Api cpdb

def cpdb_network_cal(adata:anndata.AnnData,pvals:list,celltype_key:str)->dict:
    r"""
    Calculate a CPDB (Cell Phone Database) network using gene expression data and return a dictionary of results.

    Arguments:
        adata: An AnnData object containing gene expression data.
        pvals: A list or array of p-values for each gene in the expression data.
        celltype_key: The name of the cell type key in adata.obs.

    Returns:
        cpdb_dict: A dictionary of results generated by `kpy.plot_cpdb_heatmap()` from the `ktplotspy` library.

    """
    check_kpy()
    global kpy_install
    if kpy_install==True:
        global_imports("ktplotspy","kpy")

    cpdb_dict=kpy.plot_cpdb_heatmap(
        adata = adata,
        pvals = pvals,
        celltype_key = celltype_key,
        figsize = (1,1),
        title = "",
        symmetrical = False,
        return_tables=True,
    )
    return cpdb_dict

def cpdb_plot_network(adata:anndata.AnnData,interaction_edges:pd.DataFrame,
                      celltype_key:str,nodecolor_dict=None,
                      edgeswidth_scale:int=10,nodesize_scale:int=1,
                      pos_scale:int=1,pos_size:int=10,figsize:tuple=(5,5),title:str='',
                      legend_ncol:int=3,legend_bbox:tuple=(1,0.2),legend_fontsize:int=10,
                     return_graph:bool=False)->nx.Graph:
    r"""
    Plot a network of interactions between cell types using gene expression data.

    Arguments:
        adata: An AnnData object containing gene expression data.
        interaction_edges: A DataFrame containing the edges of the network, including the source, target, and count of interactions.
        celltype_key: The name of the cell type key in adata.obs.
        nodecolor_dict: A dictionary mapping cell type names to node colors. If not provided, uses the colors specified in adata.uns.
        edgeswidth_scale: The scaling factor for edge width. Default is 10.
        nodesize_scale: The scaling factor for node size. Default is 1.
        pos_scale: The scaling factor for node positions. Default is 1.
        pos_size: The size of the node positions. Default is 10.
        figsize: The size of the plot. Default is (5,5).
        title: The title of the plot. Default is ''.
        legend_ncol: The number of columns in the legend. Default is 3.
        legend_bbox: The location of the legend. Default is (1,0.2).
        legend_fontsize: The font size of the legend. Default is 10.
        return_graph: If True, returns the network graph. Default is False.

    Returns:
        ax or G: The plot if `return_graph` is False, otherwise the network graph.

    """
    check_kpy()
    global kpy_install
    if kpy_install==True:
        global_imports("ktplotspy","kpy")

    #set Digraph of cellphonedb
    G=nx.DiGraph()
    for i in interaction_edges.index:
        G.add_edge(interaction_edges.loc[i,'SOURCE'],
                   interaction_edges.loc[i,'TARGET'],
                   weight=interaction_edges.loc[i,'COUNT'],)

    #set celltypekey's color
    if nodecolor_dict!=None:
        type_color_all=nodecolor_dict
    else:
        if '{}_colors'.format(celltype_key) in adata.uns:
            type_color_all=dict(zip(adata.obs[celltype_key].cat.categories,adata.uns['{}_colors'.format(celltype_key)]))
        else:
            if len(adata.obs[celltype_key].cat.categories)>28:
                type_color_all=dict(zip(adata.obs[celltype_key].cat.categories,sc.pl.palettes.default_102))
            else:
                type_color_all=dict(zip(adata.obs[celltype_key].cat.categories,sc.pl.palettes.zeileis_28))

    #set G_nodes_dict
    nodes=[]
    G_degree=dict(G.degree(G.nodes()))


    G_nodes_dict={}
    links = []
    for i in G.edges:
        if i[0] not in G_nodes_dict.keys():
            G_nodes_dict[i[0]]=0
        if i[1] not in G_nodes_dict.keys():
            G_nodes_dict[i[1]]=0
        links.append({"source": i[0], "target": i[1]})
        weight=G.get_edge_data(i[0],i[1])['weight']
        G_nodes_dict[i[0]]+=weight
        G_nodes_dict[i[1]]+=weight

    #plot
    fig, ax = plt.subplots(figsize=figsize) 
    pos = nx.spring_layout(G, scale=pos_scale, k=(pos_size)/np.sqrt(G.order()))
    p=dict(G.nodes)

    nodesize=np.array([G_nodes_dict[u] for u in G.nodes()])/nodesize_scale
    nodecolos=[type_color_all[u] for u in G.nodes()]
    nx.draw_networkx_nodes(G, pos, nodelist=p,node_size=nodesize,node_color=nodecolos)

    edgewidth = np.array([G.get_edge_data(u, v)['weight'] for u, v in G.edges()])/edgeswidth_scale
    nx.draw_networkx_edges(G, pos,width=edgewidth)


    #label_options = {"ec": "white", "fc": "white", "alpha": 0.6}
    #nx.draw_networkx_labels(G, pos, font_size=10,) #bbox=label_options)
    plt.grid(False)
    plt.axis("off")
    plt.xlim(-2,2)
    plt.ylim(-2,1.5)

    labels = adata.obs[celltype_key].cat.categories
    #用label和color列表生成mpatches.Patch对象，它将作为句柄来生成legend
    color = [type_color_all[u] for u in labels]
    patches = [mpatches.Patch(color=type_color_all[u], label=u) for u in labels ] 

