scitex 2.7.3__py3-none-any.whl → 2.10.0__py3-none-any.whl

scitex/fts/_fig/_backend/_render.py

@@ -0,0 +1,538 @@
+#!/usr/bin/env python3
+# File: ./src/scitex/vis/backend/render.py
+"""Render figure models to matplotlib figures using scitex.plt."""
+
+from typing import Any, Dict
+
+import numpy as np
+
+import scitex.logging as logging
+
+from .._models import AnnotationModel, AxesModel, FigureModel, GuideModel, PlotModel
+from ._parser import parse_figure_json
+
+logger = logging.getLogger(__name__)
+
+
+def render_figure(fig_model: FigureModel):
+    """
+    Render FigureModel to matplotlib figure using scitex.plt.
+
+    Parameters
+    ----------
+    fig_model : FigureModel
+        Figure model to render
+
+    Returns
+    -------
+    tuple
+        (fig, axes) where fig is FigWrapper and axes is list of AxisWrapper
+
+    Examples
+    --------
+    >>> fig_model = FigureModel(width_mm=180, height_mm=120)
+    >>> fig, axes = render_figure(fig_model)
+    """
+    import scitex as stx
+
+    # Validate before rendering
+    fig_model.validate()
+
+    # Convert mm to inches for matplotlib
+    MM_TO_INCH = 1 / 25.4
+    figsize_inch = (
+        fig_model.width_mm * MM_TO_INCH,
+        fig_model.height_mm * MM_TO_INCH,
+    )
+
+    # Create figure with mm dimensions
+    fig, axes = stx.plt.subplots(
+        nrows=fig_model.nrows,
+        ncols=fig_model.ncols,
+        figsize=figsize_inch,
+        dpi=fig_model.dpi,
+        facecolor=fig_model.facecolor,
+        edgecolor=fig_model.edgecolor,
+    )
+
+    # Ensure axes is always a list
+    if not isinstance(axes, (list, np.ndarray)):
+        axes = [axes]
+    elif isinstance(axes, np.ndarray):
+        axes = axes.flatten().tolist()
+
+    # TODO: Apply mm-based spacing (left_mm, right_mm, etc.)
+    # Currently, matplotlib subplots_adjust uses figure coordinates (0-1), not mm
+    # Future: Pass margin_left_mm, margin_right_mm etc. to stx.plt.subplots()
+    # to leverage scitex.plt's mm-control system
+
+    # Add suptitle if specified
+    if fig_model.suptitle:
+        suptitle_kwargs = {"t": fig_model.suptitle}
+        if fig_model.suptitle_fontsize is not None:
+            suptitle_kwargs["fontsize"] = fig_model.suptitle_fontsize
+        if fig_model.suptitle_fontweight is not None:
+            suptitle_kwargs["fontweight"] = fig_model.suptitle_fontweight
+        if fig_model.suptitle_y is not None:
+            suptitle_kwargs["y"] = fig_model.suptitle_y
+        fig.suptitle(**suptitle_kwargs)
+
+    # Render each axes
+    for i, axes_config in enumerate(fig_model.axes):
+        if i < len(axes):
+            axes_model = AxesModel.from_dict(axes_config)
+            render_axes(axes[i], axes_model)
+
+    return fig, axes
+
+
+def render_axes(ax, axes_model: AxesModel):
+    """
+    Render AxesModel onto a matplotlib axis.
