How to use the shap.common.convert_name function in shap

interaction_index = approximate_interactions(ind, shap_values, features)[0]
    interaction_index = convert_name(interaction_index, shap_values, feature_names)
    categorical_interaction = False

    # create a matplotlib figure, if `ax` hasn't been specified.
    if not ax:
        figsize = (7.5, 5) if interaction_index != ind else (6, 5)
        fig = pl.figure(figsize=figsize)
        ax = fig.gca()
    else:
        fig = ax.get_figure()

    # plotting SHAP interaction values
    if len(shap_values.shape) == 3 and len(ind) == 2:
        ind1 = convert_name(ind[0], shap_values, feature_names)
        ind2 = convert_name(ind[1], shap_values, feature_names)
        if ind1 == ind2:
            proj_shap_values = shap_values[:, ind2, :]
        else:
            proj_shap_values = shap_values[:, ind2, :] * 2  # off-diag values are split in half

        # TODO: remove recursion; generally the functions should be shorter for more maintainable code
        dependence_plot(
            ind1, proj_shap_values, features, feature_names=feature_names,
            interaction_index=(None if ind1 == ind2 else ind2), display_features=display_features, ax=ax, show=False,
            xmin=xmin, xmax=xmax, x_jitter=x_jitter, alpha=alpha
        )
        if ind1 == ind2:
            ax.set_ylabel(labels['MAIN_EFFECT'] % feature_names[ind1])
        else:
            ax.set_ylabel(labels['INTERACTION_EFFECT'] % (feature_names[ind1], feature_names[ind2]))

)
            stemlines.set_edgecolors([shap.plots.colors.red_rgb if v > 0 else shap.plots.colors.blue_rgb for v in vals])
            pl.setp(stemlines, 'zorder', -1)
            pl.setp(stemlines, 'linewidth', 2)
            pl.setp(markerline, 'color', "black")
            pl.setp(markerline, 'markersize', 4)

        if show:
            pl.show()
        else:
            return fig,ax1
        
        
    # this is for a 2D partial dependence plot
    else:
        ind0 = convert_name(ind[0], None, feature_names)
        ind1 = convert_name(ind[1], None, feature_names)
        xv0 = features[:,ind0]
        xv1 = features[:,ind1]
        
        xmin0 = xmin[0] if type(xmin) is tuple else xmin
        xmin1 = xmin[1] if type(xmin) is tuple else xmin
        xmax0 = xmax[0] if type(xmax) is tuple else xmax
        xmax1 = xmax[1] if type(xmax) is tuple else xmax
        
        xmin0, xmax0 = compute_bounds(xmin0, xmax0, xv0)
        xmin1, xmax1 = compute_bounds(xmin1, xmax1, xv1)
        npoints = 20 if npoints is None else npoints
        xs0 = np.linspace(xmin0, xmax0, npoints)
        xs1 = np.linspace(xmin1, xmax1, npoints)
        
        features_tmp = features.copy()

stemlines.set_edgecolors([shap.plots.colors.red_rgb if v > 0 else shap.plots.colors.blue_rgb for v in vals])
            pl.setp(stemlines, 'zorder', -1)
            pl.setp(stemlines, 'linewidth', 2)
            pl.setp(markerline, 'color', "black")
            pl.setp(markerline, 'markersize', 4)

        if show:
            pl.show()
        else:
            return fig,ax1
        
        
    # this is for a 2D partial dependence plot
    else:
        ind0 = convert_name(ind[0], None, feature_names)
        ind1 = convert_name(ind[1], None, feature_names)
        xv0 = features[:,ind0]
        xv1 = features[:,ind1]
        
        xmin0 = xmin[0] if type(xmin) is tuple else xmin
        xmin1 = xmin[1] if type(xmin) is tuple else xmin
        xmax0 = xmax[0] if type(xmax) is tuple else xmax
        xmax1 = xmax[1] if type(xmax) is tuple else xmax
        
        xmin0, xmax0 = compute_bounds(xmin0, xmax0, xv0)
        xmin1, xmax1 = compute_bounds(xmin1, xmax1, xv1)
        npoints = 20 if npoints is None else npoints
        xs0 = np.linspace(xmin0, xmax0, npoints)
        xs1 = np.linspace(xmin1, xmax1, npoints)
        
        features_tmp = features.copy()
        x0 = np.zeros((npoints, npoints))

if feature_names is None:
        feature_names = [labels['FEATURE'] % str(i) for i in range(shap_values.shape[1])]

