DEPS: set min versions

closes pandas-dev#15206, numpy >= 1.9
closes pandas-dev#15543, matplotlib >= 1.4.3
scipy >= 0.14.0

ci/requirements-2.7_COMPAT.build

@@ -1,5 +1,5 @@
 python=2.7*
-numpy=1.7.1
+numpy=1.9.2
 cython=0.23
 dateutil=1.5
 pytz=2013b

ci/requirements-2.7_COMPAT.run

@@ -1,11 +1,12 @@
-numpy=1.7.1
+numpy=1.9.2
 dateutil=1.5
 pytz=2013b
-scipy=0.11.0
+scipy=0.14.0
 xlwt=0.7.5
 xlrd=0.9.2
-numexpr=2.2.2
-pytables=3.0.0
+bottleneck=1.0.0
+numexpr=2.4.4 # this is actually unsupported for non-pytables
+pytables=3.2.2
 psycopg2
 pymysql=0.6.0
 sqlalchemy=0.7.8

ci/requirements-2.7_LOCALE.build

@@ -1,5 +1,5 @@
 python=2.7*
 python-dateutil
 pytz=2013b
-numpy=1.8.2
+numpy=1.9.2
 cython=0.23

ci/requirements-2.7_LOCALE.run

@@ -1,11 +1,12 @@
 python-dateutil
 pytz=2013b
-numpy=1.8.2
+numpy=1.9.2
 xlwt=0.7.5
 openpyxl=1.6.2
 xlsxwriter=0.5.2
 xlrd=0.9.2
-matplotlib=1.3.1
+bottleneck=1.0.0
+matplotlib=1.4.3
 sqlalchemy=0.8.1
 lxml=3.2.1
 scipy

ci/requirements-2.7_SLOW.build

@@ -1,5 +1,5 @@
 python=2.7*
 python-dateutil
 pytz
-numpy=1.8.2
+numpy=1.10*
 cython

ci/requirements-2.7_SLOW.run

@@ -1,7 +1,7 @@
 python-dateutil
 pytz
-numpy=1.8.2
-matplotlib=1.3.1
+numpy=1.10*
+matplotlib=1.4.3
 scipy
 patsy
 xlwt

doc/source/install.rst

@@ -203,7 +203,7 @@ Dependencies
 ------------
 
 * `setuptools <https://setuptools.readthedocs.io/en/latest/>`__
-* `NumPy <http://www.numpy.org>`__: 1.7.1 or higher
+* `NumPy <http://www.numpy.org>`__: 1.9.0 or higher
 * `python-dateutil <http://labix.org/python-dateutil>`__: 1.5 or higher
 * `pytz <http://pytz.sourceforge.net/>`__: Needed for time zone support
 
@@ -233,7 +233,7 @@ Optional Dependencies
 
 * `Cython <http://www.cython.org>`__: Only necessary to build development
   version. Version 0.23 or higher.
-* `SciPy <http://www.scipy.org>`__: miscellaneous statistical functions
+* `SciPy <http://www.scipy.org>`__: miscellaneous statistical functions, Version 0.14.0 or higher
 * `xarray <http://xarray.pydata.org>`__: pandas like handling for > 2 dims, needed for converting Panels to xarray objects. Version 0.7.0 or higher is recommended.
 * `PyTables <http://www.pytables.org>`__: necessary for HDF5-based storage. Version 3.0.0 or higher required, Version 3.2.1 or higher highly recommended.
 * `Feather Format <https://github.com/wesm/feather>`__: necessary for feather-based storage, version 0.3.1 or higher.
@@ -244,7 +244,7 @@ Optional Dependencies
   * `pymysql <https://github.com/PyMySQL/PyMySQL>`__: for MySQL.
   * `SQLite <https://docs.python.org/3.5/library/sqlite3.html>`__: for SQLite, this is included in Python's standard library by default.
 
-* `matplotlib <http://matplotlib.org/>`__: for plotting
+* `matplotlib <http://matplotlib.org/>`__: for plotting, Version 1.4.3 or higher.
 * For Excel I/O:
 
   * `xlrd/xlwt <http://www.python-excel.org/>`__: Excel reading (xlrd) and writing (xlwt)

doc/source/whatsnew/v0.21.0.txt

@@ -133,6 +133,26 @@ Other Enhancements
 Backwards incompatible API changes
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
+
+.. _whatsnew_0210.api_breaking.deps:
+
+Dependencies have increased minimum versions
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+We have updated our minimum supported versions of dependencies (:issue:`15206`, :issue:`15543`, :issue:`15214`). We now require:
+
+   +--------------+-----------------+
+   | Package      | Minimum Version |
+   +======================+=========+
+   | Numpy        | 1.9.0           |
+   +--------------+-----------------+
+   | Matplotlib   | 1.4.3           |
+   +--------------+-----------------+
+   | Scipy        | 0.14.0          |
+   +--------------+-----------------+
+   | Bottleneck   | 1.0.0           |
+   +--------------+-----------------+
+
 .. _whatsnew_0210.api_breaking.pandas_eval:
 
