Merge remote-tracking branch 'upstream/master'

Author: m-charlton

.travis.yml

@@ -80,11 +80,10 @@ matrix:
         apt:
           packages:
           - xsel
-    - python: 3.6-dev
+    - python: 3.6
       env:
         - PYTHON_VERSION=3.6
-        - JOB_NAME: "36_dev"
-        - JOB_TAG=_DEV
+        - JOB_NAME: "36"
         - NOSE_ARGS="not slow and not network and not disabled"
         - PANDAS_TESTING_MODE="deprecate"
       addons:
@@ -183,9 +182,9 @@ matrix:
         - CACHE_NAME="35_ascii"
         - USE_CACHE=true
 #    In allow_failures
-    - python: 2.7
+    - python: 3.5
       env:
-        - PYTHON_VERSION=2.7
+        - PYTHON_VERSION=3.5
         - JOB_NAME: "doc_build"
         - FULL_DEPS=true
         - DOC_BUILD=true
@@ -276,9 +275,9 @@ matrix:
         - LOCALE_OVERRIDE="C"
         - CACHE_NAME="35_ascii"
         - USE_CACHE=true
-      - python: 2.7
+      - python: 3.5
         env:
-        - PYTHON_VERSION=2.7
+        - PYTHON_VERSION=3.5
         - JOB_NAME: "doc_build"
         - FULL_DEPS=true
         - DOC_BUILD=true

appveyor.yml

@@ -80,6 +80,7 @@ install:
   - cmd: conda config --set ssl_verify false
 
   # add the pandas channel *before* defaults to have defaults take priority
+  - cmd: conda config --add channels conda-forge
   - cmd: conda config --add channels pandas
   - cmd: conda config --remove channels defaults
   - cmd: conda config --add channels defaults

asv_bench/benchmarks/inference.py

@@ -95,4 +95,23 @@ def setup(self, dtype, downcast):
         self.data = self.data_dict[dtype]
 
     def time_downcast(self, dtype, downcast):
-        pd.to_numeric(self.data, downcast=downcast)
+        pd.to_numeric(self.data, downcast=downcast)
+
+
+class MaybeConvertNumeric(object):
+
+    def setup(self):
+        n = 1000000
+        arr = np.repeat([2**63], n)
+        arr = arr + np.arange(n).astype('uint64')
+        arr = np.array([arr[i] if i%2 == 0 else
+                        str(arr[i]) for i in range(n)],
+                       dtype=object)
+
+        arr[-1] = -1
+        self.data = arr
+        self.na_values = set()
+
+    def time_convert(self):
+        pd.lib.maybe_convert_numeric(self.data, self.na_values,
+                                     coerce_numeric=False)

asv_bench/benchmarks/io_bench.py

@@ -128,6 +128,29 @@ def time_read_parse_dates_iso8601(self):
         read_csv(StringIO(self.data), header=None, names=['foo'], parse_dates=['foo'])
 
 
+class read_uint64_integers(object):
+    goal_time = 0.2
+
+    def setup(self):
+        self.na_values = [2**63 + 500]
+
+        self.arr1 = np.arange(10000).astype('uint64') + 2**63
+        self.data1 = '\n'.join(map(lambda x: str(x), self.arr1))
+
+        self.arr2 = self.arr1.copy().astype(object)
+        self.arr2[500] = -1
+        self.data2 = '\n'.join(map(lambda x: str(x), self.arr2))
+
+    def time_read_uint64(self):
+        read_csv(StringIO(self.data1), header=None)
+
+    def time_read_uint64_neg_values(self):
+        read_csv(StringIO(self.data2), header=None)
+
+    def time_read_uint64_na_values(self):
+        read_csv(StringIO(self.data1), header=None, na_values=self.na_values)
+
+
 class write_csv_standard(object):
     goal_time = 0.2
 

ci/install-3.6_DEV.sh

@@ -71,7 +71,8 @@ else
     conda config --set always_yes true --set changeps1 false || exit 1
     conda update -q conda
 
