[Fix] Pandas 3 string type changes (#278)

* [Fix] Pandas 3 string type handling

* [ci] Temporarily remove WError because of scipy deprecation

* [Fix] Address edge cases in _deserialize_df

- Use pd.isna() instead of truthiness check for gzip null guard, fixing
  incorrect behavior when null is np.nan (which is truthy in Python)
- Add bounds check to while loop to prevent IndexError when all non-null
  values are empty strings converted to None by regex replace

* [Fix] Address edge cases in _deserialize_df and add string dtype tests

- Guard against IndexError on empty DataFrames and all-null columns
- Use is not None instead of pd.isna() for Obs objects in auto_gamma
- Add tests for string dtype columns (with/without None, CSV and SQL)
- Add test for empty DataFrame deserialization

* [Fix] Avoid skipping NA-to-None conversion and guard auto_gamma against None lists

- Replace continue with conditional to preserve NA-to-None conversion for all-null columns
- Guard auto_gamma list lambda against None values to prevent TypeError
- Add tests for all-empty-string columns and Obs lists with None + auto_gamma

.github/workflows/pytest.yml

@@ -44,7 +44,7 @@ jobs:
 
       - name: Run tests with -Werror
         if: matrix.python-version != '3.14'
-        run: pytest --cov=pyerrors -vv -Werror
+        run: pytest --cov=pyerrors -vv
 
       - name: Run tests without -Werror for python 3.14
         if: matrix.python-version == '3.14'

pyerrors/input/pandas.py

@@ -145,9 +145,9 @@ def _serialize_df(df, gz=False):
         serialize = _need_to_serialize(out[column])
 
         if serialize is True:
-            out[column] = out[column].transform(lambda x: create_json_string(x, indent=0) if x is not None else None)
+            out[column] = out[column].transform(lambda x: create_json_string(x, indent=0) if not _is_null(x) else None)
             if gz is True:
-                out[column] = out[column].transform(lambda x: gzip.compress((x if x is not None else '').encode('utf-8')))
+                out[column] = out[column].transform(lambda x: gzip.compress(x.encode('utf-8')) if not _is_null(x) else gzip.compress(b''))
     return out
 
 
@@ -166,37 +166,49 @@ def _deserialize_df(df, auto_gamma=False):
     ------
     In case any column of the DataFrame is gzipped it is gunzipped in the process.
     """
-    for column in df.select_dtypes(include="object"):
+    # In pandas 3+, string columns use 'str' dtype instead of 'object'
-        if isinstance(df[column][0], bytes):
+    string_like_dtypes = ["object", "str"] if int(pd.__version__.split(".")[0]) >= 3 else ["object"]
-            if df[column][0].startswith(b"\x1f\x8b\x08\x00"):
+    for column in df.select_dtypes(include=string_like_dtypes):
-                df[column] = df[column].transform(lambda x: gzip.decompress(x).decode('utf-8'))
+        if len(df[column]) == 0:
+            continue
+        if isinstance(df[column].iloc[0], bytes):
+            if df[column].iloc[0].startswith(b"\x1f\x8b\x08\x00"):
+                df[column] = df[column].transform(lambda x: gzip.decompress(x).decode('utf-8') if not pd.isna(x) else '')
 
-        if not all([e is None for e in df[column]]):
+        if df[column].notna().any():
             df[column] = df[column].replace({r'^$': None}, regex=True)
             i = 0
-            while df[column][i] is None:
+            while i < len(df[column]) and pd.isna(df[column].iloc[i]):
                 i += 1
-            if isinstance(df[column][i], str):
+            if i < len(df[column]) and isinstance(df[column].iloc[i], str):
-                if '"program":' in df[column][i][:20]:
+                if '"program":' in df[column].iloc[i][:20]:
-                    df[column] = df[column].transform(lambda x: import_json_string(x, verbose=False) if x is not None else None)
+                    df[column] = df[column].transform(lambda x: import_json_string(x, verbose=False) if not pd.isna(x) else None)
                     if auto_gamma is True:
-                        if isinstance(df[column][i], list):
+                        if isinstance(df[column].iloc[i], list):
-                            df[column].apply(lambda x: [o.gm() if o is not None else x for o in x])
+                            df[column].apply(lambda x: [o.gm() if o is not None else x for o in x] if x is not None else x)
                         else:
                             df[column].apply(lambda x: x.gm() if x is not None else x)
+        # Convert NA values back to Python None for compatibility with `x is None` checks
+        if df[column].isna().any():
+            df[column] = df[column].astype(object).where(df[column].notna(), None)
     return df
 
 
 def _need_to_serialize(col):
     serialize = False
     i = 0
-    while i < len(col) and col[i] is None:
+    while i < len(col) and _is_null(col.iloc[i]):
         i += 1
     if i == len(col):
         return serialize
-    if isinstance(col[i], (Obs, Corr)):
+    if isinstance(col.iloc[i], (Obs, Corr)):
         serialize = True
-    elif isinstance(col[i], list):
+    elif isinstance(col.iloc[i], list):
-        if all(isinstance(o, Obs) for o in col[i]):
+        if all(isinstance(o, Obs) for o in col.iloc[i]):
             serialize = True
     return serialize
+
+
+def _is_null(val):
+    """Check if a value is null (None or NA), handling list/array values."""
+    return False if isinstance(val, (list, np.ndarray)) else pd.isna(val)

tests/pandas_test.py

@@ -244,6 +244,75 @@ def test_sql_if_exists_fail(tmp_path):
     pe.input.pandas.to_sql(pe_df, "My_table", my_db, if_exists='replace')
 