 Improved error handling during item assignment in pd.eval

pandas/_libs/sparse.pyx

@@ -12,8 +12,6 @@ from distutils.version import LooseVersion
 
 # numpy versioning
 _np_version = np.version.short_version
-_np_version_under1p8 = LooseVersion(_np_version) < '1.8'
-_np_version_under1p9 = LooseVersion(_np_version) < '1.9'
 _np_version_under1p10 = LooseVersion(_np_version) < '1.10'
 _np_version_under1p11 = LooseVersion(_np_version) < '1.11'
 

pandas/compat/numpy/__init__.py

@@ -9,19 +9,18 @@
 # numpy versioning
 _np_version = np.__version__
 _nlv = LooseVersion(_np_version)
-_np_version_under1p8 = _nlv < '1.8'
-_np_version_under1p9 = _nlv < '1.9'
 _np_version_under1p10 = _nlv < '1.10'
 _np_version_under1p11 = _nlv < '1.11'
 _np_version_under1p12 = _nlv < '1.12'
 _np_version_under1p13 = _nlv < '1.13'
 _np_version_under1p14 = _nlv < '1.14'
+_np_version_under1p15 = _nlv < '1.15'
 
-if _nlv < '1.7.0':
+if _nlv < '1.9':
     raise ImportError('this version of pandas is incompatible with '
-                      'numpy < 1.7.0\n'
+                      'numpy < 1.9.0\n'
                       'your numpy version is {0}.\n'
-                      'Please upgrade numpy to >= 1.7.0 to use '
+                      'Please upgrade numpy to >= 1.9.0 to use '
                       'this pandas version'.format(_np_version))
 
 
@@ -70,11 +69,10 @@ def np_array_datetime64_compat(arr, *args, **kwargs):
 
 
 __all__ = ['np',
-           '_np_version_under1p8',
-           '_np_version_under1p9',
            '_np_version_under1p10',
            '_np_version_under1p11',
            '_np_version_under1p12',
            '_np_version_under1p13',
-           '_np_version_under1p14'
+           '_np_version_under1p14',
+           '_np_version_under1p15'
            ]

pandas/core/algorithms.py

@@ -6,7 +6,6 @@
 from warnings import warn, catch_warnings
 import numpy as np
 
-from pandas import compat, _np_version_under1p8
 from pandas.core.dtypes.cast import maybe_promote
 from pandas.core.dtypes.generic import (
     ABCSeries, ABCIndex,
@@ -407,14 +406,12 @@ def isin(comps, values):
     comps, dtype, _ = _ensure_data(comps)
     values, _, _ = _ensure_data(values, dtype=dtype)
 
-    # GH11232
-    # work-around for numpy < 1.8 and comparisions on py3
     # faster for larger cases to use np.in1d
     f = lambda x, y: htable.ismember_object(x, values)
+
     # GH16012
     # Ensure np.in1d doesn't get object types or it *may* throw an exception
-    if ((_np_version_under1p8 and compat.PY3) or len(comps) > 1000000 and
-       not is_object_dtype(comps)):
+    if len(comps) > 1000000 and not is_object_dtype(comps):
         f = lambda x, y: np.in1d(x, y)
     elif is_integer_dtype(comps):
         try:

pandas/core/generic.py

@@ -1827,11 +1827,8 @@ def _box_item_values(self, key, values):
 
     def _maybe_cache_changed(self, item, value):
         """The object has called back to us saying maybe it has changed.
-
-        numpy < 1.8 has an issue with object arrays and aliasing
-        GH6026
         """
-        self._data.set(item, value, check=pd._np_version_under1p8)
+        self._data.set(item, value, check=False)
 
     @property
     def _is_cached(self):

pandas/core/groupby.py

@@ -13,7 +13,7 @@
 )
 
 from pandas import compat
-from pandas.compat.numpy import function as nv, _np_version_under1p8
+from pandas.compat.numpy import function as nv
 from pandas.compat import set_function_name
 
 from pandas.core.dtypes.common import (
@@ -3257,11 +3257,7 @@ def value_counts(self, normalize=False, sort=True, ascending=False,
             d = np.diff(np.r_[idx, len(ids)])
             if dropna:
                 m = ids[lab == -1]
-                if _np_version_under1p8:
-                    mi, ml = algorithms.factorize(m)
-                    d[ml] = d[ml] - np.bincount(mi)
-                else:
-                    np.add.at(d, m, -1)
+                np.add.at(d, m, -1)
                 acc = rep(d)[mask]
             else:
                 acc = rep(d)

pandas/core/internals.py

@@ -69,8 +69,7 @@
 import pandas.core.computation.expressions as expressions
 from pandas.util._decorators import cache_readonly
 from pandas.util._validators import validate_bool_kwarg
-
-from pandas import compat, _np_version_under1p9
+from pandas import compat
 from pandas.compat import range, map, zip, u
 