+
+    Parameters
+    ----------
+    ax : AxisWrapper or matplotlib Axes
+        Target axis
+    axes_model : AxesModel
+        Axes model to render
+    """
+    # Validate
+    axes_model.validate()
+
+    # Render plots first (background layer)
+    for plot_config in axes_model.plots:
+        plot_model = PlotModel.from_dict(plot_config)
+        render_plot(ax, plot_model)
+
+    # Render guides (middle layer - reference lines/spans over plots)
+    # Note: Use zorder in guide config to control layering if needed
+    for guide_config in axes_model.guides:
+        guide_model = GuideModel.from_dict(guide_config)
+        render_guide(ax, guide_model)
+
+    # Render annotations last (foreground layer)
+    for annotation_config in axes_model.annotations:
+        annotation_model = AnnotationModel.from_dict(annotation_config)
+        render_annotation(ax, annotation_model)
+
+    # Apply axis properties
+    if axes_model.xlabel:
+        ax.set_xlabel(axes_model.xlabel)
+
+    if axes_model.ylabel:
+        ax.set_ylabel(axes_model.ylabel)
+
+    if axes_model.title:
+        ax.set_title(axes_model.title)
+
+    # Set axis limits
+    if axes_model.xlim:
+        ax.set_xlim(axes_model.xlim)
+
+    if axes_model.ylim:
+        ax.set_ylim(axes_model.ylim)
+
+    # Set axis scales
+    ax.set_xscale(axes_model.xscale)
+    ax.set_yscale(axes_model.yscale)
+
+    # Ticks
+    if axes_model.xticks is not None:
+        ax.set_xticks(axes_model.xticks)
+
+    if axes_model.yticks is not None:
+        ax.set_yticks(axes_model.yticks)
+
+    if axes_model.xticklabels is not None:
+        ax.set_xticklabels(axes_model.xticklabels)
+
+    if axes_model.yticklabels is not None:
+        ax.set_yticklabels(axes_model.yticklabels)
+
+    # Apply style properties
+    style = axes_model.style
+
+    # Grid
+    if style.grid:
+        ax.grid(
+            alpha=style.grid_alpha,
+            linestyle=style.grid_linestyle,
+            linewidth=style.grid_linewidth,
+        )
+
+    # Spines
+    for spine, visible in style.spines_visible.items():
+        if hasattr(ax, "spines") and spine in ax.spines:
+            ax.spines[spine].set_visible(visible)
+
+    # Tick label size
+    if style.tick_fontsize is not None:
+        ax.tick_params(labelsize=style.tick_fontsize)
+
+    # Legend
+    if style.legend:
+        legend_kwargs = {
+            "loc": style.legend_loc,
+            "frameon": style.legend_frameon,
+        }
+        if style.legend_fontsize is not None:
+            legend_kwargs["fontsize"] = style.legend_fontsize
+        ax.legend(**legend_kwargs)
+
+    # Aspect
+    if style.aspect:
+        ax.set_aspect(style.aspect)
+
+    # Background color
+    if style.facecolor:
+        ax.set_facecolor(style.facecolor)
+
+
+def render_plot(ax, plot_model: PlotModel):
+    """
+    Render PlotModel onto an axis.
+
+    Parameters
+    ----------
+    ax : AxisWrapper or matplotlib Axes
+        Target axis
+    plot_model : PlotModel
+        Plot model to render
+    """
+    # Validate
+    plot_model.validate()
+
+    data = plot_model.data
+
+    # Build kwargs from style (clean separation of concerns)
+    # Only include relevant kwargs for each plot type
+    style_dict = plot_model.style.to_dict()
+
+    # Common kwargs for all plot types
+    common_keys = {"color", "alpha", "zorder"}
+
+    # Define relevant keys for each plot type
+    plot_type_keys = {
+        "line": {"linewidth", "linestyle", "marker", "markersize"},
+        "scatter": {"marker", "markersize"},
+        "errorbar": {"linewidth", "linestyle", "marker", "xerr", "yerr", "capsize"},
+        "bar": {"width"},
+        "barh": {"width"},  # Will be renamed to height
+        "hist": {"bins", "density", "cumulative"},
+        "fill_between": {"linewidth", "fill_color", "fill_alpha"},
+        "heatmap": {"cmap", "vmin", "vmax", "interpolation"},
+        "imshow": {"cmap", "vmin", "vmax", "interpolation"},
+        "contour": {"cmap", "vmin", "vmax"},
+        "contourf": {"cmap", "vmin", "vmax"},
+    }
+
+    # Get relevant keys for this plot type
+    relevant_keys = common_keys | plot_type_keys.get(plot_model.plot_type, set())
+
+    # Filter kwargs to only include relevant ones
+    kwargs = {k: v for k, v in style_dict.items() if k in relevant_keys}
+
+    # Add label if present
+    if plot_model.label:
+        kwargs["label"] = plot_model.label
+
