refactor(api): remove the remnants of now unsupported timecontext feature (ibis-project#8721)

The timecontext feature was only supported by the pandas, dask and
pyspark backends.
Since the epic split refactor none of the ported backends support this
feature.

BREAKING CHANGE: timecontext feature is removed

Author: kszucs

docs/_quarto.yml

@@ -174,7 +174,6 @@ website:
 
         - section: Configuration
           contents:
-            - reference/ContextAdjustment.qmd
             - reference/Interactive.qmd
             - reference/Options.qmd
             - reference/Repr.qmd
@@ -623,7 +622,6 @@ quartodoc:
       desc: "Ibis configuration"
       package: ibis.config
       contents:
-        - ContextAdjustment
         - Interactive
         - Options
         - Repr

ibis/backends/bigquery/__init__.py

@@ -719,8 +719,6 @@ def execute(self, expr, params=None, limit="default", **kwargs):
         """
         self._run_pre_execute_hooks(expr)
 
-        # TODO: upstream needs to pass params to raw_sql, I think.
-        kwargs.pop("timecontext", None)
         sql = self.compile(expr, limit=limit, params=params, **kwargs)
         self._log(sql)
         cursor = self.raw_sql(sql, params=params, **kwargs)

ibis/backends/dask/__init__.py

@@ -68,11 +68,7 @@ def disconnect(self) -> None:
     def version(self):
         return dask.__version__
 
-    def _validate_args(self, expr, limit, timecontext):
-        if timecontext is not None:
-            raise com.UnsupportedArgumentError(
-                "The Dask backend does not support timecontext"
-            )
+    def _validate_args(self, expr, limit):
         if limit != "default" and limit is not None:
             raise com.UnsupportedArgumentError(
                 "limit parameter to execute is not yet implemented in the "
@@ -88,12 +84,11 @@ def compile(
         expr: ir.Expr,
         params: dict | None = None,
         limit: int | None = None,
-        timecontext=None,
         **kwargs,
     ):
         from ibis.backends.dask.executor import DaskExecutor
 
-        self._validate_args(expr, limit, timecontext)
+        self._validate_args(expr, limit)
         params = params or {}
         params = {k.op() if isinstance(k, ir.Expr) else k: v for k, v in params.items()}
 
@@ -104,12 +99,11 @@ def execute(
         expr: ir.Expr,
         params: Mapping[ir.Expr, object] | None = None,
         limit: str = "default",
-        timecontext=None,
         **kwargs,
     ):
         from ibis.backends.dask.executor import DaskExecutor
 
-        self._validate_args(expr, limit, timecontext)
+        self._validate_args(expr, limit)
         params = params or {}
         params = {k.op() if isinstance(k, ir.Expr) else k: v for k, v in params.items()}
 

ibis/backends/tests/test_timecontext.py

@@ -42,20 +42,6 @@ def __call__(self, options):
         return self._with_temporary(options)
 
 
-class ContextAdjustment(Config):
-    """Options related to time context adjustment.
-
-    Attributes
-    ----------
-    time_col : str
-        Name of the timestamp column for execution with a `timecontext`. See
-        `ibis/expr/timecontext.py` for details.
-
-    """
-
-    time_col: str = "time"
-
-
 class SQL(Config):
     """SQL-related options.
 
@@ -153,8 +139,6 @@ class Options(Config):
     default_backend : Optional[ibis.backends.BaseBackend]
         The default backend to use for execution, defaults to DuckDB if not
         set.
-    context_adjustment : ContextAdjustment
-        Options related to time context adjustment.
     sql: SQL
         SQL-related options.
     clickhouse : Config | None
@@ -176,7 +160,6 @@ class Options(Config):
     verbose_log: Optional[Callable] = None
     graphviz_repr: bool = False
     default_backend: Optional[Any] = None
-    context_adjustment: ContextAdjustment = ContextAdjustment()
     sql: SQL = SQL()
     clickhouse: Optional[Config] = None
     dask: Optional[Config] = None

ibis/config.py

@@ -30,8 +30,6 @@
     from ibis.backends import BaseBackend
     from ibis.expr.visualize import EdgeAttributeGetter, NodeAttributeGetter
 
-    TimeContext = tuple[pd.Timestamp, pd.Timestamp]
-
 
 class _FixedTextJupyterMixin(JupyterMixin):
     """JupyterMixin adds a spurious newline to text, this fixes the issue."""
@@ -343,7 +341,6 @@ def _find_backend(self, *, use_default: bool = False) -> BaseBackend:
     def execute(
         self,
         limit: int | str | None = "default",
-        timecontext: TimeContext | None = None,
         params: Mapping[ir.Value, Any] | None = None,
         **kwargs: Any,
     ):
@@ -354,27 +351,18 @@ def execute(
         limit
             An integer to effect a specific row limit. A value of `None` means
             "no limit". The default is in `ibis/config.py`.
-        timecontext
-            Defines a time range of `(begin, end)`. When defined, the execution
-            will only compute result for data inside the time range. The time
-            range is inclusive of both endpoints. This is conceptually same as
-            a time filter.
-            The time column must be named `'time'` and should preserve
-            across the expression. For example, if that column is dropped then
-            execute will result in an error.
         params
             Mapping of scalar parameter expressions to value
         kwargs
             Keyword arguments
         """
         return self._find_backend(use_default=True).execute(
-            self, limit=limit, timecontext=timecontext, params=params, **kwargs
+            self, limit=limit, params=params, **kwargs
         )
 
     def compile(
         self,
         limit: int | None = None,
-        timecontext: TimeContext | None = None,
         params: Mapping[ir.Value, Any] | None = None,
         pretty: bool = False,
     ):
@@ -385,21 +373,13 @@ def compile(
         limit
             An integer to effect a specific row limit. A value of `None` means
             "no limit". The default is in `ibis/config.py`.
-        timecontext
-            Defines a time range of `(begin, end)`. When defined, the execution
-            will only compute result for data inside the time range. The time
-            range is inclusive of both endpoints. This is conceptually same as
-            a time filter.
-            The time column must be named `'time'` and should preserve
-            across the expression. For example, if that column is dropped then
-            execute will result in an error.
         params
             Mapping of scalar parameter expressions to value
         pretty
             In case of SQL backends, return a pretty formatted SQL query.
         """
         return self._find_backend().compile(
-            self, limit=limit, timecontext=timecontext, params=params, pretty=pretty
+            self, limit=limit, params=params, pretty=pretty
         )
 
     @experimental