Split pandas functions into separate module.

This frees //labm8:sqlutil from the dependency on
//third_party/py/pandas, which is currently broken.

github.com/ChrisCummins/phd/issues/51

[Exported from 2391f3a74286a844b98a5537a0e17c4a19ddf324]

labm8/BUILD

@@ -547,6 +547,28 @@ py_test(
     ],
 )
 
+py_library(
+    name = "pdutil",
+    srcs = ["pdutil.py"],
+    visibility = ["//visibility:public"],
+    deps = [
+        ":app",
+        ":sqlutil",
+        "//third_party/py/pandas",
+    ],
+)
+
+py_test(
+    name = "pdutil_test",
+    srcs = ["pdutil_test.py"],
+    deps = [
+        ":pdutil",
+        ":sqlutil",
+        ":test",
+        "//third_party/py/sqlalchemy",
+    ],
+)
+
 py_library(
     name = "ppar",
     srcs = ["ppar.py"],
@@ -627,7 +649,6 @@ py_library(
         ":pbutil",
         ":text",
         "//third_party/py/absl",
-        "//third_party/py/pandas",
         "//third_party/py/sqlalchemy",
     ],
 )

labm8/sqlutil.py

@@ -12,14 +12,13 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 """Utility code for working with sqlalchemy."""
+import time
+
 import collections
 import contextlib
 import pathlib
-import time
-import typing
-
-import pandas as pd
 import sqlalchemy as sql
+import typing
 from absl import flags as absl_flags
 from sqlalchemy import func
 from sqlalchemy import orm
@@ -468,45 +467,6 @@ def __repr__(self) -> str:
     return self.url
 
 
-def QueryToDataFrame(session: Session, query: Query) -> pd.DataFrame:
-  """Read query results to a Pandas DataFrame.
-
-  Args:
-    session: A database session.
-    query: The query to run.
-
-  Returns:
-    A Pandas DataFrame.
-  """
-  return pd.read_sql(query.statement, session.bind)
-
-
-def ModelToDataFrame(
-    session: Session,
-    model,
-    columns: typing.Optional[typing.List[str]] = None,
-    query_identity=lambda q: q,
-):
-  """Construct and execute a query reads an object's fields to a dataframe.
-
-  Args:
-    session: A database session.
-    model: A database mapped object.
-    columns: A list of column names, where each element is a column mapped to
-      the model. If not provided, all column names are used.
-    query_identity: A function which takes the produced query and returns a
-      query. Use this to implement filtering of the query results.
-
-  Returns:
-    A Pandas DataFrame with one column for each field.
-  """
-  columns = columns or ColumnNames(model)
-  query = session.query(*[getattr(model, column) for column in columns])
-  df = QueryToDataFrame(session, query_identity(query))
-  df.columns = columns
-  return df
-
-
 class TablenameFromClassNameMixin(object):
   """A class mixin which derives __tablename__ from the class name.