revert: Remove all tag based features

vega · Jan 13, 2025 · 5d8b6db · 5d8b6db
1 parent 6035b39
commit 5d8b6db
Show file tree

Hide file tree

Showing 4 changed files with 27 additions and 126 deletions.
diff --git a/altair/datasets/__init__.py b/altair/datasets/__init__.py
@@ -14,7 +14,7 @@
         from typing_extensions import LiteralString
 
     from altair.datasets._loader import _Load
-    from altair.datasets._typing import Dataset, Extension, Version
+    from altair.datasets._typing import Dataset, Extension
 
 
 __all__ = ["Loader", "load", "url"]
@@ -47,7 +47,6 @@ def url(
     name: Dataset | LiteralString,
     suffix: Extension | None = None,
     /,
-    tag: Version | None = None,
 ) -> str:
     """
     Return the address of a remote dataset.
@@ -61,15 +60,11 @@ def url(
 
         .. note::
             Only needed if ``name`` is available in multiple formats.
-    tag
-        Version identifier for a `vega-datasets release`_.
 
     .. _Path.stem:
         https://docs.python.org/3/library/pathlib.html#pathlib.PurePath.stem
     .. _Path.suffix:
         https://docs.python.org/3/library/pathlib.html#pathlib.PurePath.suffix
-    .. _vega-datasets release:
-        https://github.com/vega/vega-datasets/releases
 
     Related
     -------
@@ -83,7 +78,7 @@ def url(
     try:
         from altair.datasets._loader import load
 
-        url = load.url(name, suffix, tag=tag)
+        url = load.url(name, suffix)
     except AltairDatasetsError:
         from altair.datasets._cache import url_cache
 

diff --git a/altair/datasets/_cache.py b/altair/datasets/_cache.py
@@ -8,8 +8,6 @@
 import narwhals.stable.v1 as nw
 from narwhals.stable.v1.typing import IntoDataFrameT, IntoFrameT
 
-from altair.datasets._typing import VERSION_LATEST
-
 if sys.version_info >= (3, 12):
     from typing import Protocol
 else:
@@ -105,10 +103,7 @@ class UrlCache(CompressedCache[_KT, _VT]):
     `csv`_, `gzip`_ -based, lazy url lookup.
 
     Operates on a subset of available datasets:
-    - Only the latest version
     - Excludes `.parquet`, which `cannot be read via url`_
-    - Name collisions are pre-resolved
-        - Only provide the smallest (e.g. ``weather.json`` instead of ``weather.csv``)
 
     .. _csv:
         https://docs.python.org/3/library/csv.html
@@ -256,13 +251,10 @@ def download_all(self) -> None:
         Requires **30-50MB** of disk-space.
         """
         stems = tuple(fp.stem for fp in self)
-        latest = nw.col("tag") == nw.lit(VERSION_LATEST)
-        predicates = (~(nw.col("sha").is_in(stems)), latest) if stems else (latest,)
+        predicates = (~(nw.col("sha").is_in(stems)),) if stems else ()
         frame = (
-            self._rd._scan_metadata(
-                predicates, ext_supported=True, name_collision=False
-            )
-            .select("sha", "suffix", "url_npm")
+            self._rd._scan_metadata(predicates, is_image=False)  # type: ignore
+            .select("sha", "suffix", "url")
             .unique("sha")
             .collect()
         )
@@ -272,7 +264,7 @@ def download_all(self) -> None:
         print(f"Downloading {len(frame)} missing datasets...")
         for row in frame.iter_rows(named=True):
             fp: Path = self.path / (row["sha"] + row["suffix"])
-            with self._rd._opener.open(row["url_npm"]) as f:
+            with self._rd._opener.open(row["url"]) as f:
                 fp.touch()
                 fp.write_bytes(f.read())
         print("Finished downloads")

diff --git a/altair/datasets/_loader.py b/altair/datasets/_loader.py
@@ -21,7 +21,7 @@
     else:
         from typing_extensions import LiteralString
     from altair.datasets._readers import _Backend
-    from altair.datasets._typing import Dataset, Extension, Version
+    from altair.datasets._typing import Dataset, Extension
 
 
 __all__ = ["Loader", "load"]
@@ -111,7 +111,7 @@ def from_backend(cls, backend_name: _Backend, /) -> Loader[Any, Any]:
         Using ``pandas``, backed by ``pyarrow`` dtypes:
 
             data = Loader.from_backend("pandas[pyarrow]")
-            cars = data("cars", tag="v1.29.0")
+            cars = data("cars")
 
             >>> type(cars)  # doctest: +SKIP
             pandas.core.frame.DataFrame
@@ -137,7 +137,6 @@ def __call__(
         name: Dataset | LiteralString,
         suffix: Extension | None = None,
         /,
-        tag: Version | None = None,
         **kwds: Any,
     ) -> IntoDataFrameT:
         """
@@ -152,17 +151,13 @@ def __call__(
 
             .. note::
                 Only needed if ``name`` is available in multiple formats.
-        tag
-            Version identifier for a `vega-datasets release`_.
         **kwds
             Arguments passed to the underlying read function.
 