 
@@ -857,9 +856,6 @@ def _is_empty_indexer(indexer):
 
         # set
         else:
-            if _np_version_under1p9:
-                # Work around GH 6168 to support old numpy
-                indexer = getattr(indexer, 'values', indexer)
             values[indexer] = value
 
         # coerce and try to infer the dtypes of the result
@@ -1482,15 +1478,7 @@ def quantile(self, qs, interpolation='linear', axis=0, mgr=None):
         tuple of (axis, block)
 
         """
-        if _np_version_under1p9:
-            if interpolation != 'linear':
-                raise ValueError("Interpolation methods other than linear "
-                                 "are not supported in numpy < 1.9.")
-
-        kw = {}
-        if not _np_version_under1p9:
-            kw.update({'interpolation': interpolation})
-
+        kw = {'interpolation': interpolation}
         values = self.get_values()
         values, _, _, _ = self._try_coerce_args(values, values)
 

pandas/tests/frame/test_quantile.py

@@ -12,7 +12,6 @@
 from pandas.util.testing import assert_series_equal, assert_frame_equal
 
 import pandas.util.testing as tm
-from pandas import _np_version_under1p9
 
 from pandas.tests.frame.common import TestData
 
@@ -103,9 +102,6 @@ def test_quantile_axis_parameter(self):
 
     def test_quantile_interpolation(self):
         # see gh-10174
-        if _np_version_under1p9:
-            pytest.skip("Numpy version under 1.9")
-
         from numpy import percentile
 
         # interpolation = linear (default case)
@@ -166,44 +162,6 @@ def test_quantile_interpolation(self):
                                  index=[.25, .5], columns=['a', 'b', 'c'])
         assert_frame_equal(result, expected)
 
-    def test_quantile_interpolation_np_lt_1p9(self):
-        # see gh-10174
-        if not _np_version_under1p9:
-            pytest.skip("Numpy version is greater than 1.9")
-
-        from numpy import percentile
-
-        # interpolation = linear (default case)
-        q = self.tsframe.quantile(0.1, axis=0, interpolation='linear')
-        assert q['A'] == percentile(self.tsframe['A'], 10)
-        q = self.intframe.quantile(0.1)
-        assert q['A'] == percentile(self.intframe['A'], 10)
-
-        # test with and without interpolation keyword
-        q1 = self.intframe.quantile(0.1)
-        assert q1['A'] == np.percentile(self.intframe['A'], 10)
-        assert_series_equal(q, q1)
-
-        # interpolation method other than default linear
-        msg = "Interpolation methods other than linear"
-        df = DataFrame({"A": [1, 2, 3], "B": [2, 3, 4]}, index=[1, 2, 3])
-        with tm.assert_raises_regex(ValueError, msg):
-            df.quantile(.5, axis=1, interpolation='nearest')
-
-        with tm.assert_raises_regex(ValueError, msg):
-            df.quantile([.5, .75], axis=1, interpolation='lower')
-
-        # test degenerate case
-        df = DataFrame({'x': [], 'y': []})
-        with tm.assert_raises_regex(ValueError, msg):
-            q = df.quantile(0.1, axis=0, interpolation='higher')
-
-        # multi
-        df = DataFrame([[1, 1, 1], [2, 2, 2], [3, 3, 3]],
-                       columns=['a', 'b', 'c'])
-        with tm.assert_raises_regex(ValueError, msg):
-            df.quantile([.25, .5], interpolation='midpoint')
-
     def test_quantile_multi(self):
         df = DataFrame([[1, 1, 1], [2, 2, 2], [3, 3, 3]],
                        columns=['a', 'b', 'c'])

pandas/tests/indexes/datetimes/test_datetime.py

@@ -9,7 +9,7 @@
 from pandas.compat import lrange
 from pandas.compat.numpy import np_datetime64_compat
 from pandas import (DatetimeIndex, Index, date_range, Series, DataFrame,
-                    Timestamp, datetime, offsets, _np_version_under1p8)
+                    Timestamp, datetime, offsets)
 
 from pandas.util.testing import assert_series_equal, assert_almost_equal
 
@@ -276,11 +276,7 @@ def test_comparisons_nat(self):
                          np_datetime64_compat('2014-06-01 00:00Z'),
                          np_datetime64_compat('2014-07-01 00:00Z')])
 
-        if _np_version_under1p8:
-            # cannot test array because np.datetime('nat') returns today's date
-            cases = [(fidx1, fidx2), (didx1, didx2)]
-        else:
-            cases = [(fidx1, fidx2), (didx1, didx2), (didx1, darr)]
+        cases = [(fidx1, fidx2), (didx1, didx2), (didx1, darr)]
 
         # Check pd.NaT is handles as the same as np.nan
         with tm.assert_produces_warning(None):