_to_bboxes_clusters(smart_bboxes: list[SmartBbox]) -> list[
- list[SmartBbox]]:
- """
- Transform a list of bounding boxes into a nested list of clustered ones.
-
- Args:
- smart_bboxes: a list of `SmartBbox` instances
-
- Returns:
- a list of `SmartBbox` instances list
-
- """
- clusters_labels = compute_smart_bboxes_clusters(smart_bboxes)
- clusters_bboxes = [[] for _ in range(max(clusters_labels) + 1)]
- for smart_bbox, labels in zip(smart_bboxes, clusters_labels):
- clusters_bboxes[labels].append(smart_bbox)
- return clusters_bboxes
-
-
-def compute_smart_bboxes_clusters(smart_bboxes: list[SmartBbox]) -> list[int]:
- """
- Compute the extent of a cluster of `SmartBbox` instances.
-
- Args:
- smart_bboxes: a list of `SmartBbox` instances
-
- Returns:
- a vector of same size as `smart_bboxes` with the group numbers (int)
-
- """
- labels = len(smart_bboxes) * [None]
- group = 0
-
- def dfs(index: int):
- """
- Deep first search with o(n) complexity.
+ bboxes = [SmartBbox(bbox) for bbox in bboxes]
+ idx = index.Index()
+ for i, bbox in enumerate(bboxes):
+ idx.insert(id=i, coordinates=bbox.coords)
+ clusters = [bboxes.pop()]
+
+ def _update_clusters_idx():
+ _clusters_idx = index.Index()
+ for _i, cluster_bbox in enumerate(clusters):
+ _clusters_idx.insert(id=_i, coordinates=cluster_bbox.coords)
+ return _clusters_idx
+
+ clusters_idx = _update_clusters_idx()
+ while bboxes:
+ bbox = bboxes.pop()
+ inter_clusters = list(clusters_idx.intersection(bbox.coords))
+ if inter_clusters:
+ # We merge all intersecting clusters into one
+ clusters[inter_clusters[0]].update(bbox)
+ for i in inter_clusters[1:]:
+ clusters[inter_clusters[0]].update(clusters[i])
+ clusters = [
+ cluster
+ for i, cluster in enumerate(clusters)
+ if i not in inter_clusters[1:]
+ ]
+ else:
+ clusters.append(bbox)
- Args:
- index: vertex index.
+ clusters_idx = _update_clusters_idx()
- """
- labels[index] = group
- cur_item = smart_bboxes[index]
- for i, (item, label) in enumerate(zip(smart_bboxes, labels)):
- if i != index and cur_item.touches(item) and labels[i] is None:
- dfs(i)
-
- while any(label is None for label in labels):
- next_unmarked = next(
- i for i, label in enumerate(labels)
- if label is None
- )
- dfs(next_unmarked)
- group += 1
- return labels
+ return [cluster.coords for cluster in clusters]
class AutoSpatialExtent(pystac.SpatialExtent):
@@ -151,7 +115,7 @@ class AutoSpatialExtent(pystac.SpatialExtent):
super().__init__(*args, **kwargs)
self.clusterize_bboxes()
- def update(self, other: pystac.SpatialExtent | Coords):
+ def update(self, other: pystac.SpatialExtent | Bbox):
"""
Updates itself with a new spatial extent or bounding box. Modifies
inplace `self.bboxes`.
@@ -208,9 +172,10 @@ class AutoTemporalExtent(pystac.TemporalExtent):
all_dates = []
for interval in self.intervals:
if isinstance(interval, (list, tuple)):
- all_dates += [i for i in interval]
+ all_dates += [i for i in interval if i is not None]
elif isinstance(interval, datetime):
all_dates.append(interval)
else:
TypeError(f"Unsupported date/range of: {interval}")
- self.intervals = [[min(all_dates), max(all_dates)]]
+ self.intervals = \
+ [[min(all_dates), max(all_dates)]] if all_dates else [None, None]
--
GitLab
From 36ae451e2fd102940d8f26591edbec125f778a61 Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Mon, 11 Mar 2024 14:21:52 +0100
Subject: [PATCH 02/26] update some docstring
---
simplestac/utils.py | 10 +++++++---
1 file changed, 7 insertions(+), 3 deletions(-)
diff --git a/simplestac/utils.py b/simplestac/utils.py
index fecce72..20459d8 100644
--- a/simplestac/utils.py
+++ b/simplestac/utils.py
@@ -610,19 +610,23 @@ def update_item_properties(x: pystac.Item, remove_item_props=DEFAULT_REMOVE_PROP
x.properties.pop(k)
-def harmonize_sen2cor_offet(collection, bands=S2_SEN2COR_BANDS, inplace=False):
+def harmonize_sen2cor_offet(x, bands=S2_SEN2COR_BANDS, inplace=False):
"""
Harmonize new Sentinel-2 item collection (Sen2Cor v4+, 2022-01-25)
to the old baseline (v3-):
- adds an offset of -1000 to all band assets of items
+ adds an offset of -1000 to the asset extra field "raster:bands" of the items
with datetime >= 2022-01-25
Parameters
----------
- data: ItemCollection
+ x: ItemCollection
An item collection of S2 scenes
bands: list
A list of bands to harmonize
+
+ inplace: bool
+ Whether to modify the collection in place. Defaults to False.
+ In that case, a cloned collection is returned.
Returns
-------
--
GitLab
From 7895a526a40a1a13da55d8e436c3e7fb1a91b56f Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Mon, 11 Mar 2024 15:35:29 +0100
Subject: [PATCH 03/26] update changelog and CI
---
.gitlab-ci.yml | 3 +++
CHANGELOG.md | 18 ++++++++++++++++++
2 files changed, 21 insertions(+)
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index 8a24467..5df5950 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -16,6 +16,9 @@ tests:
- micromamba env list
- pip install -e .
- pytest -s
+ only:
+ - merge-request
+ - schedules
pages:
stage: pages
diff --git a/CHANGELOG.md b/CHANGELOG.md
index e69de29..0f2699d 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -0,0 +1,18 @@
+# v1.1.0
+
+## Add
+
+- function `write_assets`: write item assets (rasters only at the moment) of an ItemCollection locally and return the corresponding ItemCollection.
+- function `harmonize_sen2cor_offset`: adds an `offset` property to the assets so it is taken into account by `to_xarray`.
+- method `ItemCollection.drop_duplicates`: drop duplicated ID returned by pgstac.
+- method `ItemCollection.drop_non_raster`: drop non raster assets.
