core-lens 0.1.dev149__tar.gz → 0.1.dev151__tar.gz

{core_lens-0.1.dev149 → core_lens-0.1.dev151}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: core-lens
-Version: 0.1.dev149
+Version: 0.1.dev151
 Summary: Query, analyse, and visualise CoreStack's microwatershed and Earth science data through a clean, composable Python API.
 Project-URL: Homepage, https://github.com/ApoorvaKashyap/core-lens
 Project-URL: Issues, https://github.com/ApoorvaKashyap/core-lens/issues

{core_lens-0.1.dev149 → core_lens-0.1.dev151}/src/core_lens/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.1.dev149'
-__version_tuple__ = version_tuple = (0, 1, 'dev149')
+__version__ = version = '0.1.dev151'
+__version_tuple__ = version_tuple = (0, 1, 'dev151')
 
 __commit_id__ = commit_id = None

{core_lens-0.1.dev149 → core_lens-0.1.dev151}/src/core_lens/base/result.py

@@ -89,6 +89,14 @@ class Result:
         """
         return collect_lf(self.data)
 
+    def materialise(self) -> "Result":
+        """Evaluate the lazy computation graph and cache it in memory.
+
+        Returns:
+            Result: A new Result with the data cached as an in-memory LazyFrame.
+        """
+        return self._replace(data=self.df().lazy())
+
     def gdf(self) -> "gpd.GeoDataFrame":
         """Return the data as a ``GeoDataFrame``.
 
@@ -155,11 +163,25 @@ class Result:
             return self
 
         key_cols = self.key_cols
-        geo_lf = self.entity.geometry_lazy.join(
-            self.data.select(key_cols), on=key_cols, how="semi"
-        )
-        joined = self.data.join(geo_lf, on=key_cols, how="left")
-        return self._replace(data=joined, has_geometry=True)
+
+        # 1. Evaluate the heavy aggregation graph ONCE to prevent Polars from
+        # evaluating it twice during the geometry extraction and join.
+        res = self.materialise()
+        keys_df = res.df().select(key_cols)
+
+        # 2. Push down the geometry filter into the Parquet reader using is_in
+        # so we don't scan the entire static geometry file (which takes ~0.4s).
+        geo_lf = self.entity.geometry_lazy
+        if len(key_cols) == 1 and keys_df.height < 50000:
+            key = key_cols[0]
+            geo_lf = geo_lf.filter(pl.col(key).is_in(keys_df[key]))
+        else:
+            geo_lf = geo_lf.join(keys_df.lazy(), on=key_cols, how="semi")
+
+        joined = res.data.join(geo_lf, on=key_cols, how="left")
+
+        # 3. Cache the joined geometry so multiple plots are instant.
+        return res._replace(data=joined, has_geometry=True).materialise()
 
     def derive(self, name: str, expr: pl.Expr) -> "Result":
         """Return a new ``Result`` with a computed column appended.