fix #274 : implemented set_printoptions() context_manager to set printing options for arrays

Author: alixdamman

doc/source/api.rst

@@ -664,6 +664,7 @@ Miscellaneous
    from_frame
    from_series
    get_example_filepath
+   set_printoptions
    labels_array
    union
    stack

doc/source/changes/version_0_30.rst.inc

@@ -224,6 +224,41 @@ Miscellaneous improvements
 * implemented :py:obj:`LArray.reverse()` method to reverse one or several axes of an array (closes :issue:`631`).
 
 
+* added :py:obj:`set_printoptions` allowing to set options for printing arrays within a ``with`` block or globally:
+
+    >>> from larray import *
+    >>> arr = ndtest((500, 100), dtype=float) / 3
+
+    You can use ``set_options`` either as a context manager:
+
+    >>> with set_printoptions(display_width=60, edgeitems=2, precision=2):
+    ...     print(arr)
+     a\b        b0        b1  ...       b98       b99
+      a0         0      0.33  ...     32.67        33
+      a1     33.33     33.67  ...        66     66.33
+     ...       ...       ...  ...       ...       ...
+    a498     16600  16600.33  ...  16632.67     16633
+    a499  16633.33  16633.67  ...     16666  16666.33
+
+    Or to set global options:
+
+    >>> set_printoptions(display_width=40, maxlines=10, precision=2)
+    >>> print(arr)
+     a\b        b0  ...       b99
+      a0         0  ...        33
+      a1     33.33  ...     66.33
+      a2     66.67  ...     99.67
+      a3       100  ...       133
+      a4    133.33  ...    166.33
+     ...       ...  ...       ...
+    a495     16500  ...     16533
+    a496  16533.33  ...  16566.33
+    a497  16566.67  ...  16599.67
+    a498     16600  ...     16633
+    a499  16633.33  ...  16666.33
+
+  Closes :issue:`274`.
+
 Fixes
 -----
 

larray/core/array.py

@@ -65,6 +65,7 @@
 from larray.util.misc import (table2str, size2str, basestring, izip, rproduct, ReprString, duplicates,
                               float_error_handler_factory, _isnoneslice, light_product, unique_list, common_type,
                               renamed_to, deprecate_kwarg, LHDFStore, lazy_attribute)
+from larray.util.options import OPTIONS, DISPLAY_PRECISION, DISPLAY_WIDTH, MAXLINES, EDGEITEMS
 
 
 def all(values, axis=None):
@@ -2277,8 +2278,12 @@ def __str__(self):
         elif not len(self):
             return 'LArray([])'
         else:
-            table = list(self.as_table(maxlines=200, edgeitems=5))
-            return table2str(table, 'nan', fullinfo=True, maxwidth=200, keepcols=self.ndim - 1)
+            maxlines = OPTIONS[MAXLINES] if OPTIONS[MAXLINES] is not None else 200
+            maxwidth = OPTIONS[DISPLAY_WIDTH]
+            precision = OPTIONS[DISPLAY_PRECISION]
+            table = list(self.as_table(maxlines))
+            return table2str(table, 'nan', fullinfo=True, maxwidth=maxwidth, keepcols=self.ndim - 1,
+                             precision=precision)
     __repr__ = __str__
 
     def __iter__(self):
@@ -2287,8 +2292,9 @@ def __iter__(self):
     def __contains__(self, key):
         return any(key in axis for axis in self.axes)
 
-    def as_table(self, maxlines=None, edgeitems=5, light=False, wide=True, value_name='value'):
-        """
+    def as_table(self, maxlines=None, edgeitems=None, light=False, wide=True, value_name='value'):
+        r"""as_table(maxlines=None, edgeitems=5, light=False, wide=True, value_name='value')
+
         Generator. Returns next line of the table representing an array.
 
