Move SplitIntoBins, MapBins, IterateBins, get_example_bin, cell_to_st…

…ring from lena.flow to lena.structures.

docs/source/flow.rst

@@ -48,17 +48,6 @@ Flow
     Reverse
     Slice
 
-**Split into bins:**
-
-.. currentmodule:: lena.flow.split_into_bins
-.. autosummary::
-
-    IterateBins
-    MapBins
-    SplitIntoBins
-    cell_to_string
-    get_example_bin
-
 Elements
 --------
 Elements form Lena sequences.
@@ -111,7 +100,3 @@ Iterators
 .. automodule:: lena.flow.iterators
     :special-members: __call__
 
-Split into bins
----------------
-
-.. automodule:: lena.flow.split_into_bins

docs/source/structures.rst

@@ -21,6 +21,17 @@ Structures
 
     HistToGraph
 
+**Split into bins:**
+
+.. currentmodule:: lena.structures.split_into_bins
+.. autosummary::
+
+    IterateBins
+    MapBins
+    SplitIntoBins
+    cell_to_string
+    get_example_bin
+
 **Histogram functions:**
 
 .. currentmodule:: lena.structures.hist_functions
@@ -80,6 +91,11 @@ Graph
 .. autoclass:: HistToGraph
     :members:
 
+Split into bins
+---------------
+
+.. automodule:: lena.structures.split_into_bins
+
 Histogram functions
 -------------------
 .. automodule:: lena.structures.hist_functions

lena/flow/__init__.py

@@ -9,12 +9,6 @@
 from .print_ import Print
 from .progress import Progress
 from .selectors import Not, Selector
-from .split_into_bins import (
-    IterateBins,
-    MapBins,
-    SplitIntoBins,
-    get_example_bin,
-)
 from .zip import Zip
 from .filter import Filter
 
@@ -44,9 +38,4 @@
     'Selector',
     'seq_map',
     'RunIf',
-    # split into bins
-    'SplitIntoBins',
-    'IterateBins',
-    'MapBins',
-    'get_example_bin',
 ]

lena/structures/__init__.py

@@ -2,9 +2,11 @@
 from .graph import Graph
 from .histogram import histogram, Histogram
 from .hist_functions import (
+    cell_to_string,
     check_edges_increasing,
     get_bin_edges,
     get_bin_on_value_1d, get_bin_on_value, get_bin_on_index,
+    get_example_bin,
     HistCell,
     hist_to_graph,
     init_bins,
@@ -15,22 +17,36 @@
     unify_1_md
 )
 from .numpy_histogram import NumpyHistogram
+from .split_into_bins import (
+    IterateBins,
+    MapBins,
+    SplitIntoBins,
+)
 
 
 __all__ = [
+    # structures
     'Graph',
-    'histogram', 'Histogram',
+    'histogram',
+    'HistCell',
+    'Histogram',
+    'HistToGraph',
+    'NumpyHistogram',
+    # hist functions
     'check_edges_increasing',
+    'cell_to_string',
     'get_bin_edges',
     'get_bin_on_value_1d', 'get_bin_on_value', 'get_bin_on_index',
-    'HistCell',
-    'HistToGraph',
+    'get_example_bin',
     'hist_to_graph',
     'init_bins',
     'integral',
     'iter_bins',
     'iter_cells',
     'make_hist_context',
     'unify_1_md',
-    'NumpyHistogram',
+    # split into bins
+    'SplitIntoBins',
+    'IterateBins',
+    'MapBins',
 ]

lena/structures/elements.py

@@ -53,7 +53,8 @@ def run(self, flow):
         """
         from lena.flow import get_data_context
         from lena.flow import get_context
-        get_example_bin = lena.flow.split_into_bins.get_example_bin
+        # don't know differences between these two ways of imports
+        get_example_bin = lena.structures.get_example_bin
         update_nested = lena.context.update_nested
         # why can't it be a Call element, which just returns
         # unchanged values for non-histograms?

