wizata-dsapi 2.0.0.dev28__tar.gz → 2.0.0.dev29__tar.gz

{wizata_dsapi-2.0.0.dev28/wizata_dsapi.egg-info → wizata_dsapi-2.0.0.dev29}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wizata_dsapi
-Version: 2.0.0.dev28
+Version: 2.0.0.dev29
 Summary: Wizata Data Science Toolkit
 Author: Wizata S.A.
 Author-email: info@wizata.com

{wizata_dsapi-2.0.0.dev28 → wizata_dsapi-2.0.0.dev29}/wizata_dsapi/plots/common.py

@@ -13,9 +13,53 @@ from .theme import (
     TEXT_MUTED, WHITE,
     TABLE_HEADER_BG, TABLE_HEADER_TEXT, TABLE_CELL_BG, TABLE_CELL_TEXT,
     TABLE_LINE_COLOR, FONT_FAMILY, FONT_SIZE,
+    COLUMN_TYPE_TELEMETRY, COLUMN_TYPE_SETPOINT, COLUMN_TYPE_RECOMMENDATION,
+    COLUMN_TYPE_TARGET, COLUMN_TYPE_CALCULATED, COLUMN_TYPES_ORDER,
 )
 
 
+def _classify_columns(df, context):
+    """Classify each dataframe column by its role in the pipeline.
+
+    Returns a dict of {col_name: {"type": str, "tag": str, "label": str, "color": str}}.
+    Type ordering: telemetry < setpoint < recommendation < calculated < target.
+    """
+    datapoints = context.datapoints or {}
+    target = context.properties.get("target_feat") if context.properties else None
+
+    # Also check models for target_feat if not in properties
+    if target is None and hasattr(context, "models") and context.models:
+        for model in context.models.values():
+            if hasattr(model, "has_target_feat") and model.has_target_feat:
+                target_candidates = getattr(model, "target_feat", None)
+                if target_candidates:
+                    target = target_candidates if isinstance(target_candidates, str) else target_candidates[0]
+                    break
+
+    classification = {}
+    for col in df.select_dtypes(include="number").columns:
+        if col == target:
+            classification[col] = {**COLUMN_TYPE_TARGET, "type": "target"}
+        elif isinstance(col, str) and col.endswith("_recommended"):
+            classification[col] = {**COLUMN_TYPE_RECOMMENDATION, "type": "recommendation"}
+        elif col in datapoints:
+            dp = datapoints[col]
+            if dp.business_type == wizata_dsapi.BusinessType.SET_POINTS:
+                classification[col] = {**COLUMN_TYPE_SETPOINT, "type": "setpoint"}
+            else:
+                classification[col] = {**COLUMN_TYPE_TELEMETRY, "type": "telemetry"}
+        else:
+            classification[col] = {**COLUMN_TYPE_CALCULATED, "type": "calculated"}
+
+    return classification
+
+
+def _sort_columns_by_type(columns, classification):
+    """Sort columns following COLUMN_TYPES_ORDER: telemetry, setpoint, recommendation, calculated, target."""
+    order_map = {t: i for i, t in enumerate(COLUMN_TYPES_ORDER)}
+    return sorted(columns, key=lambda c: order_map.get(classification.get(c, {}).get("type", "calculated"), 3))
+
+
 def check_single_column_and_target_feat(context: wizata_dsapi.Context):
     """
     check_single_column_and_target_feat
@@ -226,10 +270,10 @@ def parallel_coordinates(context: wizata_dsapi.Context):
 def setpoint_recommendation(context: wizata_dsapi.Context):
     """Compare current vs recommended setpoint values as a grouped bar chart with a summary table.
 
-    Auto-pairs columns: for every column ending in '_recommended', matches the same column name
-    without the suffix and validates via context.datapoints BusinessType.SET_POINTS. Uses the last
-    row of the dataframe (most recent state). The top half shows a horizontal grouped bar chart
-    (Current vs Recommended), the bottom half a table with Setpoint | Current | Recommended | delta | delta (%).
+    Auto-pairs columns ending in '_recommended' with the matching setpoint column, validated via
+    context.datapoints BusinessType.SET_POINTS. Uses the last row (most recent state). The top half
+    shows a horizontal grouped bar chart (Current vs Recommended) with type-colored labels; the
+    bottom half is a summary table with Setpoint | Current | Recommended | delta | delta (%).
     """
     df = context.dataframe
     datapoints = context.datapoints or {}
@@ -402,10 +446,10 @@ def feature_importance(context: wizata_dsapi.Context):
 
 
 def process_variability(context: wizata_dsapi.Context):
-    """Parallel coordinates plot showing configurations variability across process features, colored by
-    a quality/outcome column. Each line is one row (time step) of the dataframe; each vertical axis
-    is a numeric feature. The color gradient runs from red (poor quality) through yellow to green
-    (good quality) based on the 'color_by' column.
+    """Parallel coordinates plot showing process configurations variability, colored by a quality
+    outcome. Columns are ordered by type: telemetry, setpoints, recommendations, calculated,
+    with the target/quality column last. Each axis label is prefixed with a type tag
+    ([T], [SP], [REC], [C], [Q]) for quick identification.
 
