Add data validation and auto-detection debug logging

core/data_manager.py

@@ -1,19 +1,24 @@
 """Data loading, export, and column management."""
 
 import io
+import logging
 import pandas as pd
 import numpy as np
 import streamlit as st
 from core.state import set_df, get_df, get_var_types, set_var_type
 from core import constants
 
+logger = logging.getLogger(__name__)
+
 
 def load_csv(uploaded_file):
     """Load data from a CSV file."""
     try:
         df = pd.read_csv(uploaded_file, low_memory=False)
         _apply_loaded_df(df)
         return True, None
+    except ValueError as e:
+        return False, str(e)
     except Exception:
         return False, "Unable to load CSV file. Please check the format."
 
@@ -24,6 +29,8 @@ def load_excel(uploaded_file):
         df = pd.read_excel(uploaded_file)
         _apply_loaded_df(df)
         return True, None
+    except ValueError as e:
+        return False, str(e)
     except Exception:
         return False, "Unable to load Excel file. Please check the format."
 
@@ -41,12 +48,24 @@ def load_from_paste(text):
             df = pd.read_csv(io.StringIO(text), sep=";")
         _apply_loaded_df(df)
         return True, None
+    except ValueError as e:
+        return False, str(e)
     except Exception:
         return False, "Unable to parse pasted data. Please check the format."
 
 
 def _apply_loaded_df(df):
     """Apply a loaded DataFrame to session state."""
+    # Validate minimum data dimensions
+    if len(df.columns) == 0:
+        logger.warning("Malformed upload: shape=%s", df.shape)
+        raise ValueError("Uploaded data is empty or has no columns.")
+
+    if len(df) == 0:
+        logger.warning("No data rows in upload, only columns: %s", list(df.columns))
+
+    logger.debug("Loaded DataFrame: shape=%s, columns=%s", df.shape, list(df.columns))
+
     # Pad with empty rows so user can add more data
     if len(df) < 20:
         extra = pd.DataFrame(
@@ -65,6 +84,7 @@ def _auto_detect_type(col, df):
     """Auto-detect variable type for a column."""
     series = df[col].dropna()
     if len(series) == 0:
+        logger.debug("Column '%s': empty, assigned METRIC", col)
         set_var_type(col, constants.METRIC)
         return
 
@@ -75,13 +95,17 @@ def _auto_detect_type(col, df):
     if non_null_numeric / len(series) > 0.5:
         n_unique = numeric.dropna().nunique()
         if n_unique <= 2:
+            logger.debug("Column '%s': %d unique numeric values, assigned NOMINAL", col, n_unique)
             set_var_type(col, constants.NOMINAL)
         elif n_unique <= 7 and n_unique < len(series) * 0.3:
+            logger.debug("Column '%s': %d unique numeric values (sparse), assigned ORDINAL", col, n_unique)
             set_var_type(col, constants.ORDINAL)
         else:
+            logger.debug("Column '%s': %d unique numeric values, assigned METRIC", col, n_unique)
             set_var_type(col, constants.METRIC)
     else:
         # Non-numeric columns are Nominal regardless of cardinality.
+        logger.debug("Column '%s': non-numeric (%.1f%% coercible), assigned NOMINAL", col, 100 * non_null_numeric / len(series))
         set_var_type(col, constants.NOMINAL)
 
 

tests/test_data_manager.py

@@ -7,7 +7,7 @@
 import streamlit as st
 
 from core import constants
-from core.data_manager import _auto_detect_type, add_column, load_csv
+from core.data_manager import _apply_loaded_df, _auto_detect_type, add_column, load_csv
 
 
 @pytest.fixture(autouse=True)
@@ -74,6 +74,33 @@ def test_load_csv_detects_every_column():
     assert types["label"] == constants.NOMINAL
 
 
+def test_load_csv_with_header_only_still_loads_and_pads_rows():
+    # A CSV with columns but no data rows (e.g. a fresh export template) must
+    # still load so the user can type values into the padded editor rows.
+    csv = "num,grp,label\n"
+    ok, err = load_csv(io.StringIO(csv))
+
+    assert (ok, err) == (True, None)
+    df = st.session_state["df"]
+    assert list(df.columns) == ["num", "grp", "label"]
+    assert len(df) == 20
+
+
+def test_apply_loaded_df_rejects_dataframe_with_no_columns():
+    with pytest.raises(ValueError):
+        _apply_loaded_df(pd.DataFrame())
+
+
+def test_load_csv_with_no_columns_reports_specific_error_not_generic_one():
+    # A ValueError raised while applying the loaded frame must reach the
+    # caller verbatim instead of being swallowed by the catch-all Exception
+    # handler and replaced with the generic format-error message.
+    ok, err = load_csv(io.StringIO(""))
+
+    assert ok is False
+    assert err != "Unable to load CSV file. Please check the format."
+
+
 def test_add_column_marks_the_new_column_as_metric():
     add_column()