cecil 0.0.31__tar.gz → 0.0.35__tar.gz

{cecil-0.0.31 → cecil-0.0.35}/.gitignore

@@ -9,3 +9,4 @@ __pycache__
 dist
 tmp
 venv
+.venv

cecil-0.0.35/CONTRIBUTING.md

@@ -0,0 +1,23 @@
+## Development installation
+
+Install packaging/distribution tools and linter:
+
+```shell
+pip install hatch twine black
+```
+
+From top-level repo directory, install the package in editable mode:
+
+```shell
+pip install -e .
+```
+
+Local edits to the package will immediately take effect.
+
+Get the PyPI Test API Key from 1Password and add it to `~/.pypirc`:
+
+```bash
+[testpypi]
+  username = __token__
+  password = <PyPI Test API Key>
+```

cecil-0.0.35/PKG-INFO

@@ -0,0 +1,24 @@
+Metadata-Version: 2.4
+Name: cecil
+Version: 0.0.35
+Summary: Python SDK for Cecil Earth
+License-Expression: MIT
+License-File: LICENSE.txt
+Classifier: Development Status :: 4 - Beta
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Requires-Python: >=3.10
+Requires-Dist: dask==2025.11.0
+Requires-Dist: pydantic<3.0.0,>=2.11.9
+Requires-Dist: requests<3.0.0,>=2.32.5
+Requires-Dist: rioxarray==0.19.0
+Requires-Dist: snowflake-connector-python[pandas]<4.0.0,>=3.17.4
+Requires-Dist: xarray==2025.11.0
+Description-Content-Type: text/markdown
+
+# Cecil SDK
+
+Please refer to the Cecil documentation:
+
+https://docs.cecil.earth

cecil-0.0.35/README.md

@@ -0,0 +1,5 @@
+# Cecil SDK
+
+Please refer to the Cecil documentation:
+
+https://docs.cecil.earth

{cecil-0.0.31 → cecil-0.0.35}/pyproject.toml

@@ -16,12 +16,12 @@ classifiers = [
     "Operating System :: OS Independent",
 ]
 dependencies = [
-    "dask==2025.9.1",
+    "dask==2025.11.0",
     "pydantic>=2.11.9,<3.0.0",
     "requests>=2.32.5,<3.0.0",
     "rioxarray==0.19.0",
     "snowflake-connector-python[pandas]>=3.17.4,<4.0.0",
-    "xarray==2025.6.1"
+    "xarray==2025.11.0"
 ]
 
 [tool.hatch.version]

{cecil-0.0.31 → cecil-0.0.35}/src/cecil/client.py

@@ -104,14 +104,14 @@ class Client:
         return [DataRequest(**record) for record in res["records"]]
 
     def list_subscriptions(self) -> List[Subscription]:
-        res = self._get(url="/v0/data-requests")
+        res = self._get(url="/v0/subscriptions")
         return [Subscription(**record) for record in res["records"]]
 
     def create_subscription(
         self, aoi_id: str, dataset_id: str, external_ref: Optional[str] = None
     ) -> Subscription:
         res = self._post(
-            url="/v0/data-requests",
+            url="/v0/subscriptions",
             model=SubscriptionCreate(
                 aoi_id=aoi_id, dataset_id=dataset_id, external_ref=external_ref
             ),
@@ -120,7 +120,7 @@ class Client:
         return Subscription(**res)
 
     def get_subscription(self, id: str) -> Subscription:
-        res = self._get(url=f"/v0/data-requests/{id}")
+        res = self._get(url=f"/v0/subscriptions/{id}")
         return Subscription(**res)
 
     def load_xarray(
@@ -145,7 +145,7 @@ class Client:
             subscription_id = data_request_id
 
         res = SubscriptionMetadata(
-            **self._get(url=f"/v0/data-requests/{subscription_id}/metadata")
+            **self._get(url=f"/v0/subscriptions/{subscription_id}/metadata")
         )
         return load_xarray(res)
 