+- `writer_args` to `ItemCollection.apply_...` methods and function in order to specify the outputs format, e.g. the encoding.
+- in local.py, `start_datetime` and `end_datetime` can now be used instead of `datetime` in the template used to build a local ItemCollection.
+- module `extents.py` to manipulate STAC extents.
+- tests for CI
+
+## Fix
+
+- `apply_formula` with "in" operator in apply_formula.
+- COG type of local STAC assets (instead of GTiff)
+- imports in `simplestac.utils`
--
GitLab
From fe5b77ca1fbb0691a85f66f74cd9b324c7294a29 Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Mon, 11 Mar 2024 17:00:24 +0100
Subject: [PATCH 04/26] amend previous commit
---
simplestac/utils.py | 88 ++++++++++++++++++++++-----------------------
1 file changed, 43 insertions(+), 45 deletions(-)
diff --git a/simplestac/utils.py b/simplestac/utils.py
index 20459d8..23b8e03 100644
--- a/simplestac/utils.py
+++ b/simplestac/utils.py
@@ -609,51 +609,6 @@ def update_item_properties(x: pystac.Item, remove_item_props=DEFAULT_REMOVE_PROP
for k in pop_props:
x.properties.pop(k)
-
-def harmonize_sen2cor_offet(x, bands=S2_SEN2COR_BANDS, inplace=False):
- """
- Harmonize new Sentinel-2 item collection (Sen2Cor v4+, 2022-01-25)
- to the old baseline (v3-):
- adds an offset of -1000 to the asset extra field "raster:bands" of the items
- with datetime >= 2022-01-25
-
- Parameters
- ----------
- x: ItemCollection
- An item collection of S2 scenes
- bands: list
- A list of bands to harmonize
-
- inplace: bool
- Whether to modify the collection in place. Defaults to False.
- In that case, a cloned collection is returned.
-
- Returns
- -------
- ItemCollection
- A collection of S2 scenes with extra_fields["raster:bands"]
- added/updated to each band asset with datetime >= 2022-01-25.
-
- Notes
- -----
- References:
- - https://planetarycomputer.microsoft.com/dataset/sentinel-2-l2a#Baseline-Change
- - https://github.com/microsoft/PlanetaryComputer/issues/134
- """
-
- if not inplace:
- collection = collection.copy()
- for item in collection:
- for asset in bands:
- if asset in item.assets:
- if item.properties["datetime"] >= "2022-01-25":
- item.assets[asset].extra_fields["raster:bands"] = [dict(offset=-1000)]
- else:
- item.assets[asset].extra_fields["raster:bands"] = [dict(offset=0)]
- if inplace:
- return collection
-
-
def apply_item(x, fun, name, output_dir, overwrite=False,
copy=True, bbox=None, geometry=None, writer_args=None, **kwargs):
"""
@@ -880,4 +835,47 @@ def apply_formula(x, formula):
return eval(formula)
+def harmonize_sen2cor_offet(x, bands=S2_SEN2COR_BANDS, inplace=False):
+ """
+ Harmonize new Sentinel-2 item collection (Sen2Cor v4+, 2022-01-25)
+ to the old baseline (v3-):
+ adds an offset of -1000 to the asset extra field "raster:bands" of the items
+ with datetime >= 2022-01-25
+
+ Parameters
+ ----------
+ x: ItemCollection
+ An item collection of S2 scenes
+ bands: list
+ A list of bands to harmonize
+
+ inplace: bool
+ Whether to modify the collection in place. Defaults to False.
+ In that case, a cloned collection is returned.
+
+ Returns
+ -------
+ ItemCollection
+ A collection of S2 scenes with extra_fields["raster:bands"]
+ added/updated to each band asset with datetime >= 2022-01-25.
+
+ Notes
+ -----
+ References:
+ - https://planetarycomputer.microsoft.com/dataset/sentinel-2-l2a#Baseline-Change
+ - https://github.com/microsoft/PlanetaryComputer/issues/134
+ """
+
+ if not inplace:
+ x = x.copy()
+ for item in x:
+ for asset in bands:
+ if asset in item.assets:
+ if item.properties["datetime"] >= "2022-01-25":
+ item.assets[asset].extra_fields["raster:bands"] = [dict(offset=-1000)]
+ else:
+ item.assets[asset].extra_fields["raster:bands"] = [dict(offset=0)]
+ if inplace:
+ return x
+
#######################################
\ No newline at end of file
--
GitLab
From 5951e2af26015b9b4972c316c3da13ecf248ddfc Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Mon, 11 Mar 2024 17:04:10 +0100
Subject: [PATCH 05/26] add function extract_points
---
simplestac/utils.py | 50 +++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 50 insertions(+)
diff --git a/simplestac/utils.py b/simplestac/utils.py
index 23b8e03..94d920a 100644
--- a/simplestac/utils.py
+++ b/simplestac/utils.py
@@ -878,4 +878,54 @@ def harmonize_sen2cor_offet(x, bands=S2_SEN2COR_BANDS, inplace=False):
if inplace:
return x
+def extract_points(x, df, **kwargs):
+ """_summary_
+
+ Parameters
+ ----------
+ x : xarray.DataArray or xarray.Dataset
+ df : pandas.DataFrame
+ Coordinates of the points
+
+ Returns
+ -------
+ xarray.DataArray or xarray.Dataset
+ The points values, tha can be converted to
+ dataframe with `to_dataframe` `to_dask_dataframe`
+
+ Examples
+ --------
+ >>> import xarray as xr
+ >>> import pandas as pd
+ >>> import dask.array
+ >>> import numpy as np
+ >>> da = xr.DataArray(
+ ... # np.random.random((100,200)),
+ ... dask.array.random.random((100,200,10), chunks=10),
+ ... coords = [('x', np.arange(100)+.5),
+ ... ('y', np.arange(200)+.5),
+ ... ('z', np.arange(10)+.5)]
+ ... ).rename("pixel_value")
+ >>> df = pd.DataFrame(
+ ... dict(
+ ... x=np.random.permutation(range(100))[:100]+np.random.random(100),