lena/structures/hist_functions.py

@@ -16,8 +16,6 @@
 
 import lena.core
 import lena.structures.graph
-# todo: move it to this module
-import lena.flow.split_into_bins
 
 
 class HistCell(collections.namedtuple("HistCell", ("edges, bin, index"))):
@@ -26,6 +24,49 @@ class HistCell(collections.namedtuple("HistCell", ("edges, bin, index"))):
     __slots__ = ()
 
 
+def cell_to_string(
+        cell_edges, var_context=None, coord_names=None,
+        coord_fmt="{}_lte_{}_lt_{}", coord_join="_", reverse=False):
+    """Transform cell edges into a string.
+
+    *cell_edges* is a tuple of pairs *(lower bound, upper bound)*
+    for each coordinate.
+
+    *coord_names* is a list of coordinates names.
+
+    *coord_fmt* is a string,
+    which defines how to format individual coordinates.
+
+    *coord_join* is a string, which joins coordinate pairs.
+
+    If *reverse* is True, coordinates are joined in reverse order.
+    """
+    # todo: do we really need var_context?
+    # todo: even if so, why isn't that a {}? Is that dangerous?
+    if coord_names is None:
+        if var_context is None:
+            coord_names = [
+                "coord{}".format(ind) for ind in range(len(cell_edges))
+            ]
+        else:
+            if "combine" in var_context:
+                coord_names = [var["name"]
+                               for var in var_context["combine"]]
+            else:
+                coord_names = [var_context["name"]]
+    if len(cell_edges) != len(coord_names):
+        raise lena.core.LenaValueError(
+            "coord_names must have same length as cell_edges, "
+            "{} and {} given".format(coord_names, cell_edges)
+        )
+    coord_strings = [coord_fmt.format(edge[0], coord_names[ind], edge[1])
+                     for (ind, edge) in enumerate(cell_edges)]
+    if reverse:
+        coord_strings = reversed(coord_strings)
+    coord_str = coord_join.join(coord_strings)
+    return coord_str
+
+
 def _check_edges_increasing_1d(arr):
     if len(arr) <= 1:
         raise lena.core.LenaValueError("size of edges should be more than one,"
@@ -237,6 +278,23 @@ def get_bin_on_value(arg, edges):
     return indices
 
 
+def get_example_bin(struct):
+    """Return bin with zero index on each axis of the histogram bins.
+
+    For example, if the histogram is two-dimensional, return hist[0][0].
+
+    *struct* can be a :class:`.histogram`
+    or an array of bins.
+    """
+    if isinstance(struct, lena.structures.histogram):
+        return lena.structures.get_bin_on_index([0] * struct.dim, struct.bins)
+    else:
+        bins = struct
+        while isinstance(bins, list):
+            bins = bins[0]
+        return bins
+
+
 def hist_to_graph(hist, make_value=None, get_coordinate="left", scale=None):
     """Convert a :class:`.histogram` to a :class:`.Graph`.
 
@@ -280,8 +338,7 @@ def hist_to_graph(hist, make_value=None, get_coordinate="left", scale=None):
             '"{}" provided'.format(get_coordinate)
         )
 
-    _ibe = lena.flow.split_into_bins._iter_bins_with_edges
-    for value, edges in _ibe(hist.bins, hist.edges):
+    for value, edges in _iter_bins_with_edges(hist.bins, hist.edges):
         coord = get_coord(edges)
         # todo: unclear when bin_context is present.
         bin_value = lena.flow.get_data(value)
@@ -379,6 +436,30 @@ def iter_bins(bins):
                 yield (((ind,) + sub_ind), val)
 
 
+def _iter_bins_with_edges(bins, edges):
+    """Yield *(bin content, bin edges)* pairs.
+
+    *Bin edges* is a tuple, such that at index *i*
+    its element is bin's *(lower bound, upper bound)*
+    along *i*-th the coordinate.
+    """
+    # todo: only a list or also a tuple, an array?
+    if not isinstance(edges[0], list):
+        edges = [edges]
+    bins_sizes = [len(edge)-1 for edge in edges]
+    indices = [list(range(nbins)) for nbins in bins_sizes]
+    for index in itertools.product(*indices):
+        bin_ = lena.structures.get_bin_on_index(index, bins)
+        edges_low = []
+        edges_high = []
+        for var, var_ind in enumerate(index):
+            edges_low.append(edges[var][var_ind])
+            edges_high.append(edges[var][var_ind+1])
+        yield (bin_, tuple(zip(edges_low, edges_high)))
+        # old interface:
+        # yield (bin_, (edges_low, edges_high))
+
+
 def iter_cells(hist, ranges=None, coord_ranges=None):
     """Iterate cells of a histogram *hist*, possibly in a subrange.
 