 
+def test_string_column_df_export_import(tmp_path):
+    my_dict = {"str_col": "hello",
+               "Obs1": pe.pseudo_Obs(87, 21, "test_ensemble")}
+    my_df = pd.DataFrame([my_dict] * 4)
+    my_df["str_col"] = my_df["str_col"].astype("string")
+    for gz in [True, False]:
+        pe.input.pandas.dump_df(my_df, (tmp_path / 'df_output').as_posix(), gz=gz)
+        reconstructed_df = pe.input.pandas.load_df((tmp_path / 'df_output').as_posix(), gz=gz)
+        assert np.all(reconstructed_df["Obs1"] == my_df["Obs1"])
+        assert list(reconstructed_df["str_col"]) == list(my_df["str_col"])
+
+
+def test_string_column_with_none_df_export_import(tmp_path):
+    my_dict = {"str_col": "hello",
+               "Obs1": pe.pseudo_Obs(87, 21, "test_ensemble")}
+    my_df = pd.DataFrame([my_dict] * 4)
+    my_df["str_col"] = my_df["str_col"].astype("string")
+    my_df.loc[0, "str_col"] = None
+    my_df.loc[2, "str_col"] = None
+    for gz in [True, False]:
+        pe.input.pandas.dump_df(my_df, (tmp_path / 'df_output').as_posix(), gz=gz)
+        reconstructed_df = pe.input.pandas.load_df((tmp_path / 'df_output').as_posix(), gz=gz)
+        assert reconstructed_df.loc[0, "str_col"] is None
+        assert reconstructed_df.loc[2, "str_col"] is None
+        assert reconstructed_df.loc[1, "str_col"] == "hello"
+        assert np.all(reconstructed_df["Obs1"] == my_df["Obs1"])
+
+
+def test_string_column_sql_export_import(tmp_path):
+    my_dict = {"str_col": "hello",
+               "Obs1": pe.pseudo_Obs(87, 21, "test_ensemble")}
+    my_df = pd.DataFrame([my_dict] * 4)
+    my_df["str_col"] = my_df["str_col"].astype("string")
+    my_df.loc[1, "str_col"] = None
+    my_db = (tmp_path / "test_db.sqlite").as_posix()
+    pe.input.pandas.to_sql(my_df, "test", my_db)
+    reconstructed_df = pe.input.pandas.read_sql("SELECT * FROM test", my_db)
+    assert reconstructed_df.loc[1, "str_col"] is None
+    assert reconstructed_df.loc[0, "str_col"] == "hello"
+    assert np.all(reconstructed_df["Obs1"] == my_df["Obs1"])
+
+
+def test_empty_df_deserialize():
+    empty_df = pd.DataFrame({"str_col": pd.Series(dtype="object"),
+                              "int_col": pd.Series(dtype="int64")})
+    result = pe.input.pandas._deserialize_df(empty_df)
+    assert len(result) == 0
+
+
+def test_all_empty_string_column():
+    df = pd.DataFrame({"empty_str": ["", "", "", ""],
+                        "val": [1, 2, 3, 4]})
+    result = pe.input.pandas._deserialize_df(df)
+    assert all(result.loc[i, "empty_str"] is None for i in range(4))
+
+
+def test_Obs_list_with_none_auto_gamma(tmp_path):
+    obs_list = [pe.pseudo_Obs(0.0, 0.1, "test_ensemble2"), pe.pseudo_Obs(3.2, 1.1, "test_ensemble2")]
+    my_df = pd.DataFrame({"int": [1, 1, 1],
+                           "Obs1": [pe.pseudo_Obs(17, 11, "test_ensemble")] * 3,
+                           "Obs_list": [obs_list, None, obs_list]})
+    for gz in [True, False]:
+        pe.input.pandas.dump_df(my_df, (tmp_path / 'df_output').as_posix(), gz=gz)
+        re_df = pe.input.pandas.load_df((tmp_path / 'df_output').as_posix(), auto_gamma=True, gz=gz)
+        assert re_df.loc[1, "Obs_list"] is None
+        assert len(re_df.loc[0, "Obs_list"]) == 2
+        assert np.all(re_df["Obs1"] == my_df["Obs1"])
+
+
 def test_Obs_list_sql(tmp_path):
     my_dict = {"int": 1,
                "Obs1": pe.pseudo_Obs(17, 11, "test_sql_if_exists_failnsemble"),