-    # add the pandas channel *before* defaults to have defaults take priority
+    # add the pandas channel to take priority
+    # to add extra packages
     echo "add channels"
     conda config --add channels pandas || exit 1
     conda config --remove channels defaults || exit 1
@@ -90,7 +91,15 @@ if [ -e ${INSTALL} ]; then
 else
 
     # create new env
-    time conda create -n pandas python=$PYTHON_VERSION nose coverage flake8 || exit 1
+    time conda create -n pandas python=$PYTHON_VERSION nose || exit 1
+
+    if [ "$COVERAGE" ]; then
+        pip install coverage
+    fi
+    if [ "$LINT" ]; then
+        conda install flake8
+        pip install cpplint
+    fi
 fi
 
 # build deps

ci/install_travis.sh

@@ -7,7 +7,6 @@ source activate pandas
 RET=0
 
 if [ "$LINT" ]; then
-    pip install cpplint
 
     # pandas/rpy is deprecated and will be removed.
     # pandas/src is C code, so no need to search there.

ci/lint.sh

@@ -3,7 +3,7 @@ pytz
 numpy=1.10*
 xlwt
 numexpr
-pytables
+pytables==3.2.2
 matplotlib
 openpyxl
 xlrd

ci/requirements-2.7-64.run

@@ -0,0 +1,7 @@
+#!/bin/bash
+
+source activate pandas
+
+echo "install 27"
+
+conda install -n pandas -c conda-forge feather-format

ci/requirements-2.7.sh

@@ -1,11 +1,12 @@
 python-dateutil
 pytz
-numpy=1.10*
+numpy
 openpyxl
 xlsxwriter
 xlrd
 xlwt
 scipy
+feather-format
 numexpr
 pytables
 matplotlib

ci/requirements-3.5-64.run

@@ -18,6 +18,4 @@ pymysql
 psycopg2
 xarray
 s3fs
-
-# incompat with conda ATM
-# beautiful-soup
+beautifulsoup4

ci/requirements-3.5.run

@@ -0,0 +1,7 @@
+#!/bin/bash
+
+source activate pandas
+
+echo "install 35"
+
+conda install -n pandas -c conda-forge feather-format

ci/requirements-3.5.sh

@@ -1,4 +1,4 @@
-dateutil
+python-dateutil
 pytz
 numpy
 cython

ci/requirements-2.7_DOC_BUILD.build renamed to ci/requirements-3.5_DOC_BUILD.build

@@ -7,7 +7,7 @@ notebook
 matplotlib
 scipy
 lxml
-beautiful-soup
+beautifulsoup4
 html5lib
 pytables
 openpyxl=1.8.5

ci/requirements-2.7_DOC_BUILD.run renamed to ci/requirements-3.5_DOC_BUILD.run

@@ -0,0 +1,7 @@
+#!/bin/bash
+
+source activate pandas
+
+echo "install DOC_BUILD"
+
+conda install -n pandas -c conda-forge feather-format

ci/requirements-3.5_DOC_BUILD.sh

@@ -13,6 +13,4 @@ jinja2
 bottleneck
 xarray
 s3fs
-
-# incompat with conda ATM
-# beautiful-soup
+beautifulsoup4

ci/requirements-3.5_OSX.run

@@ -0,0 +1,7 @@
+#!/bin/bash
+
+source activate pandas
+
+echo "install 35_OSX"
+
+conda install -n pandas -c conda-forge feather-format

ci/requirements-3.5_OSX.sh

@@ -0,0 +1,4 @@
+python-dateutil
+pytz
+numpy
+cython

ci/requirements-3.6.build

@@ -0,0 +1,4 @@
+python-dateutil
+pytz
+numpy
+scipy

ci/requirements-3.6.run

@@ -3,4 +3,4 @@
 
 .. currentmodule:: {{ module.split('.')[0] }}
 
-.. automethod:: {{ [module.split('.')[1], objname]|join('.') }}
+.. autoaccessor:: {{ (module.split('.')[1:] + [objname]) | join('.') }}