         Parameters
@@ -2300,6 +2306,9 @@ def as_table(self, maxlines=None, edgeitems=5, light=False, wide=True, value_nam
             only the first and last `edgeitems` lines are displayed.
             Only active if `maxlines` is not None.
             Equals to 5 by default.
+        light: bool, optional
+            Whether or not printing the array in the same way as a pandas DataFrame with a MultiIndex
+            (see example below). Defaults to False.
         wide : boolean, optional
             Whether or not to write arrays in "wide" format. If True, arrays are exported with the last axis
             represented horizontally. If False, arrays are exported in "narrow" format: one column per axis plus one
@@ -2317,13 +2326,13 @@ def as_table(self, maxlines=None, edgeitems=5, light=False, wide=True, value_nam
         --------
         >>> arr = ndtest((2, 2, 3))
         >>> list(arr.as_table())  # doctest: +NORMALIZE_WHITESPACE
-        [['a', 'b\\\\c', 'c0', 'c1', 'c2'],
+        [['a', 'b\\c', 'c0', 'c1', 'c2'],
          ['a0', 'b0', 0, 1, 2],
          ['a0', 'b1', 3, 4, 5],
          ['a1', 'b0', 6, 7, 8],
          ['a1', 'b1', 9, 10, 11]]
         >>> list(arr.as_table(light=True))  # doctest: +NORMALIZE_WHITESPACE
-        [['a', 'b\\\\c', 'c0', 'c1', 'c2'],
+        [['a', 'b\\c', 'c0', 'c1', 'c2'],
          ['a0', 'b0', 0, 1, 2],
          ['', 'b1', 3, 4, 5],
          ['a1', 'b0', 6, 7, 8],
@@ -2346,6 +2355,12 @@ def as_table(self, maxlines=None, edgeitems=5, light=False, wide=True, value_nam
         if not self.ndim:
             return
 
+        # get default options
+        if maxlines is None:
+            maxlines = OPTIONS[MAXLINES]
+        if edgeitems is None:
+            edgeitems = OPTIONS[EDGEITEMS]
+
         # ert     unit  geo\time  2012    2011    2010
         # NEER27  I05   AT        101.41  101.63  101.63
         # NEER27  I05   AU        134.86  125.29  117.08
@@ -2398,7 +2413,9 @@ def as_table(self, maxlines=None, edgeitems=5, light=False, wide=True, value_nam
                 yield list(tick) + dataline.tolist()
 
     def dump(self, header=True, wide=True, value_name='value'):
-        """Dump array as a 2D nested list
+        """dump(header=True, wide=True, value_name='value')
+
+        Dump array as a 2D nested list
 
         Parameters
         ----------

larray/tests/test_options.py

@@ -0,0 +1,30 @@
+from __future__ import absolute_import, division, print_function
+import pytest
+import larray
+from larray.util.options import OPTIONS
+
+
+def test_invalid_option_raises():
+    with pytest.raises(ValueError):
+        larray.set_printoptions(not_a_valid_options=True)
+
+
+def test_set_options_as_global():
+    original_ops = OPTIONS.copy()
+    arr = larray.ndtest((500, 100))
+    larray.set_printoptions(display_width=40, maxlines=10)
+    expected = """\
+ a\\b     b0     b1  ...    b98    b99
+  a0      0      1  ...     98     99
+  a1    100    101  ...    198    199
+  a2    200    201  ...    298    299
+  a3    300    301  ...    398    399
+  a4    400    401  ...    498    499
+ ...    ...    ...  ...    ...    ...
+a495  49500  49501  ...  49598  49599
+a496  49600  49601  ...  49698  49699
+a497  49700  49701  ...  49798  49799
+a498  49800  49801  ...  49898  49899
+a499  49900  49901  ...  49998  49999"""
+    assert str(arr) == expected
+    larray.set_printoptions(**original_ops)

larray/util/__init__.py

@@ -1 +1,3 @@
 from __future__ import absolute_import, division, print_function
+
+from larray.util.options import *

larray/util/misc.py

@@ -81,18 +81,23 @@ def prod(values):
     return reduce(operator.mul, values, 1)
 
 
-def format_value(value, missing, fullinfo=False):
+def format_value(value, missing, fullinfo=False, precision=None):
     if isinstance(value, float) and not fullinfo:
         # nans print as "-1.#J", let's use something nicer
         if value != value:
             return missing
+        elif precision is not None:
+            return '{:2.{precision}f}'.format(value, precision=precision).rstrip('0').rstrip('.')
         else:
             return '%2.f' % value
     elif isinstance(value, np.ndarray) and value.shape:
         # prevent numpy's default wrapping
         return str(list(value)).replace(',', '')
     else:
-        return str(value)
+        if isinstance(value, float) and precision is not None:
+            return '{:2.{precision}f}'.format(value, precision=precision).rstrip('0').rstrip('.')
+        else:
+            return str(value)
 
 
 def get_col_width(table, index):
@@ -131,8 +136,9 @@ def get_min_width(table, index):
     return max(longest_word(row[index]) for row in table)
 