    # allow vectors to be passed
    if len(shap_values.shape) == 1:
        shap_values = np.reshape(shap_values, len(shap_values), 1)
    if len(features.shape) == 1:
        features = np.reshape(features, len(features), 1)

    ind = convert_name(ind, shap_values, feature_names)

    # guess what other feature as the stongest interaction with the plotted feature
    if interaction_index == "auto":
        interaction_index = approximate_interactions(ind, shap_values, features)[0]
    interaction_index = convert_name(interaction_index, shap_values, feature_names)
    categorical_interaction = False

    # create a matplotlib figure, if `ax` hasn't been specified.
    if not ax:
        figsize = (7.5, 5) if interaction_index != ind else (6, 5)
        fig = pl.figure(figsize=figsize)
        ax = fig.gca()
    else:
        fig = ax.get_figure()

    # plotting SHAP interaction values
    if len(shap_values.shape) == 3 and len(ind) == 2:
        ind1 = convert_name(ind[0], shap_values, feature_names)
        ind2 = convert_name(ind[1], shap_values, feature_names)
        if ind1 == ind2:
            proj_shap_values = shap_values[:, ind2, :]

if interaction_index == "auto":
        interaction_index = approximate_interactions(ind, shap_values, features)[0]
    interaction_index = convert_name(interaction_index, shap_values, feature_names)
    categorical_interaction = False

    # create a matplotlib figure, if `ax` hasn't been specified.
    if not ax:
        figsize = (7.5, 5) if interaction_index != ind else (6, 5)
        fig = pl.figure(figsize=figsize)
        ax = fig.gca()
    else:
        fig = ax.get_figure()

    # plotting SHAP interaction values
    if len(shap_values.shape) == 3 and len(ind) == 2:
        ind1 = convert_name(ind[0], shap_values, feature_names)
        ind2 = convert_name(ind[1], shap_values, feature_names)
        if ind1 == ind2:
            proj_shap_values = shap_values[:, ind2, :]
        else:
            proj_shap_values = shap_values[:, ind2, :] * 2  # off-diag values are split in half

        # TODO: remove recursion; generally the functions should be shorter for more maintainable code
        dependence_plot(
            ind1, proj_shap_values, features, feature_names=feature_names,
            interaction_index=(None if ind1 == ind2 else ind2), display_features=display_features, ax=ax, show=False,
            xmin=xmin, xmax=xmax, x_jitter=x_jitter, alpha=alpha
        )
        if ind1 == ind2:
            ax.set_ylabel(labels['MAIN_EFFECT'] % feature_names[ind1])
        else:
            ax.set_ylabel(labels['INTERACTION_EFFECT'] % (feature_names[ind1], feature_names[ind2]))

The names of the features in the shap_values array.

    method : "pca" or numpy.array
        How to reduce the dimensions of the shap_values to 2D. If "pca" then the 2D
        PCA projection of shap_values is used. If a numpy array then is should be
        (# samples x 2) and represent the embedding of that values. 

    alpha : float
        The transparency of the data points (between 0 and 1). This can be useful to the
        show density of the data points when using a large dataset.
    """
    
    if feature_names is None:
        feature_names = [labels['FEATURE'] % str(i) for i in range(shap_values.shape[1])]
    
    ind = convert_name(ind, shap_values, feature_names)
    if ind == "sum()":
        cvals = shap_values.sum(1)
        fname = "sum(SHAP values)"
    else:
        cvals = shap_values[:,ind]
        fname = feature_names[ind]
    
    # see if we need to compute the embedding
    if type(method) == str and method == "pca":
        pca = sklearn.decomposition.PCA(2)
        embedding_values = pca.fit_transform(shap_values)
    elif hasattr(method, "shape") and method.shape[1] == 2:
        embedding_values = method
    else:
        print("Unsupported embedding method:", method)