+    # Merge extra kwargs
+    kwargs.update(plot_model.extra_kwargs)
+
+    # Render based on plot type
+    if plot_model.plot_type == "line":
+        ax.plot(data["x"], data["y"], **kwargs)
+
+    elif plot_model.plot_type == "scatter":
+        # Rename markersize -> s for scatter
+        if "markersize" in kwargs:
+            kwargs["s"] = kwargs.pop("markersize")
+        ax.scatter(data["x"], data["y"], **kwargs)
+
+    elif plot_model.plot_type == "errorbar":
+        ax.errorbar(data["x"], data["y"], **kwargs)
+
+    elif plot_model.plot_type == "bar":
+        height = data.get("height", data.get("y"))
+        ax.bar(data["x"], height, **kwargs)
+
+    elif plot_model.plot_type == "barh":
+        # Rename width -> height for horizontal bars
+        if "width" in kwargs:
+            kwargs["height"] = kwargs.pop("width")
+        width = data.get("width", data.get("x"))
+        ax.barh(data["y"], width, **kwargs)
+
+    elif plot_model.plot_type == "hist":
+        ax.hist(data["x"], **kwargs)
+
+    elif plot_model.plot_type == "fill_between":
+        # Use fill_alpha and fill_color if present
+        if "fill_alpha" in kwargs:
+            kwargs["alpha"] = kwargs.pop("fill_alpha")
+        if "fill_color" in kwargs:
+            kwargs["color"] = kwargs.pop("fill_color")
+        ax.fill_between(data["x"], data["y1"], data["y2"], **kwargs)
+
+    elif plot_model.plot_type in ["heatmap", "imshow"]:
+        img_data = data.get("z", data.get("img"))
+        ax.imshow(img_data, **kwargs)
+
+    elif plot_model.plot_type == "contour":
+        ax.contour(data["x"], data["y"], data["z"], **kwargs)
+
+    elif plot_model.plot_type == "contourf":
+        ax.contourf(data["x"], data["y"], data["z"], **kwargs)
+
+    else:
+        raise ValueError(f"Unsupported plot type: {plot_model.plot_type}")
+
+
+def render_guide(ax, guide_model: GuideModel):
+    """
+    Render GuideModel onto an axis.
+
+    Parameters
+    ----------
+    ax : AxisWrapper or matplotlib Axes
+        Target axis
+    guide_model : GuideModel
+        Guide model to render
+    """
+    # Validate
+    guide_model.validate()
+
+    # Build kwargs from style
+    kwargs = guide_model.style.to_dict()
+
+    # Add label if present
+    if guide_model.label:
+        kwargs["label"] = guide_model.label
+
+    # Render based on guide type
+    if guide_model.guide_type == "axhline":
+        ax.axhline(guide_model.y, **kwargs)
+
+    elif guide_model.guide_type == "axvline":
+        ax.axvline(guide_model.x, **kwargs)
+
+    elif guide_model.guide_type == "axhspan":
+        ax.axhspan(guide_model.ymin, guide_model.ymax, **kwargs)
+
+    elif guide_model.guide_type == "axvspan":
+        ax.axvspan(guide_model.xmin, guide_model.xmax, **kwargs)
+
+
+def render_annotation(ax, annotation_model: AnnotationModel):
+    """
+    Render AnnotationModel onto an axis.
+
+    Parameters
+    ----------
+    ax : AxisWrapper or matplotlib Axes
+        Target axis
+    annotation_model : AnnotationModel
+        Annotation model to render
+    """
+    # Validate
+    annotation_model.validate()
+
+    # Build kwargs from style
+    kwargs = annotation_model.style.to_dict()
+
+    # Render based on annotation type
+    if annotation_model.annotation_type == "text":
+        ax.text(
+            annotation_model.x,
+            annotation_model.y,
+            annotation_model.text,
+            **kwargs,
+        )
+
+    elif annotation_model.annotation_type == "annotate":
+        ax.annotate(
+            annotation_model.text,
+            xy=(annotation_model.x, annotation_model.y),
+            xytext=annotation_model.xytext,
+            **kwargs,
+        )
+
+
+def build_figure_from_json(fig_json: Dict[str, Any]):
+    """
+    Build matplotlib figure from figure JSON.
+
+    This is the main entry point for rendering.
+
+    Parameters
+    ----------
+    fig_json : Dict[str, Any]
+        Figure JSON specification
+
+    Returns
+    -------
+    tuple
+        (fig, axes) where fig is FigWrapper and axes is list of AxisWrapper
+
+    Examples
+    --------
+    >>> fig_json = {...}
+    >>> fig, axes = build_figure_from_json(fig_json)
+    >>> import scitex as stx
+    >>> stx.io.save(fig, "output.png")
+    """
+    fig_model = parse_figure_json(fig_json)
+    return render_figure(fig_model)
+
+
+def _find_column(data, col_name):
+    """Find a column in data, supporting prefixed column names.