     Column selection:
       - color_by: property name of the quality column used for coloring (defaults to
@@ -422,12 +466,19 @@ def process_variability(context: wizata_dsapi.Context):
     if len(numeric_cols) < 2:
         raise ValueError("process_variability: need at least 2 numeric columns")
 
+    # Classify and order columns by type
+    classification = _classify_columns(df, context)
+    ordered_cols = _sort_columns_by_type(
+        [c for c in numeric_cols if c in classification],
+        classification
+    )
+
     # Resolve color column
     color_by = context.properties.get("color_by")
     if color_by is None:
         color_by = context.properties.get("target_feat")
     if color_by is None:
-        color_by = numeric_cols[-1]
+        color_by = ordered_cols[-1] if ordered_cols else numeric_cols[-1]
     if color_by not in df.columns:
         raise ValueError(f"process_variability: color_by column '{color_by}' not found in dataframe")
 
@@ -435,14 +486,16 @@ def process_variability(context: wizata_dsapi.Context):
     if isinstance(color_reverse, str):
         color_reverse = color_reverse.lower() not in ("false", "0", "no")
 
-    # Build dimensions for all numeric columns
+    # Build dimensions with type-tagged labels, in order
     dimensions = []
-    for col in numeric_cols:
+    for col in ordered_cols:
         col_vals = df[col].dropna()
         if col_vals.empty:
             continue
+        info = classification[col]
+        tagged_label = f"[{info['tag']}] {col}"
         dimensions.append(dict(
-            label=col,
+            label=tagged_label,
             values=df[col].values,
             range=[float(col_vals.min()), float(col_vals.max())],
         ))
@@ -463,7 +516,21 @@ def process_variability(context: wizata_dsapi.Context):
         dimensions=dimensions,
     ))
 
-    fig.update_layout(title="Configurations Variability")
+    # Add a legend-like annotation for the type tags
+    type_legend = "  ".join(
+        f"[{t['tag']}] {t['label']}"
+        for t in [COLUMN_TYPE_TELEMETRY, COLUMN_TYPE_SETPOINT, COLUMN_TYPE_RECOMMENDATION,
+                  COLUMN_TYPE_CALCULATED, COLUMN_TYPE_TARGET]
+    )
+    fig.update_layout(
+        title=dict(text="Configurations Variability"),
+        annotations=[dict(
+            text=type_legend,
+            xref="paper", yref="paper", x=0, y=-0.08,
+            showarrow=False,
+            font=dict(size=10, color=TEXT_MUTED),
+        )],
+    )
     apply_theme(fig)
 
     context.set_plot(figure=fig, name="process_variability")

{wizata_dsapi-2.0.0.dev28 → wizata_dsapi-2.0.0.dev29}/wizata_dsapi/plots/theme.py

@@ -53,6 +53,16 @@ TABLE_CELL_BG = [TERTIARY, FOURTH]
 TABLE_CELL_TEXT = TEXT_MUTED
 TABLE_LINE_COLOR = "rgba(255,255,255,0.12)"
 
+# -- Column type classification -----------------------------------------------
+# Used by plots to visually distinguish column roles in the process pipeline.
+COLUMN_TYPE_TELEMETRY = {"tag": "T", "label": "Telemetry", "color": "#4992FF"}
+COLUMN_TYPE_SETPOINT = {"tag": "SP", "label": "Setpoint", "color": "#E64600"}
+COLUMN_TYPE_RECOMMENDATION = {"tag": "REC", "label": "Recommendation", "color": "#7CFFB2"}
+COLUMN_TYPE_TARGET = {"tag": "Q", "label": "Target / Quality", "color": "#FDDD60"}
+COLUMN_TYPE_CALCULATED = {"tag": "C", "label": "Calculated", "color": "#9E7CFF"}
+
+COLUMN_TYPES_ORDER = ["telemetry", "setpoint", "recommendation", "calculated", "target"]
+
 # -- Font ---------------------------------------------------------------------
 FONT_FAMILY = "Inter, -apple-system, BlinkMacSystemFont, sans-serif"
 FONT_SIZE = 12

wizata_dsapi-2.0.0.dev29/wizata_dsapi/version.py

@@ -0,0 +1 @@
+__version__ = "2.0.0.dev29"

{wizata_dsapi-2.0.0.dev28 → wizata_dsapi-2.0.0.dev29/wizata_dsapi.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: wizata_dsapi
-Version: 2.0.0.dev28
+Version: 2.0.0.dev29
 Summary: Wizata Data Science Toolkit
 Author: Wizata S.A.
 Author-email: info@wizata.com

wizata_dsapi-2.0.0.dev28/wizata_dsapi/version.py

@@ -1 +0,0 @@
-__version__ = "2.0.0.dev28"