doc/_templates/autosummary/accessor.rst

@@ -3,4 +3,4 @@
 
 .. currentmodule:: {{ module.split('.')[0] }}
 
-.. autoaccessorattribute:: {{ [module.split('.')[1], objname]|join('.') }}
+.. autoaccessorattribute:: {{ (module.split('.')[1:] + [objname]) | join('.') }}

doc/_templates/autosummary/accessor_attribute.rst

@@ -3,4 +3,4 @@
 
 .. currentmodule:: {{ module.split('.')[0] }}
 
-.. autoaccessorcallable:: {{ [module.split('.')[1], objname]|join('.') }}.__call__
+.. autoaccessorcallable:: {{ (module.split('.')[1:] + [objname]) | join('.') }}.__call__

doc/_templates/autosummary/accessor_callable.rst

@@ -3,4 +3,4 @@
 
 .. currentmodule:: {{ module.split('.')[0] }}
 
-.. autoaccessormethod:: {{ [module.split('.')[1], objname]|join('.') }}
+.. autoaccessormethod:: {{ (module.split('.')[1:] + [objname]) | join('.') }}

doc/_templates/autosummary/accessor_method.rst

@@ -83,6 +83,14 @@ HDFStore: PyTables (HDF5)
    HDFStore.get
    HDFStore.select
 
+Feather
+~~~~~~~
+
+.. autosummary::
+   :toctree: generated/
+
+   read_feather
+
 SAS
 ~~~
 
@@ -1015,6 +1023,7 @@ Serialization / IO / Conversion
    DataFrame.to_excel
    DataFrame.to_json
    DataFrame.to_html
+   DataFrame.to_feather
    DataFrame.to_latex
    DataFrame.to_stata
    DataFrame.to_msgpack

doc/cheatsheet/Pandas_Cheat_Sheet.pdf

@@ -129,8 +129,7 @@ To get back to the original Series or `numpy` array, use ``Series.astype(origina
     s
     s2 = s.astype('category')
     s2
-    s3 = s2.astype('string')
-    s3
+    s2.astype(str)
     np.asarray(s2)
 
 If you have already `codes` and `categories`, you can use the :func:`~pandas.Categorical.from_codes`

doc/cheatsheet/Pandas_Cheat_Sheet.pptx

@@ -326,6 +326,23 @@
 from sphinx.ext.autosummary import Autosummary
 
 
+class AccessorDocumenter(MethodDocumenter):
+    """
+    Specialized Documenter subclass for accessors.
+    """
+
+    objtype = 'accessor'
+    directivetype = 'method'
+
+    # lower than MethodDocumenter so this is not chosen for normal methods
+    priority = 0.6
+
+    def format_signature(self):
+        # this method gives an error/warning for the accessors, therefore
+        # overriding it (accessor has no arguments)
+        return ''
+
+
 class AccessorLevelDocumenter(Documenter):
     """
     Specialized Documenter subclass for objects on accessor level (methods,
@@ -381,12 +398,17 @@ class AccessorAttributeDocumenter(AccessorLevelDocumenter, AttributeDocumenter):
     objtype = 'accessorattribute'
     directivetype = 'attribute'
 
+    # lower than AttributeDocumenter so this is not chosen for normal attributes
+    priority = 0.6
 
 class AccessorMethodDocumenter(AccessorLevelDocumenter, MethodDocumenter):
 
     objtype = 'accessormethod'
     directivetype = 'method'
 
+    # lower than MethodDocumenter so this is not chosen for normal methods
+    priority = 0.6
+
 
 class AccessorCallableDocumenter(AccessorLevelDocumenter, MethodDocumenter):
     """
@@ -483,6 +505,7 @@ def remove_flags_docstring(app, what, name, obj, options, lines):
 
 def setup(app):
     app.connect("autodoc-process-docstring", remove_flags_docstring)
+    app.add_autodocumenter(AccessorDocumenter)
     app.add_autodocumenter(AccessorAttributeDocumenter)
     app.add_autodocumenter(AccessorMethodDocumenter)
     app.add_autodocumenter(AccessorCallableDocumenter)