+ ... y=np.random.permutation(range(100))[:100]+np.random.random(100),
+ ... other=range(100),
+ ... )
+ ... )
+ >>> df.index.rename("id_point", inplace=True)
+ >>> extraction = extract_points(da, df, method="nearest", tolerance=.5)
+ >>> ext_df = extraction.to_dataframe()
+ >>> ext_df.reset_index(drop=False, inplace=True)
+ >>> ext_df.rename({k: k+"_pixel" for k in da.dims}, axis=1, inplace=True)
+ >>> # join extraction to original dataframe
+ >>> df.merge(ext_df, on=["id_point"])
+
+ """
+ # x = da
+ xk = x.dims
+ coords_cols = [c for c in df.keys() if c in xk]
+ coords = df[coords_cols]
+ points = x.sel(coords.to_xarray(), **kwargs)
+ return points
#######################################
\ No newline at end of file
--
GitLab
From e5e40f3178d2d75ec3185c1df1157c17341c5d6f Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Mon, 11 Mar 2024 21:30:02 +0100
Subject: [PATCH 06/26] add drop_non_raster to message for stackstac.stack
ValueError
---
simplestac/utils.py | 8 +++++++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/simplestac/utils.py b/simplestac/utils.py
index 94d920a..87ec81e 100644
--- a/simplestac/utils.py
+++ b/simplestac/utils.py
@@ -90,7 +90,13 @@ class ExtendPystacClasses:
# times = pd.to_datetime(
with warnings.catch_warnings():
warnings.filterwarnings("ignore", category=UserWarning)
- arr = stackstac.stack(self, xy_coords=xy_coords, **kwargs)
+ try:
+ arr = stackstac.stack(self, xy_coords=xy_coords, **kwargs)
+ except ValueError as e:
+ if "Cannot automatically compute the resolution" in str(e):
+ raise ValueError(str(e)+"\nOr drop non-raster assets from collection with ItemCollection.drop_non_raster()")
+ else:
+ raise e
if bbox is not None:
arr = arr.rio.clip_box(*bbox)
--
GitLab
From 4f5143119c38a5ffd671cc45c75692461d6c2a1c Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Mon, 11 Mar 2024 21:31:26 +0100
Subject: [PATCH 07/26] fix crs comparison
---
simplestac/utils.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/simplestac/utils.py b/simplestac/utils.py
index 87ec81e..463a4c6 100644
--- a/simplestac/utils.py
+++ b/simplestac/utils.py
@@ -101,7 +101,7 @@ class ExtendPystacClasses:
if bbox is not None:
arr = arr.rio.clip_box(*bbox)
if geometry is not None:
- if hasattr(geometry, 'crs') and geometry.crs != arr.rio.crs:
+ if hasattr(geometry, 'crs') and not geometry.crs.equals(arr.rio.crs):
logger.debug(f"Reprojecting geometry from {geometry.crs} to {arr.rio.crs}")
geometry = geometry.to_crs(arr.rio.crs)
arr = arr.rio.clip(geometry)
--
GitLab
From 7d21fcc0fe11883b9b2c9f41ddde3e3f8fe573d2 Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Mon, 11 Mar 2024 22:39:59 +0100
Subject: [PATCH 08/26] add method ItemCollection.extract_points
---
simplestac/utils.py | 104 ++++++++++++++++++++++++++++++++++++++++----
1 file changed, 95 insertions(+), 9 deletions(-)
diff --git a/simplestac/utils.py b/simplestac/utils.py
index 463a4c6..e1dd81a 100644
--- a/simplestac/utils.py
+++ b/simplestac/utils.py
@@ -19,6 +19,7 @@ from tqdm import tqdm
from typing import Union
import warnings
import datetime
+import geopandas as gpd
from simplestac.local import stac_asset_info_from_raster, properties_from_assets
@@ -492,6 +493,77 @@ class ExtendPystacClasses:
if not inplace:
return x
+ def extract_points(self, points, method="nearest", tolerance="pixel", drop=False, **kwargs):
+ """Extract points from xarray
+
+ Parameters
+ ----------
+ x : xarray.DataArray or xarray.Dataset
+ points : geopandas.GeoDataFrame or pandas.DataFrame
+ Points or coordinates of the points
+ method, tolerance, drop : see xarray.DataArray.sel
+ Additional keyword arguments passed to xarray.DataArray.sel
+
+ Returns
+ -------
+ xarray.DataArray or xarray.Dataset
+ The points values with points index as coordinate.
+ The returned xarray can then be converted to
+ dataframe with `to_dataframe` or `to_dask_dataframe`.
+
+ Examples
+ --------
+ >>> import xarray as xr
+ >>> import pandas as pd
+ >>> import dask.array
+ >>> import numpy as np
+ >>> da = xr.DataArray(
+ ... # np.random.random((100,200)),
+ ... dask.array.random.random((100,200,10), chunks=10),
+ ... coords = [('x', np.arange(100)+.5),
+ ... ('y', np.arange(200)+.5),
+ ... ('z', np.arange(10)+.5)]
+ ... ).rename("pixel_value")
+ >>> df = pd.DataFrame(
+ ... dict(
+ ... x=np.random.permutation(range(100))[:100]+np.random.random(100),
+ ... y=np.random.permutation(range(100))[:100]+np.random.random(100),
+ ... other=range(100),
+ ... )
+ ... )
+ >>> df.index.rename("id_point", inplace=True)
+ >>> extraction = extract_points(da, df, method="nearest", tolerance=.5)
+ >>> ext_df = extraction.to_dataframe()
+ >>> ext_df.reset_index(drop=False, inplace=True)
+ >>> ext_df.rename({k: k+"_pixel" for k in da.dims}, axis=1, inplace=True)
+ >>> # join extraction to original dataframe
+ >>> df.merge(ext_df, on=["id_point"])
+ """
+ """_summary_
+
+ Parameters
+ ----------
+ points : _type_
+ _description_
+ tolerance : float or str, optional
+
+ method, tolerance, drop : see xarray.DataArray.sel
+ Additional keyword arguments passed to xarray.DataArray.sel
+ If tolerance is "pixel", it is set to half the resolution of the xarray.