+
+    Parameters
+    ----------
+    data : pd.DataFrame
+        Data to search
+    col_name : str
+        Column name to find (can be simple like "x" or prefixed)
+
+    Returns
+    -------
+    str or None
+        Actual column name in data, or None if not found
+    """
+    if col_name in data.columns:
+        return col_name
+
+    # Try to find a column that ends with the variable name
+    # e.g., "ax-row-0-col-0_trace-id-plot-0_variable-x" matches "x"
+    suffix = f"_variable-{col_name}"
+    for col in data.columns:
+        if col.endswith(suffix):
+            return col
+
+    # Try partial match at the end
+    for col in data.columns:
+        if col.endswith(f"-{col_name}"):
+            return col
+
+    return None
+
+
+def _is_special_trace(trace_id):
+    """Check if trace is a special type that doesn't require x/y encoding.
+
+    Parameters
+    ----------
+    trace_id : str
+        Trace identifier
+
+    Returns
+    -------
+    bool
+        True if this is a special trace type
+    """
+    special_patterns = [
+        "fill-between",
+        "fill_between",
+        "axhline",
+        "axvline",
+        "axhspan",
+        "axvspan",
+        "text",
+        "annotation",
+    ]
+    trace_lower = trace_id.lower()
+    return any(pattern in trace_lower for pattern in special_patterns)
+
+
+def render_traces(ax, trace, data, theme=None):
+    """
+    Render a TraceEncoding onto an axis.
+
+    Parameters
+    ----------
+    ax : matplotlib Axes
+        Target axis
+    trace : TraceEncoding
+        Trace encoding specification
+    data : pd.DataFrame or None
+        Data to plot
+    theme : Theme, optional
+        Theme for styling
+
+    Raises
+    ------
+    ValueError
+        If data is None or required columns are missing (for standard traces)
+    """
+    # Skip special traces that don't require standard x/y encoding
+    if _is_special_trace(trace.trace_id):
+        logger.debug(f"Skipping special trace '{trace.trace_id}' (no x/y rendering)")
+        return
+
+    if data is None:
+        logger.warning(f"No data provided for trace '{trace.trace_id}', skipping")
+        return
+
+    # Get style from theme
+    color = None
+    linewidth = 1.0
+    alpha = 1.0
+
+    if theme:
+        # Check for trace-specific theme
+        if hasattr(theme, "traces") and theme.traces:
+            for t in theme.traces:
+                if t.trace_id == trace.trace_id:
+                    color = t.color
+                    if t.line_width_pt:
+                        linewidth = t.line_width_pt
+                    if t.alpha:
+                        alpha = t.alpha
+                    break
+
+        # Fall back to default colors
+        if not color and hasattr(theme, "colors"):
+            palette = theme.colors.palette
+            if palette:
+                color = palette[0]
+
+        if hasattr(theme, "lines"):
+            linewidth = theme.lines.width_pt
+
+    # Default color if none set
+    if not color:
+        color = "#1f77b4"
+
+    # Get column names from encoding
+    x_col_name = trace.x.column if trace.x else None
+    y_col_name = trace.y.column if trace.y else None
+
+    if not x_col_name or not y_col_name:
+        logger.warning(f"Trace '{trace.trace_id}' missing x or y encoding, skipping")
+        return
+
+    # Find actual columns in data (handles prefixed names)
+    x_col = _find_column(data, x_col_name)
+    y_col = _find_column(data, y_col_name)
+
+    if x_col is None:
+        logger.warning(f"Column '{x_col_name}' not found in data for trace '{trace.trace_id}'. Available: {list(data.columns)[:5]}...")
+        return
+    if y_col is None:
+        logger.warning(f"Column '{y_col_name}' not found in data for trace '{trace.trace_id}'. Available: {list(data.columns)[:5]}...")
+        return
+
+    # Plot the data
+    ax.plot(
+        data[x_col],
+        data[y_col],
+        color=color,
+        linewidth=linewidth,
+        alpha=alpha,
+        label=trace.trace_id,
+    )
+
+
+# EOF