@@ -171,7 +171,7 @@ class Client:
             subscription_id = data_request_id
 
         res = SubscriptionListFiles(
-            **self._get(url=f"/v0/data-requests/{subscription_id}/files/tiff")
+            **self._get(url=f"/v0/subscriptions/{subscription_id}/files/tiff")
         )
         return load_xarray_v2(res)
 
@@ -197,7 +197,7 @@ class Client:
             subscription_id = data_request_id
 
         res = SubscriptionParquetFiles(
-            **self._get(url=f"/v0/data-requests/{subscription_id}/parquet-files")
+            **self._get(url=f"/v0/subscriptions/{subscription_id}/parquet-files")
         )
         df = pd.concat((pd.read_parquet(f) for f in res.files))
         return df[
@@ -308,12 +308,12 @@ class Client:
     def update_organisation_settings(
         self,
         *,
-        monthly_data_request_limit,
+        monthly_subscription_limit,
     ) -> OrganisationSettings:
         res = self._post(
             url="/v0/organisation/settings",
             model=OrganisationSettings(
-                monthly_data_request_limit=monthly_data_request_limit,
+                monthly_subscription_limit=monthly_subscription_limit,
             ),
         )
         return OrganisationSettings(**res)

{cecil-0.0.31 → cecil-0.0.35}/src/cecil/models.py

@@ -1,7 +1,7 @@
 import datetime
 from typing import Dict, Optional, List
 
-from pydantic import BaseModel, ConfigDict, SecretStr
+from pydantic import BaseModel, ConfigDict, Field, SecretStr
 from pydantic.alias_generators import to_camel
 
 
@@ -49,7 +49,9 @@ class DataRequestCreate(BaseModel):
 
 class OrganisationSettings(BaseModel):
     model_config = ConfigDict(alias_generator=to_camel, populate_by_name=True)
-    monthly_data_request_limit: Optional[int] = None
+    monthly_subscription_limit: Optional[int] = Field(
+        alias="monthlyDataRequestLimit",
+    )
 
 
 class RecoverAPIKey(BaseModel):

cecil-0.0.35/src/cecil/version.py

@@ -0,0 +1 @@
+__version__ = "0.0.35"