+ Returns
+ -------
+ _type_
+ _description_
+ """
+ # avoid starting anything if not all points
+ if isinstance(points, (gpd.GeoDataFrame, gpd.GeoSeries)):
+ if not points.geom_type.isin(['Point', 'MultiPoint']).all():
+ raise ValueError("All geometries must be of type Point or MultiPoint")
+
+ arr = self.to_xarray(**kwargs)#geometry=points)
+ if tolerance == "pixel":
+ tolerance = arr.rio.resolution()[0] / 2
+ return extract_points(arr, points, method=method, tolerance=tolerance, drop=drop)
class ItemCollection(pystac.ItemCollection, ExtendPystacClasses):
pass
@@ -884,20 +956,23 @@ def harmonize_sen2cor_offet(x, bands=S2_SEN2COR_BANDS, inplace=False):
if inplace:
return x
-def extract_points(x, df, **kwargs):
- """_summary_
+def extract_points(x, points, method=None, tolerance=None, drop=False):
+ """Extract points from xarray
Parameters
----------
x : xarray.DataArray or xarray.Dataset
- df : pandas.DataFrame
- Coordinates of the points
+ points : geopandas.GeoDataFrame or pandas.DataFrame
+ Points or coordinates of the points
+ method, tolerance, drop : see xarray.DataArray.sel
+ Additional keyword arguments passed to xarray.DataArray.sel
Returns
-------
xarray.DataArray or xarray.Dataset
- The points values, tha can be converted to
- dataframe with `to_dataframe` `to_dask_dataframe`
+ The points values with points index as coordinate.
+ The returned xarray can then be converted to
+ dataframe with `to_dataframe` or `to_dask_dataframe`.
Examples
--------
@@ -929,9 +1004,20 @@ def extract_points(x, df, **kwargs):
"""
# x = da
+ valid_types = (gpd.GeoDataFrame, gpd.GeoSeries)
+ if isinstance(points, valid_types):
+ if not points.geom_type.isin(['Point', 'MultiPoint']).all():
+ raise ValueError("All geometries must be of type Point")
+
+ if isinstance(points, valid_types):
+ if hasattr(points, 'crs') and not points.crs.equals(x.rio.crs):
+ logger.debug(f"Reprojecting points from {points.crs} to {x.rio.crs}")
+ points = points.to_crs(x.rio.crs)
+ points = points.get_coordinates()
+
xk = x.dims
- coords_cols = [c for c in df.keys() if c in xk]
- coords = df[coords_cols]
- points = x.sel(coords.to_xarray(), **kwargs)
+ coords_cols = [c for c in points.keys() if c in xk]
+ coords = points[coords_cols]
+ points = x.sel(coords.to_xarray(), method=method, tolerance=tolerance, drop=drop)
return points
#######################################
\ No newline at end of file
--
GitLab
From 4821f2080e93b65fbfef50a1ec9fbfeeeb3de5b6 Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Mon, 11 Mar 2024 22:40:36 +0100
Subject: [PATCH 09/26] add tests
---
tests/test_local.py | 20 +++++++++++++++++---
tests/test_remote.py | 33 ++++++++++++++++++++++++++++++++-
2 files changed, 49 insertions(+), 4 deletions(-)
diff --git a/tests/test_local.py b/tests/test_local.py
index 676448c..3471050 100644
--- a/tests/test_local.py
+++ b/tests/test_local.py
@@ -2,6 +2,9 @@ from simplestac.local import collection_format, build_item_collection
from simplestac.utils import write_raster, apply_formula
import xarray as xr
import pystac
+from shapely.geometry import MultiPoint
+import geopandas as gpd
+
def test_formatting():
fmt = collection_format()
@@ -86,7 +89,7 @@ def test_apply_items_raster_args(s2scene_dir, roi):
dtype="int16",
scale_factor=0.001,
add_offset=0.0,
- _FillValue= 2**15 - 1,
+ _FillValue= -2**15,
),
)
)
@@ -95,9 +98,20 @@ def test_apply_items_raster_args(s2scene_dir, roi):
assert rb["datatype"] == "int16"
assert rb["scale"] == 0.001
assert rb["offset"] == 0.0
- assert rb["nodata"] == 2**15 - 1
+ assert rb["nodata"] == -2**15
+
+def test_extract_points(s2scene_dir, roi):
+ col = build_item_collection(s2scene_dir, collection_format())
+ points = roi.geometry.apply(lambda x: MultiPoint(list(x.exterior.coords)))
+ points.index.rename("id_point", inplace=True)
+ ext = col.extract_points(points)
+ assert ext.id_point.isin(points.index.values).all()
+ coords = points.get_coordinates().reset_index(drop=True)
+ points = gpd.GeoSeries(gpd.points_from_xy(**coords, crs=roi.crs))
+ points.index.rename("id_point", inplace=True)
+ ext = col.extract_points(points)
+ assert ext.id_point.isin(points.index.values).all()
-
############################################################
diff --git a/tests/test_remote.py b/tests/test_remote.py
index 3fced92..beb2b82 100644
--- a/tests/test_remote.py
+++ b/tests/test_remote.py
@@ -3,6 +3,21 @@ import planetary_computer as pc
import pystac_client
URL = "https://planetarycomputer.microsoft.com/api/stac/v1"
+
+def test_to_xarray(roi, s2scene_pc_dir):
+ time_range = "2022-01-01/2022-01-31"
+
+ catalog = pystac_client.Client.open(URL, modifier=pc.sign_inplace)
+ search = catalog.search(
+ collections=["sentinel-2-l2a"],
+ bbox=roi.to_crs(4326).total_bounds,
+ datetime=time_range,
+ sortby="datetime",
+ )
+ col = ItemCollection(search.item_collection())
+ x = col.drop_non_raster().to_xarray()
+ assert len(x.time) == len(col)
+
def test_offset_harmonization(roi, s2scene_pc_dir):
time_range = "2022-01-20/2022-01-31"
@@ -21,7 +36,23 @@ def test_offset_harmonization(roi, s2scene_pc_dir):
assert of0 == 0
assert ofN == -1000
+def test_drop_duplicates(roi, s2scene_pc_dir):
+ time_range = "2022-01-20/2022-01-31"
+ catalog = pystac_client.Client.open(URL, modifier=pc.sign_inplace)
+ search = catalog.search(
+ collections=["sentinel-2-l2a"],
+ bbox=roi.to_crs(4326).total_bounds,
+ datetime=time_range,
+ sortby="datetime",
+ )
+ col = search.item_collection()
+ col1 = ItemCollection(col.clone()+col.clone())
+ col1.drop_duplicates(inplace=True)
+ assert len(col1) == len(col)
+
def test_write_assets(roi, s2scene_pc_dir):
+
+ s2scene_pc_dir.rmtree_p().mkdir_p()
time_range = "2016-01-01/2016-01-31"
catalog = pystac_client.Client.open(URL)
@@ -35,7 +66,6 @@ def test_write_assets(roi, s2scene_pc_dir):
col = ItemCollection(search.item_collection()).drop_non_raster()
bbox = roi.to_crs(col.to_xarray().rio.crs).total_bounds
col = pc.sign(col)
- s2scene_pc_dir.rmtree_p().mkdir_p()
encoding=dict(
dtype="int16",
scale_factor=0.001,
@@ -48,6 +78,7 @@ def test_write_assets(roi, s2scene_pc_dir):
item = new_col[0]
assert item.id == col[0].id
assert len(item.assets) == len(s2scene_pc_dir.dirs()[0].files("*.tif"))
+ assert item.assets["B08"].href.startswith(s2scene_pc_dir)
assert new_col[0].assets["B08"].extra_fields["raster:bands"][0]["scale"] == 0.001
--
GitLab
From 543760a9634b9aeb12fcfa9ff9e9330ee7b21a33 Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Mon, 11 Mar 2024 23:05:33 +0100
Subject: [PATCH 10/26] fix extract_points docstring
---
simplestac/utils.py | 21 +++++----------------
1 file changed, 5 insertions(+), 16 deletions(-)
diff --git a/simplestac/utils.py b/simplestac/utils.py
index e1dd81a..b038582 100644
--- a/simplestac/utils.py
+++ b/simplestac/utils.py
@@ -503,7 +503,11 @@ class ExtendPystacClasses:
Points or coordinates of the points
method, tolerance, drop : see xarray.DataArray.sel
Additional keyword arguments passed to xarray.DataArray.sel
-
+ If tolerance is "pixel", it is set to half the resolution
+ of the xarray, supposing it is a rioxarray.