 
+# XXX: what fullinfo is used for and when ?
 def table2str(table, missing, fullinfo=False, summarize=True, maxwidth=80, numedges='auto', sep='  ', cont='...',
-              keepcols=0):
+              keepcols=0, precision=None):
     """
     table is a list of lists
     :type table: list of list
@@ -144,7 +150,7 @@ def table2str(table, missing, fullinfo=False, summarize=True, maxwidth=80, numed
     for row in table:
         if len(row) < numcol:
             row.extend([''] * (numcol - len(row)))
-    formatted = [[format_value(value, missing, fullinfo) for value in row]
+    formatted = [[format_value(value, missing, fullinfo, precision) for value in row]
                  for row in table]
     maxwidths = [get_col_width(formatted, i) for i in range(numcol)]
 

larray/util/options.py

@@ -0,0 +1,103 @@
+from __future__ import absolute_import, division, print_function
+
+from larray.util.misc import basestring
+
+
+__all__ = ['set_printoptions']
+
+
+DISPLAY_PRECISION = 'precision'
+DISPLAY_WIDTH = 'display_width'
+MAXLINES = 'maxlines'
+EDGEITEMS = 'edgeitems'
+
+
+OPTIONS = {
+    DISPLAY_PRECISION: None,
+    DISPLAY_WIDTH: 200,
+    MAXLINES: None,
+    EDGEITEMS: 5,
+}
+
+
+def _positive_integer(value):
+    if not (isinstance(value, int) and value > 0):
+        raise ValueError("Expected positive integer")
+
+
+def _positive_integer_or_none(value):
+    if value is None:
+        return
+    else:
+        _positive_integer(value)
+
+
+_VALIDATORS = {
+    DISPLAY_PRECISION: _positive_integer_or_none,
+    DISPLAY_WIDTH: _positive_integer,
+    MAXLINES: _positive_integer_or_none,
+    EDGEITEMS: _positive_integer,
+}
+
+
+# idea taken from xarray. See xarray/core/options.py module for original implementation.
+class set_printoptions(object):
+    r"""Set options for printing arrays in a controlled context.
+
+    Currently supported options:
+
+    - ``precision``: number of digits of precision for floating point output. Defaults to 8.
+    - ``display_width``: maximum display width for ``repr`` on xarray objects. Defaults to 80.
+    - ``maxlines``: Maximum number of lines to show. Default behavior shows all lines.
+    - ``edgeitems`` : if number of lines to display is greater than ``maxlines``, only the first and last
+      ``edgeitems`` lines are displayed. Only active if ``maxlines`` is not None. Defaults to 5.
+
+    Examples
+    --------
+    >>> from larray import *
+    >>> arr = ndtest((500, 100), dtype=float) / 3
+
+    You can use ``set_options`` either as a context manager:
+
+    >>> with set_printoptions(display_width=60, edgeitems=2, precision=2):
+    ...     print(arr)
+     a\b        b0        b1  ...       b98       b99
+      a0         0      0.33  ...     32.67        33
+      a1     33.33     33.67  ...        66     66.33
+     ...       ...       ...  ...       ...       ...
+    a498     16600  16600.33  ...  16632.67     16633
+    a499  16633.33  16633.67  ...     16666  16666.33
+
+    Or to set global options:
+
+    >>> set_printoptions(display_width=40, maxlines=10, precision=2) # doctest: +SKIP
+    >>> print(arr) # doctest: +SKIP
+     a\b        b0  ...       b99
+      a0         0  ...        33
+      a1     33.33  ...     66.33
+      a2     66.67  ...     99.67
+      a3       100  ...       133
+      a4    133.33  ...    166.33
+     ...       ...  ...       ...
+    a495     16500  ...     16533
+    a496  16533.33  ...  16566.33
+    a497  16566.67  ...  16599.67
+    a498     16600  ...     16633
+    a499  16633.33  ...  16666.33
+    """
+
+    def __init__(self, **kwargs):
+        self.old = {}
+        for k, v in kwargs.items():
+            if k not in OPTIONS:
+                raise ValueError('Argument {} is not in the set of valid options {}'.format(k, set(OPTIONS)))
+            if k in _VALIDATORS:
+                _VALIDATORS[k](v)
+            self.old[k] = OPTIONS[k]
+        OPTIONS.update(kwargs)
+
+    def __enter__(self):
+        return
+
+    def __exit__(self, type, value, traceback):
+        OPTIONS.update(self.old)