cecil-0.0.35/src/cecil/xarray.py

@@ -0,0 +1,193 @@
+import re
+import time
+from datetime import datetime
+
+import boto3
+import dask
+import rasterio
+import rasterio.session
+import rioxarray
+import xarray
+
+from .models import SubscriptionMetadata, SubscriptionListFiles
+
+# v1
+
+
+def load_xarray(metadata: SubscriptionMetadata) -> xarray.Dataset:
+    data_vars = {}
+
+    for f in metadata.files:
+        try:
+            dataset = _retry_with_exponential_backoff(_load_file, 5, 1, 2, f.url)
+        except Exception as e:
+            raise ValueError(f"failed to load file: {e}")
+
+        for b in f.bands:
+            band = dataset.sel(band=b.number, drop=True)
+
+            if b.time and b.time_pattern:
+                t = datetime.strptime(b.time, b.time_pattern)
+                band = band.expand_dims("time")
+                band = band.assign_coords(time=[t])
+
+            band.name = b.variable_name
+
+            if b.variable_name not in data_vars:
+                data_vars[b.variable_name] = []
+
+            data_vars[b.variable_name].append(band)
+
+    for variable_name, time_series in data_vars.items():
+        if "time" in time_series[0].dims:
+            data_vars[variable_name] = xarray.concat(
+                time_series, dim="time", join="exact"
+            )
+        else:
+            data_vars[variable_name] = time_series[0]
+
+    return xarray.Dataset(
+        data_vars=data_vars,
+        attrs={
+            "provider_name": metadata.provider_name,
+            "dataset_name": metadata.dataset_name,
+            "dataset_id": metadata.dataset_id,
+            "aoi_id": metadata.aoi_id,
+            "subscription_id": metadata.data_request_id,
+        },
+    )
+
+
+def _retry_with_exponential_backoff(
+    func, retries, start_delay, multiplier, *args, **kwargs
+):
+    delay = start_delay
+    for attempt in range(1, retries + 1):
+        try:
+            return func(*args, **kwargs)
+        except Exception as e:
+            if attempt == retries:
+                raise e
+            time.sleep(delay)
+            delay *= multiplier
+    return None
+
+
+def _load_file(url: str):
+    return rioxarray.open_rasterio(
+        url,
+        chunks={"x": 2000, "y": 2000},
+    )
+
+
+# v2
+
+
+def load_xarray_v2(res: SubscriptionListFiles) -> xarray.Dataset:
+    session = boto3.session.Session(
+        aws_access_key_id=res.credentials.access_key_id,
+        aws_secret_access_key=res.credentials.secret_access_key,
+        aws_session_token=res.credentials.session_token,
+    )
+
+    keys = _list_keys_v2(session, res.bucket.name, res.bucket.prefix)
+
+    if not keys:
+        return xarray.Dataset()
+
+    timestamp_pattern = re.compile(r"\d{4}/\d{2}/\d{2}/\d{2}/\d{2}/\d{2}")
+    data_vars = {}
+
+    with rasterio.env.Env(
+        session=rasterio.session.AWSSession(session),
+    ):
+        first_file = rioxarray.open_rasterio(
+            f"s3://{res.bucket.name}/{keys[0]}", chunks="auto"
+        )
+
+    for key in keys:
+        filename = key.split("/")[-1]
+
+        file_info = res.file_mapping.get(filename)
+        if not file_info:
+            continue
+
+        timestamp_str = timestamp_pattern.search(key).group()
+
+        for band_num, var_name in enumerate(file_info.bands, start=1):
+            lazy_array = dask.array.from_delayed(
+                dask.delayed(_load_file_v2)(
+                    session, f"s3://{res.bucket.name}/{key}", band_num
+                ),
+                shape=(
+                    first_file.rio.height,
+                    first_file.rio.width,
+                ),
+                dtype=file_info.type,
+            )
+            band_da = xarray.DataArray(
+                lazy_array,
+                dims=("y", "x"),
+                coords={
+                    "y": first_file.y.values,
+                    "x": first_file.x.values,
+                },
+                # attrs=first_file.attrs.copy() # TODO: is it the same for all files?
+            )
+            # band_da.encoding = first_file.encoding.copy() # TODO: is it the same for all files?
+            band_da.rio.write_crs(first_file.rio.crs, inplace=True)
+            band_da.rio.write_transform(first_file.rio.transform(), inplace=True)
+
+            band_da.name = var_name
+
+            # Dataset with time dimension
+            if timestamp_str != "0000/00/00/00/00/00":
+                t = datetime.strptime(timestamp_str, "%Y/%m/%d/%H/%M/%S")
+                band_da = band_da.expand_dims("time")
+                band_da = band_da.assign_coords(time=[t])
+
+            if var_name not in data_vars:
+                data_vars[var_name] = []
+
+            data_vars[var_name].append(band_da)
+
+    for var_name, time_series in data_vars.items():
+        if "time" in time_series[0].dims:
+            data_vars[var_name] = xarray.concat(time_series, dim="time", join="exact")
+        else:
+            data_vars[var_name] = time_series[0]
+
+    return xarray.Dataset(
+        data_vars=data_vars,
+        attrs={
+            "provider_name": res.provider_name,
+            "dataset_name": res.dataset_name,
+            "dataset_id": res.dataset_id,
+            "aoi_id": res.aoi_id,
+            "subscription_id": res.data_request_id,
+        },
+    )
+
+
+def _load_file_v2(aws_session: boto3.session.Session, url: str, band_num: int):
+    with rasterio.env.Env(
+        session=rasterio.session.AWSSession(aws_session),
+    ):
+        with rasterio.open(url) as src:
+            return src.read(band_num)
+
+
+def _list_keys_v2(session: boto3.session.Session, bucket_name, prefix) -> list[str]:
+    s3_client = session.client("s3")
+    paginator = s3_client.get_paginator("list_objects_v2")
+    page_iterator = paginator.paginate(
+        Bucket=bucket_name,
+        Prefix=prefix,
+    )
+
+    keys = []
+    for page in page_iterator:
+        for obj in page.get("Contents", []):
+            keys.append(obj["Key"])
+
+    return keys

cecil-0.0.31/CONTRIBUTING.md

@@ -1,21 +0,0 @@
-## Development installation
-
-Install packaging/distribution tools:
-
-```shell
-pip install hatch twine
-```
-
-Install linter
-
-```shell
-pip install black
-```
-
-From top-level repo directory, install the package in editable mode:
-
-```shell
-pip install -e .
-```
-
-Local edits to the package will immediately take effect.