+ **kwargs:
+ Additional keyword arguments passed to `ItemCollection.to_xarray()`
+
Returns
-------
xarray.DataArray or xarray.Dataset
@@ -539,22 +543,7 @@ class ExtendPystacClasses:
>>> # join extraction to original dataframe
>>> df.merge(ext_df, on=["id_point"])
"""
- """_summary_
-
- Parameters
- ----------
- points : _type_
- _description_
- tolerance : float or str, optional
- method, tolerance, drop : see xarray.DataArray.sel
- Additional keyword arguments passed to xarray.DataArray.sel
- If tolerance is "pixel", it is set to half the resolution of the xarray.
- Returns
- -------
- _type_
- _description_
- """
# avoid starting anything if not all points
if isinstance(points, (gpd.GeoDataFrame, gpd.GeoSeries)):
if not points.geom_type.isin(['Point', 'MultiPoint']).all():
--
GitLab
From 9a5500b433d4c33dddbc4e8840d9b114f12c3958 Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Mon, 11 Mar 2024 23:25:43 +0100
Subject: [PATCH 11/26] try other CI rules
---
.gitlab-ci.yml | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/.gitlab-ci.yml b/.gitlab-ci.yml
index 5df5950..41bc8c8 100644
--- a/.gitlab-ci.yml
+++ b/.gitlab-ci.yml
@@ -16,9 +16,9 @@ tests:
- micromamba env list
- pip install -e .
- pytest -s
- only:
- - merge-request
- - schedules
+ rules:
+ - if: $CI_PIPELINE_SOURCE == "merge_request_event"
+ - if: $CI_COMMIT_BRANCH =~ /.*/
pages:
stage: pages
--
GitLab
From 4a4adea9f28ca74c2b0193065f5f1cbb654d898a Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Mon, 11 Mar 2024 23:42:24 +0100
Subject: [PATCH 12/26] update changelog
---
CHANGELOG.md | 1 +
1 file changed, 1 insertion(+)
diff --git a/CHANGELOG.md b/CHANGELOG.md
index 0f2699d..6387fd9 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -6,6 +6,7 @@
- function `harmonize_sen2cor_offset`: adds an `offset` property to the assets so it is taken into account by `to_xarray`.
- method `ItemCollection.drop_duplicates`: drop duplicated ID returned by pgstac.
- method `ItemCollection.drop_non_raster`: drop non raster assets.
+- function `extract_points` and method `ItemCollection.extract_points` to extract points time series.
- `writer_args` to `ItemCollection.apply_...` methods and function in order to specify the outputs format, e.g. the encoding.
- in local.py, `start_datetime` and `end_datetime` can now be used instead of `datetime` in the template used to build a local ItemCollection.
- module `extents.py` to manipulate STAC extents.
--
GitLab
From 6aef6c38490b5b163ba17ef71104fbf3e905f429 Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Tue, 12 Mar 2024 00:48:31 +0100
Subject: [PATCH 13/26] add version number
---
simplestac/__init__.py | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/simplestac/__init__.py b/simplestac/__init__.py
index e69de29..e60be81 100644
--- a/simplestac/__init__.py
+++ b/simplestac/__init__.py
@@ -0,0 +1,7 @@
+from importlib.metadata import version, PackageNotFoundError
+
+try:
+ __version__ = version("simplestac")
+except PackageNotFoundError:
+ # package is not installed
+ pass
\ No newline at end of file
--
GitLab
From d540f1cddb8ef49302654543c66b069c1f896ea9 Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Tue, 12 Mar 2024 00:52:31 +0100
Subject: [PATCH 14/26] update gitignore
---
.gitignore | 1 +
1 file changed, 1 insertion(+)
diff --git a/.gitignore b/.gitignore
index 8d4c383..5c1a962 100644
--- a/.gitignore
+++ b/.gitignore
@@ -7,3 +7,4 @@ __pycache__
tests/data*
draft
examples/data
+docs/examples
\ No newline at end of file
--
GitLab
From 1c4d2825520eacefdcd5a44e54e888568a113902 Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Tue, 12 Mar 2024 01:38:07 +0100
Subject: [PATCH 15/26] add light mode in pages
---
pyproject_doc.toml | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/pyproject_doc.toml b/pyproject_doc.toml
index 1fa654e..6d948de 100644
--- a/pyproject_doc.toml
+++ b/pyproject_doc.toml
@@ -53,7 +53,14 @@ features=["navigation.top"]
icon = "material/theme-light-dark"
name = "Switch to light mode"
+ [[tool.portray.mkdocs.theme.palette]]
+ scheme = "default"
+ primary = "teal"
+ accent = "blue"
+ [tool.portray.mkdocs.theme.palette.toggle]
+ icon = "material/theme-light-dark"
+ name = "Switch to dark mode"
# favicon = "docs/icon/....png"
# logo = "docs/icon/....png"
--
GitLab
From bec9ebba7712d5b9ed26c18b2b95323104ec70bc Mon Sep 17 00:00:00 2001
From: Cresson Remi <remi.cresson@irstea.fr>
Date: Tue, 12 Mar 2024 12:16:19 +0100
Subject: [PATCH 16/26] Update 3 files
- /simplestac/extents.py
- /pyproject.toml
- /tests/extents.py
---
pyproject.toml | 1 -
simplestac/extents.py | 17 +++--------------
tests/extents.py | 20 ++++++++++++++++++++
3 files changed, 23 insertions(+), 15 deletions(-)
create mode 100644 tests/extents.py
diff --git a/pyproject.toml b/pyproject.toml
index 405ec2f..7dd4f3d 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -26,7 +26,6 @@ dependencies = [
"rioxarray",
"stac_static@git+https://github.com/jsignell/stac-static",
