🔖 0.6.2 (#88)

* 🚑 Fix #87 boolean operator losing index

* 📝 Update doc for case_when (#87)

* 📝 Update doc for case_when (#87)

* 📝 Fix links in reference map

* 🔖 0.6.2

* ✨ Add `base.diff()`

* 📝 Update docs for `dplyr.base`

* 📝 Update mkdocs.yml

* 📝 Update CHANGELOG

* 🚑 Fix false alarm from `rename()`/`relocate()` for missing grouping variables (#89)

* 📝 Update CHANGELOG

* 📝 Add deps for docs building

datar/__init__.py

@@ -30,7 +30,7 @@
 )
 
 __all__ = ("f", "get_versions")
-__version__ = "0.6.1"
+__version__ = "0.6.2"
 
 
 def get_versions(prnt: bool = True) -> _VersionsTuple:

datar/all.py

@@ -14,6 +14,7 @@
 from .dplyr import _warn as _
 from .tibble import *
 from .tidyr import *
+from .base import rank  # overwrite dplyr.rank
 
 _builtin_names = _base_builtin_names.copy()
 _builtin_names.update(_dplyr_builtin_names)

datar/base/__init__.py

@@ -59,6 +59,7 @@
 from .funs import (
     cut,
     data_context,
+    diff,
     expandgrid,
     identity,
     make_unique,

datar/base/funs.py

@@ -12,6 +12,7 @@
 
 from ..core.middlewares import WithDataEnv
 from ..core.contexts import Context
+from ..core.factory import func_factory
 from ..core.tibble import Tibble
 from ..core.utils import arg_match, name_of
 from ..core.names import repair_names
@@ -66,6 +67,38 @@ def cut(
     )
 
 
+@func_factory("agg", "x")
+def diff(x, lag: int = 1, differences: int = 1):
+    """Calculates suitably lagged and iterated differences.
+
+    If the data is a vector of length n and differences = 1, then the computed
+    result is equal to the successive differences
+    `x[lag:] – x[:-lag]`.
+
+    Examples:
+        >>> rv = [52, 21, 10, 11, 19]
+        >>> data = diff(rv)
+        >>> # -31 -11 1 8
+        >>> # rv[1:] - rv[:-1]
+        >>> # rv[1:]  [21, 10, 11, 19]
+        >>> # rv[:-1] [52, 21, 10, 11]
+
+    Args:
+        x: The data
+        lag: The lag to use. Could be negative.
+            It always calculates `x[lag:] - x[:-lag]` even when `lag` is negative
+        differences: The order of the difference
+
+    Returns:
+        An array of `x[lag:] – x[:-lag]`.
+        If `differences > 1`, the rule applies `differences` times on `x`
+    """
+    x = x.values
+    for _ in range(differences):
+        x = x[lag:] - x[:-lag]
+    return x
+
+
 @register_func(None, context=Context.EVAL)
 def identity(x):
     """Return whatever passed in

datar/core/broadcast.py

@@ -668,7 +668,9 @@ def init_tibble_from(value, name: str) -> Tibble:
 
 @init_tibble_from.register(Series)
 def _(value: Series, name: str) -> Tibble:
-    name = name or value.name
+    # Deprecate warning, None will be used as series name in the future
+    # So use 0 as default here
+    name = name or value.name or 0
     return Tibble(value.to_frame(name=name), copy=False)
 
 

datar/core/operator.py

@@ -14,8 +14,15 @@
 def _binop(op, left, right, fill_false=False):
     left, right, grouper, is_rowwise = broadcast2(left, right)
     if fill_false:
-        left = Series(left).fillna(False).values
-        right = Series(right).fillna(False).values
+        if isinstance(left, Series):
+            left = left.fillna(False)
+        else:
+            left = Series(left).fillna(False).values
+
+        if isinstance(right, Series):
+            right = right.fillna(False)
+        else:
+            right = Series(right).fillna(False).values
 
     out = op(left, right)
     if grouper:

datar/dplyr/relocate.py

@@ -52,6 +52,7 @@ def relocate(
         *args,
         **kwargs,
         _group_vars=gvars,
+        _missing_gvars_inform=False,
     )
 
     to_move = list(to_move)
@@ -65,6 +66,7 @@ def relocate(
                 all_columns,
                 _before,
                 _group_vars=[],
+                _missing_gvars_inform=False,
             )[0]
         )
         if where not in to_move:
@@ -76,6 +78,7 @@ def relocate(
                 all_columns,
                 _after,
                 _group_vars=[],
+                _missing_gvars_inform=False,
             )[0]
         )
         if where not in to_move:

datar/dplyr/rename.py

@@ -29,6 +29,7 @@ def rename(_data, **kwargs):
     selected, new_names = _eval_select(
         all_columns,
         _group_vars=gvars,
+        _missing_gvars_inform=False,
         **kwargs,
     )
 

datar/dplyr/select.py

@@ -66,6 +66,7 @@ def _eval_select(
     _all_columns: Index,
     *args: Any,
     _group_vars: Sequence[str],
+    _missing_gvars_inform: bool = True,
     **kwargs: Any,
 ) -> Tuple[Sequence[int], Mapping[str, str]]:
     """Evaluate selections to get locations
@@ -79,9 +80,10 @@ def _eval_select(
         *kwargs.values(),
     )
 
-    missing = regcall(setdiff, _group_vars, _all_columns[selected_idx])
-    if len(missing) > 0:
-        logger.info("Adding missing grouping variables: %s", missing)
+    if _missing_gvars_inform:
+        missing = regcall(setdiff, _group_vars, _all_columns[selected_idx])
+        if len(missing) > 0:
+            logger.info("Adding missing grouping variables: %s", missing)
 
     selected_idx = regcall(
         union,

docs/CHANGELOG.md

@@ -1,3 +1,12 @@
+## 0.6.2
+
+- 🚑 Fix #87 boolean operator losing index
+- 🚑 Fix false alarm from `rename()`/`relocate()` for missing grouping variables (#89)
+- ✨ Add `base.diff()`
+- 📝 [doc] Update/Fix doc for case_when (#87)
+- 📝 [doc] Fix links in reference map
+- 📝 [doc] Update docs for `dplyr.base`
+
 ## 0.6.1
 
 - 🐛 Fix `rep(df, n)` producing a nested df