cecil-0.0.31/PKG-INFO

@@ -1,122 +0,0 @@
-Metadata-Version: 2.4
-Name: cecil
-Version: 0.0.31
-Summary: Python SDK for Cecil Earth
-License-Expression: MIT
-License-File: LICENSE.txt
-Classifier: Development Status :: 4 - Beta
-Classifier: License :: OSI Approved :: MIT License
-Classifier: Operating System :: OS Independent
-Classifier: Programming Language :: Python :: 3
-Requires-Python: >=3.10
-Requires-Dist: dask==2025.9.1
-Requires-Dist: pydantic<3.0.0,>=2.11.9
-Requires-Dist: requests<3.0.0,>=2.32.5
-Requires-Dist: rioxarray==0.19.0
-Requires-Dist: snowflake-connector-python[pandas]<4.0.0,>=3.17.4
-Requires-Dist: xarray==2025.6.1
-Description-Content-Type: text/markdown
-
-# Cecil SDK
-
-[![PyPI - Version](https://img.shields.io/pypi/v/cecil-sdk.svg)](https://pypi.org/project/cecil-sdk)
-[![PyPI - Python Version](https://img.shields.io/pypi/pyversions/cecil-sdk.svg)](https://pypi.org/project/cecil-sdk)
-
------
-
-## Table of Contents
-
-- [Installation](#installation)
-- [Authentication](#authentication)
-- [License](#license)
-- [Examples](#examples)
-
-## Installation
-
-```shell
-pip install cecil
-```
-
-## Authentication
-
-Set `CECIL_API_KEY` environment variable to your Cecil API key.
-
-## Examples
-
-### Create an AOI and data request using the Cecil client
-
-```python
-import cecil
-
-client = cecil.Client()
-
-my_aoi = client.create_aoi(
-    name="My AOI",
-    geometry={
-        "type": "Polygon",
-        "coordinates": [
-            [
-                [145.410408835, -42.004083838],
-                [145.410408835, -42.004203978],
-                [145.410623191, -42.004203978],
-                [145.410623191, -42.004083838],
-                [145.410408835, -42.004083838],
-            ]
-        ],
-    },
-)
-
-# Get dataset ID from docs.cecil.earth -> Datasets
-planet_forest_carbon_diligence_id = "c2dd4f55-56f6-4d05-aae3-ba7c1dcd812f"
-
-my_data_request = client.create_data_request(
-    aoi_id=my_aoi.id,
-    dataset_id=planet_forest_carbon_diligence_id,
-)
-
-print(client.get_data_request(my_data_request.id))
-```
-
-### Create a transformation using the Cecil client
-
-```python
-my_transformation = client.create_transformation(
-    data_request_id=my_data_request.id,
-    crs="EPSG:4326",
-    spatial_resolution=0.005,
-)
-
-print(client.get_transformation(my_transformation.id))
-```
-
-### Query data (once transformation is completed)
-
-```python
-df = client.query(f'''
-    SELECT *
-    FROM
-        planet.forest_carbon_diligence
-    WHERE
-        transformation_id = '{my_transformation.id}'
-''')
-```
-
-### Other client methods:
-
-```python
-client.list_aois()
-
-client.get_aoi(my_aoi.id)
-
-client.list_data_requests()
-
-client.get_data_request(my_data_request.id)
-
-client.list_transformations()
-
-client.get_transformation(my_transformation.id)
-```
-
-## License
-
-`cecil` is distributed under the terms of the [MIT](https://spdx.org/licenses/MIT.html) license.

