From a451d7af282dbcf111a423f0d0d3f6375ac4f2a4 Mon Sep 17 00:00:00 2001
From: Yan Wong <yan.wong@bdi.ox.ac.uk>
Date: Thu, 2 Mar 2023 15:22:21 +0000
Subject: [PATCH] str(null_schema) returns "Null_schema" not None

---
 python/CHANGELOG.rst        |  6 ++++--
 python/tests/test_tables.py | 19 +++++++++++++++----
 python/tskit/metadata.py    |  9 ++++++++-
 3 files changed, 27 insertions(+), 7 deletions(-)

diff --git a/python/CHANGELOG.rst b/python/CHANGELOG.rst
index 1fd56f3b01..c12283b522 100644
--- a/python/CHANGELOG.rst
+++ b/python/CHANGELOG.rst
@@ -24,6 +24,9 @@
   in the presence of non-ancestral material (very rare).
   (:user:`petrelharp`, :issue:`2983`, :pr:`1623`)
 
+- Printing ``tskit.MetadataSchema(schema=None)`` now shows ``"Null_schema"`` rather
+  than ``None``, to avoid confusion (:user:`hyanwong`, :pr:`2720`)
+
 **Features**
 
 - Add ``TreeSequence.extend_haplotypes`` method that extends ancestral haplotypes
@@ -103,8 +106,7 @@
 - Add ``__repr__`` for variants to return a string representation of the raw data
   without spewing megabytes of text (:user:`chriscrsmith`, :pr:`2695`, :issue:`2694`)
 
-- Add ``keep_rows`` method to table classes to support efficient in-place
-  table subsetting (:user:`jeromekelleher`, :pr:`2700`)
+**Breaking Changes**
 
 **Bugfixes**
 
diff --git a/python/tests/test_tables.py b/python/tests/test_tables.py
index 7551f3250c..d283f458d8 100644
--- a/python/tests/test_tables.py
+++ b/python/tests/test_tables.py
@@ -1396,9 +1396,10 @@ def test_metadata_schema(self, table_5row):
             table2.metadata_schema = tskit.MetadataSchema({"codec": "json"})
             with pytest.raises(
                 AssertionError,
-                match=f"{type(table_5row).__name__} metadata schemas differ:",
+                match=f"{type(table_5row).__name__} metadata schemas differ: "
+                "self=Null_schema other=OrderedDict([('codec', 'json')])",
             ):
-                table_5row.assert_equals(table2)
+                 table_5row.assert_equals(table2)
             table_5row.assert_equals(table2, ignore_metadata=True)
 
     def test_row_changes(self, table_5row, test_rows):
@@ -4089,7 +4090,10 @@ def test_metadata_schema(self, t1, t2):
         t2.metadata_schema = tskit.MetadataSchema(None)
         with pytest.raises(
             AssertionError,
-            match=re.escape("Metadata schemas differ"),
+            match=re.escape(
+                "Metadata schemas differ: self=OrderedDict([('codec', 'json')]) "
+                "other=Null_schema"
+            ),
         ):
             t1.assert_equals(t2)
         t1.assert_equals(t2, ignore_metadata=True)
@@ -4136,6 +4140,10 @@ def test_ignore_metadata(self, t1, t2, table_name):
             with pytest.raises(
                 AssertionError,
                 match=re.escape(f"{type(table).__name__} metadata schemas differ:"),
+                match=re.escape(
+                    f"{type(table).__name__} metadata schemas differ: "
+                    f"self=OrderedDict([('codec', 'json')]) other=Null_schema"
+                ),
             ):
                 t1.assert_equals(t2)
             t1.assert_equals(t2, ignore_metadata=True)
@@ -4193,7 +4201,10 @@ def test_ignore_reference_sequence(self, t1, t2):
         t2.reference_sequence.clear()
         with pytest.raises(
             AssertionError,
-            match=re.escape("Metadata schemas differ"),
+            match=re.escape(
+                "Metadata schemas differ: "
+                "self=OrderedDict([('codec', 'json')]) other=Null_schema"
+            ),
         ):
             t1.assert_equals(t2)
         t1.assert_equals(t2, ignore_reference_sequence=True)
diff --git a/python/tskit/metadata.py b/python/tskit/metadata.py
index f9e8c0c6c7..781daf5ee3 100644
--- a/python/tskit/metadata.py
+++ b/python/tskit/metadata.py
@@ -664,7 +664,14 @@ def __repr__(self) -> str:
         return self._string
 
     def __str__(self) -> str:
-        return pprint.pformat(self._schema)
+        if isinstance(self._schema, collections.OrderedDict):
+            s = pprint.pformat(dict(self._schema))
+        else:
+            s = pprint.pformat(self._schema)
+        if "\n" in s:
+            return f"MetadataSchema(\n{s}\n)"
+        else:
+            return f"MetadataSchema({s})"
 
     def __eq__(self, other) -> bool:
         return self._string == other._string