Merge pull request nltk#3054 from tomaarsen/tests/doctests

Resolve the last failing doctests, add doctests to CI

Author: stevenbird

.github/workflows/ci.yaml

@@ -63,7 +63,7 @@ jobs:
         id: restore-cache
         with:
           path: ~/third
-          key: third_${{ secrets.CACHE_VERSION }}
+          key: third_${{ hashFiles('tools/github_actions/third-party.sh') }}_${{ secrets.CACHE_VERSION }}
 
       - name: Download third party data
         run: |
@@ -119,10 +119,10 @@ jobs:
         uses: actions/cache@v3
         with:
           path: ~/third
-          key: third_${{ secrets.CACHE_VERSION }}
+          key: third_${{ hashFiles('tools/github_actions/third-party.sh') }}_${{ secrets.CACHE_VERSION }}
         if: runner.os == 'Linux'
 
       - name: Run pytest
         shell: bash
         run: |
-          pytest --numprocesses auto -rsx --doctest-modules nltk/test
+          pytest --numprocesses auto -rsx --doctest-modules nltk

nltk/classify/senna.py

@@ -30,9 +30,9 @@
 Note: Unit tests for this module can be found in test/unit/test_senna.py
 
 >>> from nltk.classify import Senna
->>> pipeline = Senna('/usr/share/senna-v3.0', ['pos', 'chk', 'ner'])
+>>> pipeline = Senna('/usr/share/senna-v3.0', ['pos', 'chk', 'ner'])  # doctest: +SKIP
 >>> sent = 'Dusseldorf is an international business center'.split()
->>> [(token['word'], token['chk'], token['ner'], token['pos']) for token in pipeline.tag(sent)] # doctest: +SKIP
+>>> [(token['word'], token['chk'], token['ner'], token['pos']) for token in pipeline.tag(sent)]  # doctest: +SKIP
 [('Dusseldorf', 'B-NP', 'B-LOC', 'NNP'), ('is', 'B-VP', 'O', 'VBZ'), ('an', 'B-NP', 'O', 'DT'),
 ('international', 'I-NP', 'O', 'JJ'), ('business', 'I-NP', 'O', 'NN'), ('center', 'I-NP', 'O', 'NN')]
 """
@@ -62,7 +62,7 @@ def __init__(self, senna_path, operations, encoding="utf-8"):
                 self._path = path.normpath(environ["SENNA"]) + sep
                 exe_file_2 = self.executable(self._path)
                 if not path.isfile(exe_file_2):
-                    raise OSError(
+                    raise LookupError(
                         "Senna executable expected at %s or %s but not found"
                         % (exe_file_1, exe_file_2)
                     )
@@ -115,7 +115,7 @@ def tag_sents(self, sentences):
         encoding = self._encoding
 
         if not path.isfile(self.executable(self._path)):
-            raise OSError(
+            raise LookupError(
                 "Senna executable expected at %s but not found"
                 % self.executable(self._path)
             )

nltk/corpus/reader/markdown.py

@@ -1,8 +1,6 @@
 from collections import namedtuple
 from functools import partial, wraps
 
-from yaml import safe_load
-
 from nltk.corpus.reader.api import CategorizedCorpusReader
 from nltk.corpus.reader.plaintext import PlaintextCorpusReader
 from nltk.corpus.reader.util import concat, read_blankline_block
@@ -206,6 +204,8 @@ def concatenated_view(self, reader, fileids, categories):
         )
 
     def metadata_reader(self, stream):
+        from yaml import safe_load
+
         return [
             safe_load(t.content)
             for t in self.parser.parse(stream.read())

nltk/draw/table.py

@@ -76,13 +76,8 @@ def __init__(self, master, columns, column_weights=None, cnf={}, **kw):
         :param cnf, kw: Configuration parameters for this widget.
             Use ``label_*`` to configure all labels; and ``listbox_*``
             to configure all listboxes.  E.g.:
-                >>> from nltk.draw.util import TextWidget, CanvasFrame, SpaceWidget
-                >>> from nltk.draw.table import MultiListbox
-                >>> cf = CanvasFrame(closeenough=10, width=300, height=300)
-                >>> c = cf.canvas()
-                >>> master = TextWidget(c, "hiya there", draggable=1)
-                >>> master = SpaceWidget(c, 0, 30)
-                >>> mlb = MultiListbox(master, 5, label_foreground='red')
+                >>> root = Tk()  # doctest: +SKIP
+                >>> MultiListbox(root, ["Subject", "Sender", "Date"], label_foreground='red').pack()  # doctest: +SKIP
         """
         # If columns was specified as an int, convert it to a list.
         if isinstance(columns, int):
@@ -139,7 +134,7 @@ def __init__(self, master, columns, column_weights=None, cnf={}, **kw):
             # the default listbox behavior, which scrolls):
             lb.bind("<B1-Leave>", lambda e: "break")
             # Columns can be resized by dragging them:
-            l.bind("<Button-1>", self._resize_column)
+            lb.bind("<Button-1>", self._resize_column)
 
         # Columns can be resized by dragging them.  (This binding is
         # used if they click on the grid between columns:)
@@ -305,9 +300,10 @@ def configure(self, cnf={}, **kw):
         Configure this widget.  Use ``label_*`` to configure all
         labels; and ``listbox_*`` to configure all listboxes.  E.g.:
 
-                >>> mlb = MultiListbox(master, 5)
-                >>> mlb.configure(label_foreground='red')
-                >>> mlb.configure(listbox_foreground='red')
+                >>> master = Tk()  # doctest: +SKIP
+                >>> mlb = MultiListbox(master, 5)  # doctest: +SKIP
+                >>> mlb.configure(label_foreground='red')  # doctest: +SKIP
+                >>> mlb.configure(listbox_foreground='red')  # doctest: +SKIP
         """
         cnf = dict(list(cnf.items()) + list(kw.items()))
         for (key, val) in list(cnf.items()):
@@ -592,13 +588,13 @@ class Table:
     refers to the j-th column of the i-th row.  This can be used to
     both read and write values from the table.  E.g.:
 
-        >>> table[i,j] = 'hello'
+        >>> table[i,j] = 'hello'  # doctest: +SKIP
 
     The column (j) can be given either as an index number, or as a
     column name.  E.g., the following prints the value in the 3rd row
     for the 'First Name' column:
 
-        >>> print(table[3, 'First Name'])
+        >>> print(table[3, 'First Name'])  # doctest: +SKIP
         John
 
     You can configure the colors for individual rows, columns, or

nltk/draw/util.py

@@ -80,17 +80,17 @@ class CanvasWidget(metaclass=ABCMeta):
     arguments of the form ``attribute=value``:
 
         >>> from nltk.draw.util import TextWidget
-        >>> cn = TextWidget(Canvas(), 'test', color='red')
+        >>> cn = TextWidget(Canvas(), 'test', color='red')  # doctest: +SKIP
 
     Attribute values can also be changed after a canvas widget has
     been constructed, using the ``__setitem__`` operator:
 
-        >>> cn['font'] = 'times'
+        >>> cn['font'] = 'times'  # doctest: +SKIP
 
     The current value of an attribute value can be queried using the
     ``__getitem__`` operator:
 
-        >>> cn['color']
+        >>> cn['color']  # doctest: +SKIP
         'red'
 
     For a list of the attributes supported by a type of canvas widget,