cecil-0.0.31/README.md

@@ -1,103 +0,0 @@
-# Cecil SDK
-
-[![PyPI - Version](https://img.shields.io/pypi/v/cecil-sdk.svg)](https://pypi.org/project/cecil-sdk)
-[![PyPI - Python Version](https://img.shields.io/pypi/pyversions/cecil-sdk.svg)](https://pypi.org/project/cecil-sdk)
-
------
-
-## Table of Contents
-
-- [Installation](#installation)
-- [Authentication](#authentication)
-- [License](#license)
-- [Examples](#examples)
-
-## Installation
-
-```shell
-pip install cecil
-```
-
-## Authentication
-
-Set `CECIL_API_KEY` environment variable to your Cecil API key.
-
-## Examples
-
-### Create an AOI and data request using the Cecil client
-
-```python
-import cecil
-
-client = cecil.Client()
-
-my_aoi = client.create_aoi(
-    name="My AOI",
-    geometry={
-        "type": "Polygon",
-        "coordinates": [
-            [
-                [145.410408835, -42.004083838],
-                [145.410408835, -42.004203978],
-                [145.410623191, -42.004203978],
-                [145.410623191, -42.004083838],
-                [145.410408835, -42.004083838],
-            ]
-        ],
-    },
-)
-
-# Get dataset ID from docs.cecil.earth -> Datasets
-planet_forest_carbon_diligence_id = "c2dd4f55-56f6-4d05-aae3-ba7c1dcd812f"
-
-my_data_request = client.create_data_request(
-    aoi_id=my_aoi.id,
-    dataset_id=planet_forest_carbon_diligence_id,
-)
-
-print(client.get_data_request(my_data_request.id))
-```
-
-### Create a transformation using the Cecil client
-
-```python
-my_transformation = client.create_transformation(
-    data_request_id=my_data_request.id,
-    crs="EPSG:4326",
-    spatial_resolution=0.005,
-)
-
-print(client.get_transformation(my_transformation.id))
-```
-
-### Query data (once transformation is completed)
-
-```python
-df = client.query(f'''
-    SELECT *
-    FROM
-        planet.forest_carbon_diligence
-    WHERE
-        transformation_id = '{my_transformation.id}'
-''')
-```
-
-### Other client methods:
-
-```python
-client.list_aois()
-
-client.get_aoi(my_aoi.id)
-
-client.list_data_requests()
-
-client.get_data_request(my_data_request.id)
-
-client.list_transformations()
-
-client.get_transformation(my_transformation.id)
-```
-
-## License
-
-`cecil` is distributed under the terms of the [MIT](https://spdx.org/licenses/MIT.html) license.

cecil-0.0.31/src/cecil/version.py

@@ -1 +0,0 @@
-__version__ = "0.0.31"