"stackstac",
- "rtree",
]
[project.optional-dependencies]
diff --git a/simplestac/extents.py b/simplestac/extents.py
index 379898d..f31094e 100644
--- a/simplestac/extents.py
+++ b/simplestac/extents.py
@@ -6,7 +6,6 @@ from dataclasses import dataclass
from datetime import datetime
from typing import Union
from stacflow.common_types import Bbox
-from rtree import index
@dataclass
@@ -65,21 +64,13 @@ def clusterize_bboxes(bboxes: list[Bbox]) -> list[Bbox]:
"""
bboxes = [SmartBbox(bbox) for bbox in bboxes]
- idx = index.Index()
- for i, bbox in enumerate(bboxes):
- idx.insert(id=i, coordinates=bbox.coords)
clusters = [bboxes.pop()]
- def _update_clusters_idx():
- _clusters_idx = index.Index()
- for _i, cluster_bbox in enumerate(clusters):
- _clusters_idx.insert(id=_i, coordinates=cluster_bbox.coords)
- return _clusters_idx
-
- clusters_idx = _update_clusters_idx()
while bboxes:
bbox = bboxes.pop()
- inter_clusters = list(clusters_idx.intersection(bbox.coords))
+ inter_clusters = [
+ i for i, cluster in enumerate(clusters) if bbox.touches(cluster)
+ ]
if inter_clusters:
# We merge all intersecting clusters into one
clusters[inter_clusters[0]].update(bbox)
@@ -93,8 +84,6 @@ def clusterize_bboxes(bboxes: list[Bbox]) -> list[Bbox]:
else:
clusters.append(bbox)
- clusters_idx = _update_clusters_idx()
-
return [cluster.coords for cluster in clusters]
diff --git a/tests/extents.py b/tests/extents.py
new file mode 100644
index 0000000..5b52296
--- /dev/null
+++ b/tests/extents.py
@@ -0,0 +1,20 @@
+import pytest
+from simplestac.extents import AutoSpatialExtent
+
+def test_spatial_extent():
+ """
+ Test the `AutoSpatialExtent` class.
+
+ Two clusters of bboxes (i.e. lists of bboxes) composed respectively with 2 and 1 bbox are
+ created (by definition, the clusters are disjoint: their bboxes don't overlap)
+ We instanciate an `AutoSpatialExtent` and we check that the two expected clusters are found.
+ """
+ # first cluster (e.g. "france mainland")
+ bbox1 = [4, 42, 6, 44]
+ bbox2 = [3, 41, 5, 43]
+
+ # second cluster (e.g. "corse")
+ bbox3 = [7, 42, 8, 50]
+
+ ase = AutoSpatialExtent(bboxes=[bbox1, bbox2, bbox3])
+ assert ase.bboxes == [[7, 42, 8, 50], [3, 41, 6, 44]]
--
GitLab
From 0e1a1948dae2a7b714a3f5eeb1e92a500cb08b59 Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Tue, 12 Mar 2024 15:47:49 +0100
Subject: [PATCH 17/26] fix harmonize_sen2cor_offset function name
---
simplestac/utils.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/simplestac/utils.py b/simplestac/utils.py
index b038582..bc17843 100644
--- a/simplestac/utils.py
+++ b/simplestac/utils.py
@@ -902,7 +902,7 @@ def apply_formula(x, formula):
return eval(formula)
-def harmonize_sen2cor_offet(x, bands=S2_SEN2COR_BANDS, inplace=False):
+def harmonize_sen2cor_offset(x, bands=S2_SEN2COR_BANDS, inplace=False):
"""
Harmonize new Sentinel-2 item collection (Sen2Cor v4+, 2022-01-25)
to the old baseline (v3-):
--
GitLab
From c5c2af828d8a328ab401fcb352e584f5cc3820e3 Mon Sep 17 00:00:00 2001
From: Cresson Remi <remi.cresson@irstea.fr>
Date: Tue, 12 Mar 2024 16:34:48 +0100
Subject: [PATCH 18/26] Update file extents.py
---
simplestac/extents.py | 10 ++++++++++
1 file changed, 10 insertions(+)
diff --git a/simplestac/extents.py b/simplestac/extents.py
index f31094e..cff0889 100644
--- a/simplestac/extents.py
+++ b/simplestac/extents.py
@@ -1,5 +1,15 @@
"""
Deal with STAC Extents.
+
+The `AutoSpatialExtent` is an extension of the `pystac.SpatialExtent`,
+it allows to automatically compute clusters of bboxes from a list of bboxes.
+It can be instanciated with the same arguments as `pystac.SpatialExtent`.
+Bounding boxes lists will be automatically processed: all touching bboxes
+will be merged and updated.
+
+The `AutoTemporalExtent` is an extension of the `pystac.TemporalExtent`.
+It simply computes the date min and date max of a set of dates or dates ranges.
+
"""
import pystac
from dataclasses import dataclass
--
GitLab
From d5bef12665ae2b959d0f0011cb6abb358cf33eb8 Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Tue, 12 Mar 2024 16:39:13 +0100
Subject: [PATCH 19/26] fix harmonize_sen2cor_offset
---
simplestac/utils.py | 21 ++++++++++++++++++---
1 file changed, 18 insertions(+), 3 deletions(-)
diff --git a/simplestac/utils.py b/simplestac/utils.py
index bc17843..9ed7813 100644
--- a/simplestac/utils.py
+++ b/simplestac/utils.py
@@ -571,6 +571,7 @@ def write_assets(x: Union[ItemCollection, pystac.Item],
remove_item_props=DEFAULT_REMOVE_PROPS,
overwrite=False,
progress=True,
+ writer_args=None,
**kwargs):
"""
Writes item(s) assets to the specified output directory.
@@ -591,9 +592,16 @@ def write_assets(x: Union[ItemCollection, pystac.Item],
If None, no properties are removed.
overwrite : bool, optional
Whether to overwrite existing files. Defaults to False.
+ writer_args : dict, optional
+ Arguments to pass to write_raster for each asset. Defaults to `None`.
