ARROW-7892: [Python] Add FileSystemDataset.format attribute

Closes #6462 from jorisvandenbossche/dataset-various-api and squashes the following commits:

2a40709 <Joris Van den Bossche> add more informative repr to Expression
865ccb5 <Joris Van den Bossche> ARROW-7892:  Add FilesystemSource.format attribute

Authored-by: Joris Van den Bossche <jorisvandenbossche@gmail.com>
Signed-off-by: Wes McKinney <wesm+git@apache.org>

python/pyarrow/_dataset.pyx

@@ -312,6 +312,11 @@ cdef class FileSystemDataset(Dataset):
         cdef vector[c_string] files = self.filesystem_dataset.files()
         return [frombytes(f) for f in files]
 
+    @property
+    def format(self):
+        """The FileFormat of this source."""
+        return FileFormat.wrap(self.filesystem_dataset.format())
+
 
 cdef class FileFormat:
 
@@ -327,7 +332,7 @@ cdef class FileFormat:
         self.format = sp.get()
 
     @staticmethod
-    cdef wrap(shared_ptr[CFileFormat]& sp):
+    cdef wrap(const shared_ptr[CFileFormat]& sp):
         cdef FileFormat self
 
         typ = frombytes(sp.get().type_name())
@@ -1139,6 +1144,11 @@ cdef class Expression:
     def __str__(self):
         return frombytes(self.expr.ToString())
 
+    def __repr__(self):
+        return "<pyarrow.dataset.{0} {1}>".format(
+            self.__class__.__name__, str(self)
+        )
+
     def validate(self, Schema schema not None):
         """Validate this expression for execution against a schema.
 

python/pyarrow/includes/libarrow_dataset.pxd

@@ -228,6 +228,7 @@ cdef extern from "arrow/dataset/api.h" namespace "arrow::dataset" nogil:
         vector[c_string] files()
         shared_ptr[CFragmentIterator] GetFragments(
             shared_ptr[CScanOptions] options)
+        const shared_ptr[CFileFormat] format()
 
     cdef cppclass CParquetFileFormatReaderOptions \
             "arrow::dataset::ParquetFileFormat::ReaderOptions":

python/pyarrow/tests/test_dataset.py

@@ -182,6 +182,7 @@ def test_filesystem_dataset(mockfs):
         paths_or_selector=paths,
         partitions=partitions
     )
+    assert isinstance(source.format, ds.ParquetFileFormat)
 
     root_partition = ds.ComparisonExpression(
         ds.CompareOperator.Equal,
@@ -360,6 +361,7 @@ def test_expression():
     assert condition.assume(i64_is_5).equals(ds.ScalarExpression(False))
     assert condition.assume(i64_is_7).equals(ds.ScalarExpression(True))
     assert str(condition) == "(i64 > 5:int64)"
+    assert "(i64 > 5:int64)" in repr(condition)
 
 
 def test_expression_ergonomics():