cecil-0.0.31/src/cecil/xarray.py

@@ -1,415 +0,0 @@
-import re
-import time
-from datetime import datetime
-
-import boto3
-import dask
-import rasterio
-import rasterio.session
-import rioxarray
-import xarray
-
-from .errors import Error
-from .models import SubscriptionMetadata, SubscriptionListFiles
-
-# v1
-
-
-def load_xarray(metadata: SubscriptionMetadata) -> xarray.Dataset:
-    data_vars = {}
-
-    for f in metadata.files:
-        try:
-            dataset = _retry_with_exponential_backoff(_load_file, 5, 1, 2, f.url)
-        except Exception as e:
-            raise ValueError(f"failed to load file: {e}")
-
-        for b in f.bands:
-            band = dataset.sel(band=b.number, drop=True)
-
-            if b.time and b.time_pattern:
-                t = datetime.strptime(b.time, b.time_pattern)
-                band = band.expand_dims("time")
-                band = band.assign_coords(time=[t])
-
-            band.name = b.variable_name
-
-            if b.variable_name not in data_vars:
-                data_vars[b.variable_name] = []
-
-            data_vars[b.variable_name].append(band)
-
-    for variable_name, time_series in data_vars.items():
-        if "time" in time_series[0].dims:
-            data_vars[variable_name] = xarray.concat(
-                time_series, dim="time", join="exact"
-            )
-        else:
-            data_vars[variable_name] = time_series[0]
-
-    return xarray.Dataset(
-        data_vars=data_vars,
-        attrs={
-            "provider_name": metadata.provider_name,
-            "dataset_name": metadata.dataset_name,
-            "dataset_id": metadata.dataset_id,
-            "aoi_id": metadata.aoi_id,
-            "subscription_id": metadata.data_request_id,
-        },
-    )
-
-
-def _retry_with_exponential_backoff(
-    func, retries, start_delay, multiplier, *args, **kwargs
-):
-    delay = start_delay
-    for attempt in range(1, retries + 1):
-        try:
-            return func(*args, **kwargs)
-        except Exception as e:
-            if attempt == retries:
-                raise e
-            time.sleep(delay)
-            delay *= multiplier
-    return None
-
-
-def _load_file(url: str):
-    return rioxarray.open_rasterio(
-        url,
-        chunks={"x": 2000, "y": 2000},
-    )
-
-
-# v2
-
-
-def load_xarray_v2(res: SubscriptionListFiles) -> xarray.Dataset:
-    session = boto3.session.Session(
-        aws_access_key_id=res.credentials.access_key_id,
-        aws_secret_access_key=res.credentials.secret_access_key,
-        aws_session_token=res.credentials.session_token,
-    )
-
-    keys = _list_keys_v2(session, res.bucket.name, res.bucket.prefix)
-
-    if not keys:
-        return xarray.Dataset()
-
-    timestamp_pattern = re.compile(r"\d{4}/\d{2}/\d{2}/\d{2}/\d{2}/\d{2}")
-    data_vars = {}
-
-    for key in keys:
-        try:
-            file_da = _retry_with_exponential_backoff(
-                _load_file_v2,
-                5,
-                1,
-                2,
-                session,
-                f"s3://{res.bucket.name}/{key}",
-            )
-        except Exception as e:
-            raise ValueError(f"failed to load file: {e}")
-
-        filename = key.split("/")[-1]
-
-        file_info = res.file_mapping.get(filename)
-        if not file_info:
-            continue
-
-        timestamp_str = timestamp_pattern.search(key).group()
-
-        for band_num, var_name in enumerate(file_info.bands, start=1):
-            band_da = file_da.sel(band=band_num, drop=True)
-            band_da.name = var_name
-
-            # Dataset with time dimension
-            if timestamp_str != "0000/00/00/00/00/00":
-                t = datetime.strptime(timestamp_str, "%Y/%m/%d/%H/%M/%S")
-                band_da = band_da.expand_dims("time")
-                band_da = band_da.assign_coords(time=[t])
-
-            if var_name not in data_vars:
-                data_vars[var_name] = []
-
-            data_vars[var_name].append(band_da)
-
-    for var_name, time_series in data_vars.items():
-        if "time" in time_series[0].dims:
-            data_vars[var_name] = xarray.concat(time_series, dim="time", join="exact")
-        else:
-            data_vars[var_name] = time_series[0]