+ See Notes for an example.
**kwargs
Additional keyword arguments passed to write_raster.
+ Returns
+ -------
+ ItemCollection
+ The item collection with the metadata updated with local asset paths.
"""
if isinstance(x, pystac.Item):
x = [x]
@@ -602,16 +610,23 @@ def write_assets(x: Union[ItemCollection, pystac.Item],
items = []
for item in tqdm(x, disable=not progress):
ic = ItemCollection([item], clone_items=True)
- arr = ic.to_xarray(bbox=bbox, xy_coords=xy_coords).squeeze("time")
+ arr = ic.to_xarray(bbox=bbox, xy_coords=xy_coords, ).squeeze("time")
item_dir = (output_dir / item.id).mkdir_p()
for b in arr.band.values:
filename = '_'.join([item.id, b+'.tif'])
file = item_dir / f"{filename}"
+
+ # Use specific writer args if available
+ if writer_args is not None and b in writer_args:
+ wa = writer_args[b]
+ else:
+ wa = kwargs
+
try:
if file.exists() and not overwrite:
logger.debug(f"File already exists, skipping asset: {file}")
else:
- write_raster(arr.sel(band=b), file, **kwargs)
+ write_raster(arr.sel(band=b), file, **wa)
# update stac asset info
stac_info = stac_asset_info_from_raster(file)
@@ -934,7 +949,7 @@ def harmonize_sen2cor_offset(x, bands=S2_SEN2COR_BANDS, inplace=False):
"""
if not inplace:
- x = x.copy()
+ x = x.clone()
for item in x:
for asset in bands:
if asset in item.assets:
--
GitLab
From 5df23b0816da0fb2078b46203fc71d9c4996254a Mon Sep 17 00:00:00 2001
From: Florian de Boissieu <fdeboiss@gmail.com>
Date: Tue, 12 Mar 2024 16:45:25 +0100
Subject: [PATCH 20/26] fix tests
---
tests/test_remote.py | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tests/test_remote.py b/tests/test_remote.py
index beb2b82..3333982 100644
--- a/tests/test_remote.py
+++ b/tests/test_remote.py
@@ -1,4 +1,4 @@
-from simplestac.utils import write_assets, ItemCollection, harmonize_sen2cor_offet
+from simplestac.utils import write_assets, ItemCollection, harmonize_sen2cor_offset
import planetary_computer as pc
import pystac_client
@@ -29,7 +29,7 @@ def test_offset_harmonization(roi, s2scene_pc_dir):
sortby="datetime",
)
col = search.item_collection()
- harmonize_sen2cor_offet(col, inplace=True)
+ harmonize_sen2cor_offset(col, inplace=True)
of0 = col[0].assets["B02"].extra_fields["raster:bands"][0]["offset"]
ofN = col[-1].assets["B02"].extra_fields["raster:bands"][0]["offset"]
--
GitLab
From 6f020803c1542b83e84279e10c5f5504c6d89239 Mon Sep 17 00:00:00 2001
From: Cresson Remi <remi.cresson@irstea.fr>
Date: Tue, 12 Mar 2024 22:19:00 +0100
Subject: [PATCH 21/26] Update file extents.py
---
simplestac/extents.py | 29 +++++++++++++++++++++++------
1 file changed, 23 insertions(+), 6 deletions(-)
diff --git a/simplestac/extents.py b/simplestac/extents.py
index cff0889..069e409 100644
--- a/simplestac/extents.py
+++ b/simplestac/extents.py
@@ -1,14 +1,31 @@
"""
Deal with STAC Extents.
-The `AutoSpatialExtent` is an extension of the `pystac.SpatialExtent`,
-it allows to automatically compute clusters of bboxes from a list of bboxes.
-It can be instanciated with the same arguments as `pystac.SpatialExtent`.
-Bounding boxes lists will be automatically processed: all touching bboxes
-will be merged and updated.
+# AutoSpatialExtent
+
+The vanilla `pystac.SpatialExtent` enables to describe a spatial extent
+from several bounding boxes. While this is useful, sometimes we want to
+merge together bounding boxes that are partially overlapping. For instance,
+we can take the example of the France mainland, covered by multiple remote
+sensing products that are generally partially overlapping, and the Corse
+island that is also covered by a number of RS products, but spatially
+disjoint from the France mainland RS products bounding boxes. In this
+particular exemple, we would like to regroup all RS products bounding
+boxes so that there is one bbox for the France mainland, and another
+bbox for the Corse island. This is particularly useful when a STAC
+collection covers sparsely a broad area (e.g. worldwide), with several
+isolated regions.
+
+The `AutoSpatialExtent` is an extension of the `pystac.SpatialExtent`.
+
+Instances are initialized with the same arguments as `pystac.SpatialExtent`.
+Internally, bounding boxes lists are processed at initialisation, so all
+partially overlapping bounding boxes are merged and updated as a single one.
+
+# AutoTemporalExtent
The `AutoTemporalExtent` is an extension of the `pystac.TemporalExtent`.
-It simply computes the date min and date max of a set of dates or dates ranges.
+It computes the date min and date max of a set of dates or dates ranges.
"""
import pystac
--
GitLab
From dc6db14c1d6b9d92d6b8160cae5d9561e684a483 Mon Sep 17 00:00:00 2001
From: Cresson Remi <remi.cresson@irstea.fr>
Date: Tue, 12 Mar 2024 23:13:05 +0100
Subject: [PATCH 22/26] Update file extents.py
---
simplestac/extents.py | 25 ++++++++++++++++++++++++-
1 file changed, 24 insertions(+), 1 deletion(-)
diff --git a/simplestac/extents.py b/simplestac/extents.py
index 069e409..d49f1e1 100644
--- a/simplestac/extents.py
+++ b/simplestac/extents.py
@@ -32,7 +32,30 @@ import pystac
from dataclasses import dataclass
from datetime import datetime
from typing import Union
-from stacflow.common_types import Bbox
+from numbers import Number
+from annotated_types import Predicate
+from typing_extensions import Annotated
+
+
+def is_bbox(l: List) -> bool:
+ """
+ Predicate to test is the input list represents a bounding box in WGS84.