         .. _Path.stem:
             https://docs.python.org/3/library/pathlib.html#pathlib.PurePath.stem
         .. _Path.suffix:
             https://docs.python.org/3/library/pathlib.html#pathlib.PurePath.suffix
-        .. _vega-datasets release:
-            https://github.com/vega/vega-datasets/releases
 
         Examples
         --------
@@ -171,7 +166,7 @@ def __call__(
             from altair.datasets import Loader
 
             data = Loader.from_backend("polars")
-            source = data("iowa-electricity", tag="v2.10.0")
+            source = data("iowa-electricity")
 
             >>> source.columns  # doctest: +SKIP
             ['year', 'source', 'net_generation']
@@ -199,7 +194,7 @@ def __call__(
         Using ``pandas``:
 
             data = Loader.from_backend("pandas")
-            source = data("iowa-electricity", tag="v2.10.0")
+            source = data("iowa-electricity")
 
             >>> source.columns  # doctest: +SKIP
             Index(['year', 'source', 'net_generation'], dtype='object')
@@ -223,7 +218,7 @@ def __call__(
         Using ``pyarrow``:
 
             data = Loader.from_backend("pyarrow")
-            source = data("iowa-electricity", tag="v2.10.0")
+            source = data("iowa-electricity")
 
             >>> source.column_names  # doctest: +SKIP
             ['year', 'source', 'net_generation']
@@ -238,14 +233,13 @@ def __call__(
             source: [["Fossil Fuels","Fossil Fuels","Fossil Fuels","Fossil Fuels","Fossil Fuels",...,"Renewables","Renewables","Renewables","Renewables","Renewables"]]
             net_generation: [[35361,35991,36234,36205,36883,...,16476,17452,19091,21241,21933]]
         """
-        return self._reader.dataset(name, suffix, tag=tag, **kwds)
+        return self._reader.dataset(name, suffix, **kwds)
 
     def url(
         self,
         name: Dataset | LiteralString,
         suffix: Extension | None = None,
         /,
-        tag: Version | None = None,
     ) -> str:
         """
         Return the address of a remote dataset.
@@ -259,15 +253,11 @@ def url(
 
             .. note::
                 Only needed if ``name`` is available in multiple formats.
-        tag
-            Version identifier for a `vega-datasets release`_.
 
         .. _Path.stem:
             https://docs.python.org/3/library/pathlib.html#pathlib.PurePath.stem
         .. _Path.suffix:
             https://docs.python.org/3/library/pathlib.html#pathlib.PurePath.suffix
-        .. _vega-datasets release:
-            https://github.com/vega/vega-datasets/releases
 
         Examples
         --------
@@ -277,15 +267,15 @@ def url(
             from altair.datasets import Loader
 
             data = Loader.from_backend("polars")
-            >>> data.url("cars", tag="v2.9.0")  # doctest: +SKIP
-            'https://cdn.jsdelivr.net/npm/vega-datasets@v2.9.0/data/cars.json'
+            >>> data.url("cars")  # doctest: +SKIP
+            'https://cdn.jsdelivr.net/npm/vega-datasets@v2.11.0/data/cars.json'
 
         We can pass the result directly to a chart:
 
-            url = data.url("cars", tag="v2.9.0")
+            url = data.url("cars")
             alt.Chart(url).mark_point().encode(x="Horsepower:Q", y="Miles_per_Gallon:Q")
         """
-        return self._reader.url(name, suffix, tag=tag)
+        return self._reader.url(name, suffix)
 
     @property
     def cache(self) -> DatasetCache[IntoDataFrameT, IntoFrameT]:
@@ -318,7 +308,6 @@ def __call__(  # pyright: ignore[reportOverlappingOverload]
         name: Dataset | LiteralString,
         suffix: Extension | None = ...,
         /,
-        tag: Version | None = ...,
         backend: None = ...,
         **kwds: Any,
     ) -> IntoDataFrameT: ...
@@ -328,7 +317,6 @@ def __call__(
         name: Dataset | LiteralString,
         suffix: Extension | None = ...,
         /,
-        tag: Version | None = ...,
         backend: Literal["polars"] = ...,
         **kwds: Any,
     ) -> pl.DataFrame: ...
@@ -338,7 +326,6 @@ def __call__(
         name: Dataset | LiteralString,
         suffix: Extension | None = ...,
         /,
-        tag: Version | None = ...,
         backend: Literal["pandas", "pandas[pyarrow]"] = ...,
         **kwds: Any,
     ) -> pd.DataFrame: ...
@@ -348,7 +335,6 @@ def __call__(
         name: Dataset | LiteralString,
         suffix: Extension | None = ...,
         /,
-        tag: Version | None = ...,
         backend: Literal["pyarrow"] = ...,
         **kwds: Any,
     ) -> pa.Table: ...
@@ -357,14 +343,13 @@ def __call__(
         name: Dataset | LiteralString,
         suffix: Extension | None = None,
         /,
-        tag: Version | None = None,
         backend: _Backend | None = None,
         **kwds: Any,
     ) -> IntoDataFrameT | pl.DataFrame | pd.DataFrame | pa.Table:
         if backend is None:
-            return super().__call__(name, suffix, tag, **kwds)
+            return super().__call__(name, suffix, **kwds)
         else:
-            return self.from_backend(backend)(name, suffix, tag=tag, **kwds)
+            return self.from_backend(backend)(name, suffix, **kwds)
 
 
 load: _Load[Any, Any]