-
-    return xarray.Dataset(
-        data_vars=data_vars,
-        attrs={
-            "provider_name": res.provider_name,
-            "dataset_name": res.dataset_name,
-            "dataset_id": res.dataset_id,
-            "aoi_id": res.aoi_id,
-            "subscription_id": res.data_request_id,
-        },
-    )
-
-
-def _list_keys_v2(session: boto3.session.Session, bucket_name, prefix) -> list[str]:
-    s3_client = session.client("s3")
-    paginator = s3_client.get_paginator("list_objects_v2")
-    page_iterator = paginator.paginate(
-        Bucket=bucket_name,
-        Prefix=prefix,
-    )
-
-    keys = []
-    for page in page_iterator:
-        for obj in page.get("Contents", []):
-            keys.append(obj["Key"])
-
-    return keys
-
-
-def _load_file_v2(aws_session: boto3.session.Session, url: str):
-    with rasterio.env.Env(
-        session=rasterio.session.AWSSession(aws_session),
-        GDAL_DISABLE_READDIR_ON_OPEN=True,
-    ):
-        return rioxarray.open_rasterio(
-            url,
-            chunks={"x": 2000, "y": 2000},
-        )
-
-
-# v3
-
-
-def load_xarray_v3(res: SubscriptionListFiles) -> xarray.Dataset:
-    session = boto3.session.Session(
-        aws_access_key_id=res.credentials.access_key_id,
-        aws_secret_access_key=res.credentials.secret_access_key,
-        aws_session_token=res.credentials.session_token,
-    )
-
-    keys = _list_keys_v3(session, res.bucket.name, res.bucket.prefix)
-
-    if not keys:
-        return xarray.Dataset()
-
-    timestamp_pattern = re.compile(r"\d{4}/\d{2}/\d{2}/\d{2}/\d{2}/\d{2}")
-    data_vars = {}
-
-    with rasterio.env.Env(
-        session=rasterio.session.AWSSession(session),
-        GDAL_DISABLE_READDIR_ON_OPEN=True,
-    ):
-        first_file = rioxarray.open_rasterio(
-            f"s3://{res.bucket.name}/{keys[0]}", chunks="auto"
-        )
-
-    for key in keys:
-        filename = key.split("/")[-1]
-
-        file_info = res.file_mapping.get(filename)
-        if not file_info:
-            continue
-
-        lazy_array = dask.array.from_delayed(
-            dask.delayed(_load_file_v3)(session, f"s3://{res.bucket.name}/{key}"),
-            shape=first_file.shape,
-            dtype=file_info.type,
-        )
-        lazy_da = xarray.DataArray(
-            lazy_array,
-            dims=first_file.dims,
-            coords=dict(first_file.coords),
-            # attrs=first_file.attrs.copy() # TODO: not the same for all files
-        )
-        # lazy_da.encoding = first_file.encoding.copy()
-        # lazy_da.rio.write_crs(first_file.rio.crs, inplace=True)
-        # lazy_da.rio.write_transform(first_file.rio.transform(), inplace=True)
-
-        timestamp_str = timestamp_pattern.search(key).group()
-
-        for band_num, var_name in enumerate(file_info.bands, start=1):
-            band_da = lazy_da.sel(band=band_num, drop=True)
-            band_da.name = var_name
-
-            # Dataset with time dimension
-            if timestamp_str != "0000/00/00/00/00/00":
-                t = datetime.strptime(timestamp_str, "%Y/%m/%d/%H/%M/%S")
-                band_da = band_da.expand_dims("time")
-                band_da = band_da.assign_coords(time=[t])
-
-            if var_name not in data_vars:
-                data_vars[var_name] = []
-
-            data_vars[var_name].append(band_da)
-
-    for var_name, time_series in data_vars.items():
-        if "time" in time_series[0].dims:
-            data_vars[var_name] = xarray.concat(time_series, dim="time", join="exact")
-        else:
-            data_vars[var_name] = time_series[0]
-
-    return xarray.Dataset(
-        data_vars=data_vars,
-        attrs={
-            "provider_name": res.provider_name,
-            "dataset_name": res.dataset_name,
-            "dataset_id": res.dataset_id,
-            "aoi_id": res.aoi_id,
-            "subscription_id": res.data_request_id,
-        },
-    )
-
-
-def _load_file_v3(aws_session: boto3.session.Session, url: str):
-    with rasterio.env.Env(
-        session=rasterio.session.AWSSession(aws_session),
-        GDAL_DISABLE_READDIR_ON_OPEN=True,
-    ):
-        return rioxarray.open_rasterio(
-            url,
-            chunks="auto",