+
+ Args:
+ l: a list of `Number`
+
+ Returns:
+
+ """
+ if len(l) == 4:
+ if all(isinstance(i, Number) for i in l):
+ if -180 <= l[0] and -90 <= l[1] and l[2] <= 180 and l[3] <= 90:
+ if l[0] <= l[2] and l[1] <= l[3]:
+ return True
+ return False
+
+
+Bbox = Annotated[list, Predicate(is_bbox)]
@dataclass
--
GitLab
From 602d2e01a0e67d6e17d07b195ae86fa6d70de805 Mon Sep 17 00:00:00 2001
From: Cresson Remi <remi.cresson@irstea.fr>
Date: Wed, 13 Mar 2024 08:13:45 +0100
Subject: [PATCH 23/26] Update 2 files
- /tests/extents.py
- /tests/test_extents.py
---
tests/{extents.py => test_extents.py} | 0
1 file changed, 0 insertions(+), 0 deletions(-)
rename tests/{extents.py => test_extents.py} (100%)
diff --git a/tests/extents.py b/tests/test_extents.py
similarity index 100%
rename from tests/extents.py
rename to tests/test_extents.py
--
GitLab
From 210b1d4db827bfa7f0068e24f8539c46544d8180 Mon Sep 17 00:00:00 2001
From: Cresson Remi <remi.cresson@irstea.fr>
Date: Wed, 13 Mar 2024 08:18:19 +0100
Subject: [PATCH 24/26] Update file extents.py
---
simplestac/extents.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/simplestac/extents.py b/simplestac/extents.py
index d49f1e1..77e9c79 100644
--- a/simplestac/extents.py
+++ b/simplestac/extents.py
@@ -31,7 +31,7 @@ It computes the date min and date max of a set of dates or dates ranges.
import pystac
from dataclasses import dataclass
from datetime import datetime
-from typing import Union
+from typing import Union, List
from numbers import Number
from annotated_types import Predicate
from typing_extensions import Annotated
--
GitLab
From 8a918fd58c746f462b44b09a5f83e52abaac6ef7 Mon Sep 17 00:00:00 2001
From: Cresson Remi <remi.cresson@irstea.fr>
Date: Wed, 13 Mar 2024 09:30:01 +0100
Subject: [PATCH 25/26] Update 2 files
- /tests/test_extents.py
- /simplestac/extents.py
---
simplestac/extents.py | 17 ++++++++++-------
tests/test_extents.py | 40 ++++++++++++++++++++++++++++++++++++++++
2 files changed, 50 insertions(+), 7 deletions(-)
diff --git a/simplestac/extents.py b/simplestac/extents.py
index 77e9c79..ef69713 100644
--- a/simplestac/extents.py
+++ b/simplestac/extents.py
@@ -209,12 +209,15 @@ class AutoTemporalExtent(pystac.TemporalExtent):
def make_single_interval(self):
all_dates = []
- for interval in self.intervals:
- if isinstance(interval, (list, tuple)):
- all_dates += [i for i in interval if i is not None]
- elif isinstance(interval, datetime):
- all_dates.append(interval)
- else:
- TypeError(f"Unsupported date/range of: {interval}")
+ for dates_or_intervals in self.intervals:
+ # Because base class (`pystac.SpatialExtent`) converts everything
+ # into [[...]]
+ for date_or_interval in dates_or_intervals:
+ if isinstance(date_or_interval, (list, tuple)):
+ all_dates += [i for i in date_or_interval if i is not None]
+ elif isinstance(date_or_interval, datetime):
+ all_dates.append(date_or_interval)
+ else:
+ TypeError(f"Unsupported date/range of: {date_or_interval}")
self.intervals = \
[[min(all_dates), max(all_dates)]] if all_dates else [None, None]
diff --git a/tests/test_extents.py b/tests/test_extents.py
index 5b52296..db93b23 100644
--- a/tests/test_extents.py
+++ b/tests/test_extents.py
@@ -1,6 +1,8 @@
import pytest
from simplestac.extents import AutoSpatialExtent
+from datetime import datetime
+
def test_spatial_extent():
"""
Test the `AutoSpatialExtent` class.
@@ -8,6 +10,7 @@ def test_spatial_extent():
Two clusters of bboxes (i.e. lists of bboxes) composed respectively with 2 and 1 bbox are
created (by definition, the clusters are disjoint: their bboxes don't overlap)
We instanciate an `AutoSpatialExtent` and we check that the two expected clusters are found.
+
"""
# first cluster (e.g. "france mainland")
bbox1 = [4, 42, 6, 44]
@@ -18,3 +21,40 @@ def test_spatial_extent():
ase = AutoSpatialExtent(bboxes=[bbox1, bbox2, bbox3])
assert ase.bboxes == [[7, 42, 8, 50], [3, 41, 6, 44]]
+
+def test_temporal_extent():
+ """
+ Test the `AutoTemporalExtent`.
+
+ """
+ # dates only (as plain list)
+ dates = [
+ datetime(year=2020, month=1, day=1),
+ datetime(year=2022, month=1, day=1),
+ datetime(year=2023, month=1, day=1),
+ ]
+ auto_text = AutoTemporalExtent(dates)
+ assert auto_text.intervals == [[
+ datetime(year=2020, month=1, day=1),
+ datetime(year=2023, month=1, day=1)
+ ]]
+
+ # dates only (as nested list)
+ auto_text = AutoTemporalExtent([dates])
+ assert auto_text.intervals == [[
+ datetime(year=2020, month=1, day=1),
+ datetime(year=2023, month=1, day=1)
+ ]]
+
+ # mixed dates + ranges
+ auto_text = AutoTemporalExtent([
+ datetime(year=2020, month=1, day=1),
+ [None, None],
+ [datetime(year=2019, month=1, day=1), None],
+ [None, datetime(year=2024, month=1, day=1)],
+ datetime(year=2023, month=1, day=1)
+ ])
+ assert auto_text.intervals == [[
+ datetime(year=2019, month=1, day=1),
+ datetime(year=2024, month=1, day=1)
+ ]]
--
GitLab
From 75e084c989eacf25543a8b16b340d44244b4d7f7 Mon Sep 17 00:00:00 2001
From: Cresson Remi <remi.cresson@irstea.fr>
Date: Wed, 13 Mar 2024 09:37:13 +0100
Subject: [PATCH 26/26] Update file test_extents.py
---
tests/test_extents.py | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/tests/test_extents.py b/tests/test_extents.py
index db93b23..f5d8fb2 100644
--- a/tests/test_extents.py
+++ b/tests/test_extents.py
@@ -1,5 +1,5 @@
import pytest
-from simplestac.extents import AutoSpatialExtent
+from simplestac.extents import AutoSpatialExtent, AutoTemporalExtent
from datetime import datetime
--
GitLab