lena/flow/split_into_bins.py → lena/structures/split_into_bins.py

@@ -8,30 +8,9 @@
 import lena.math
 import lena.structures
 import lena.variables
-
-
-def _iter_bins_with_edges(bins, edges):
-    """Yield *(bin content, bin edges)* pairs.
-
-    *Bin edges* is a tuple, such that at index *i*
-    its element is bin's *(lower bound, upper bound)*
-    along *i*-th the coordinate.
-    """
-    # todo: only a list or also a tuple, an array?
-    if not isinstance(edges[0], list):
-        edges = [edges]
-    bins_sizes = [len(edge)-1 for edge in edges]
-    indices = [list(range(nbins)) for nbins in bins_sizes]
-    for index in itertools.product(*indices):
-        bin_ = lena.structures.get_bin_on_index(index, bins)
-        edges_low = []
-        edges_high = []
-        for var, var_ind in enumerate(index):
-            edges_low.append(edges[var][var_ind])
-            edges_high.append(edges[var][var_ind+1])
-        yield (bin_, tuple(zip(edges_low, edges_high)))
-        # old interface:
-        # yield (bin_, (edges_low, edges_high))
+from .hist_functions import (
+    cell_to_string, get_example_bin, _iter_bins_with_edges
+)
 
 
 class _MdSeqMap(object):
@@ -58,66 +37,6 @@ def __iter__(self):
         return self
 
 
-def cell_to_string(
-        cell_edges, var_context=None, coord_names=None,
-        coord_fmt="{}_lte_{}_lt_{}", coord_join="_", reverse=False):
-    """Transform cell edges into a string.
-
-    *cell_edges* is a tuple of pairs *(lower bound, upper bound)*
-    for each coordinate.
-
-    *coord_names* is a list of coordinates names.
-
-    *coord_fmt* is a string,
-    which defines how to format individual coordinates.
-
-    *coord_join* is a string, which joins coordinate pairs.
-
-    If *reverse* is True, coordinates are joined in reverse order.
-    """
-    # todo: do we really need var_context?
-    # todo: even if so, why isn't that a {}? Is that dangerous?
-    if coord_names is None:
-        if var_context is None:
-            coord_names = [
-                "coord{}".format(ind) for ind in range(len(cell_edges))
-            ]
-        else:
-            if "combine" in var_context:
-                coord_names = [var["name"]
-                               for var in var_context["combine"]]
-            else:
-                coord_names = [var_context["name"]]
-    if len(cell_edges) != len(coord_names):
-        raise lena.core.LenaValueError(
-            "coord_names must have same length as cell_edges, "
-            "{} and {} given".format(coord_names, cell_edges)
-        )
-    coord_strings = [coord_fmt.format(edge[0], coord_names[ind], edge[1])
-                     for (ind, edge) in enumerate(cell_edges)]
-    if reverse:
-        coord_strings = reversed(coord_strings)
-    coord_str = coord_join.join(coord_strings)
-    return coord_str
-
-
-def get_example_bin(struct):
-    """Return bin with zero index on each axis of the histogram bins.
-
-    For example, if the histogram is two-dimensional, return hist[0][0].
-
-    *struct* can be a :class:`.histogram`
-    or an array of bins.
-    """
-    if isinstance(struct, lena.structures.histogram):
-        return lena.structures.get_bin_on_index([0] * struct.dim, struct.bins)
-    else:
-        bins = struct
-        while isinstance(bins, list):
-            bins = bins[0]
-        return bins
-
-
 class IterateBins(object):
     """Iterate bins of histograms."""