-        ).values
-        # ).sel(band=num_band, drop=True)
-        # ).sel(band=num_band, drop=True).values
-        # ).isel(band=num_band-1).values
-
-
-def _list_keys_v3(session: boto3.session.Session, bucket_name, prefix) -> list[str]:
-    s3_client = session.client("s3")
-    paginator = s3_client.get_paginator("list_objects_v2")
-    page_iterator = paginator.paginate(
-        Bucket=bucket_name,
-        Prefix=prefix,
-    )
-
-    keys = []
-    for page in page_iterator:
-        for obj in page.get("Contents", []):
-            keys.append(obj["Key"])
-
-    return keys
-
-
-# v4
-
-
-def load_xarray_v4(res: SubscriptionListFiles) -> xarray.Dataset:
-
-    session = boto3.session.Session(
-        aws_access_key_id=res.credentials.access_key_id,
-        aws_secret_access_key=res.credentials.secret_access_key,
-        aws_session_token=res.credentials.session_token,
-    )
-
-    keys = _list_keys_v2(session, res.bucket.name, res.bucket.prefix)
-
-    if not keys:
-        return xarray.Dataset()
-
-    first_file_metadata = _get_file_metadata_v4(session, res.bucket.name, keys[0])
-
-    timestamp_pattern = re.compile(r"\d{4}/\d{2}/\d{2}/\d{2}/\d{2}/\d{2}")
-
-    data_vars = {}
-    for key in keys:
-        filename = key.split("/")[-1].rsplit(".", 1)[0]
-
-        file_info = res.file_mapping.get(filename)
-        if not file_info:
-            continue
-
-        timestamp_str = timestamp_pattern.search(key).group()
-
-        for band_num, band_name in enumerate(file_info.bands, start=1):
-            array = _create_dask_array_v4(
-                session,
-                f"s3://{res.bucket.name}/{key}",
-                band_num,
-                first_file_metadata["height"],
-                first_file_metadata["width"],
-                file_info.type,
-            )
-            da = xarray.DataArray(
-                array,
-                dims=("y", "x"),
-            )
-            da.name = band_name
-
-            # Dataset with time dimension
-            if timestamp_str != "0000/00/00/00/00/00":
-                time = datetime.strptime(timestamp_str, "%Y/%m/%d/%H/%M/%S")
-                da = da.expand_dims("time")
-                da = da.assign_coords(time=[time])
-
-            if band_name not in data_vars:
-                data_vars[band_name] = []
-
-            data_vars[band_name].append(da)
-
-    for variable_name, time_series in data_vars.items():
-        if "time" in time_series[0].dims:
-            data_vars[variable_name] = xarray.concat(
-                time_series,
-                dim="time",
-                join="exact",
-            )
-        else:
-            data_vars[variable_name] = time_series[0]
-
-    ds = xarray.Dataset(
-        data_vars=data_vars,
-        coords={
-            "y": first_file_metadata["y"],
-            "x": first_file_metadata["x"],
-        },
-        attrs={
-            "provider_name": res.provider_name,
-            "dataset_name": res.dataset_name,
-            "dataset_id": res.dataset_id,
-            "aoi_id": res.aoi_id,
-            "subscription_id": res.data_request_id,
-        },
-    )
-    ds = ds.rio.write_crs(first_file_metadata["crs"])
-
-    return ds
-
-
-def _get_file_metadata_v4(session, bucket: str, path: str):
-    with rasterio.env.Env(
-        rasterio.session.AWSSession(session), GDAL_DISABLE_READDIR_ON_OPEN=True
-    ):
-        da = xarray.open_dataarray(f"s3://{bucket}/{path}", engine="rasterio")
-
-    return {
-        "crs": da.rio.crs,
-        "height": da.rio.height,
-        "width": da.rio.width,
-        "x": da.x.values,
-        "y": da.y.values,
-    }
-
-
-def _create_dask_array_v4(
-    session: boto3.session.Session,
-    file_path: str,
-    band_num: int,
-    height: int,
-    width: int,
-    dtype: str,
-):
-    rasterio_session = rasterio.session.AWSSession(session)
-
-    def read_chunk():
-        with rasterio.env.Env(
-            session=rasterio_session, GDAL_DISABLE_READDIR_ON_OPEN=True
-        ):
-            with rasterio.open(file_path) as src:
-                return src.read(band_num)
-
-    return dask.array.from_delayed(
-        dask.delayed(read_chunk)(), shape=(height, width), dtype=dtype
-    )