    #plt.xlim(-0.05, 1.05)
    #plt.ylim(-0.05, 1.05)
    plt.axis("off")
    plt.title(title)
    plt.legend(handles=patches,
               bbox_to_anchor=legend_bbox, 
               ncol=legend_ncol,
               fontsize=legend_fontsize)
    if return_graph==True:
        return G
    else:
        return ax

def cpdb_plot_interaction(adata:anndata.AnnData,cell_type1:str,cell_type2:str,
                          means:pd.DataFrame,pvals:pd.DataFrame,
                          celltype_key:str,genes=None,
                         keep_significant_only:bool=True,figsize:tuple = (4,8),title:str="",
                         max_size:int=1,highlight_size:float = 0.75,standard_scale:bool = True,
                         cmap_name:str='viridis',
                         ytickslabel_fontsize:int=8,xtickslabel_fontsize:int=8,title_fontsize:int=10)->matplotlib.axes._axes.Axes:
    r"""
    Plot a CellPhoneDB interaction.

    Arguments:
        adata: AnnData object containing the data.
        cell_type1: Name of cell type 1.
        cell_type2: Name of cell type 2.
        means: DataFrame containing the means for each interaction.
        pvals: DataFrame containing the p-values for each interaction.
        celltype_key: Key in adata.obs that contains the cell type information.
        genes: List of genes to include in the plot. If None, all genes are included.
        keep_significant_only: Whether to keep only significant interactions.
        figsize: Figure size.
        title: Title of the plot.
        max_size: Maximum size of the dots.
        highlight_size: Size of the dots for the highlighted interaction.
        standard_scale: Whether to standard scale the data.
        cmap_name: Name of the colormap to use.
        ytickslabel_fontsize: Fontsize of the yticks labels.
        xtickslabel_fontsize: Fontsize of the xticks labels.
        title_fontsize: Fontsize of the title.

    Returns:
        ax: Axes object containing the plot.

    """
    check_kpy()
    global kpy_install
    if kpy_install==True:
        global_imports("ktplotspy","kpy")

    fig=kpy.plot_cpdb(
        adata = adata,
        cell_type1 = cell_type1,
        cell_type2 = cell_type2, 
        means = means,
        pvals = pvals,
        celltype_key = celltype_key,
        genes = genes,
        keep_significant_only=keep_significant_only,
        figsize = figsize,
        title = "",
        max_size = max_size,
        highlight_size = highlight_size,
        standard_scale = standard_scale,
        cmap_name=cmap_name
    ).draw()

    #ytickslabels
    labels=fig.get_axes()[0].yaxis.get_ticklabels()
    plt.setp(labels, fontsize=ytickslabel_fontsize)

    #xtickslabels
    labels=fig.get_axes()[0].xaxis.get_ticklabels()
    plt.setp(labels, fontsize=xtickslabel_fontsize)

    fig.get_axes()[0].set_title(title,fontsize=title_fontsize)

    return fig.get_axes()[0]

def cpdb_interaction_filtered(adata:anndata.AnnData,cell_type1:str,cell_type2:str,
                              means:pd.DataFrame,pvals:pd.DataFrame,celltype_key:str,genes=None,
                         keep_significant_only:bool=True,figsize:tuple = (0,0),
                         max_size:int=1,highlight_size:float = 0.75,standard_scale:bool = True,cmap_name:str='viridis',)->list:
    r"""
    Returns a list of unique interaction groups from a CellPhoneDB analysis filtered by significance.

    Arguments:
        adata: Annotated data matrix containing normalized gene expression data.
        cell_type1: The cell type name for the first interacting partner.
        cell_type2: The cell type name for the second interacting partner.
        means: Dataframe containing the means of gene expression for each interaction.
        pvals: Dataframe containing the p-values for each interaction.
        celltype_key: The column name in the `adata.obs` dataframe containing cell type annotations.
        genes: List of gene names to filter on. If None, no filtering is applied.
        keep_significant_only: Whether to filter on significant interactions only (i.e. interactions with FDR < 0.05).
        figsize: Figure size. Width and height of the figure in inches. 
        max_size: Maximum size of the markers in the plot.
        highlight_size: Size of the markers for the highlighted interaction.
        standard_scale: Whether to standard scale the data.
        cmap_name: Name of the colormap to use.

    Returns:
        genes: List of unique interaction groups from a CellPhoneDB analysis filtered by significance.

    """
    check_kpy()
    global kpy_install
    if kpy_install==True:
        global_imports("ktplotspy","kpy")

    res=kpy.plot_cpdb(
        adata = adata,
        cell_type1 = cell_type1,
        cell_type2 = cell_type2, 
        means = means,
        pvals = pvals,
        celltype_key = celltype_key,
        genes = genes,
        keep_significant_only=keep_significant_only,
        figsize = figsize,
        title = "",
        max_size = max_size,
        highlight_size = highlight_size,
        standard_scale = standard_scale,
        cmap_name=cmap_name,
        return_table=True
    )

    return list(set(res['interaction_group']))

def cpdb_submeans_exacted(means:pd.DataFrame,cell_names:str,cell_type:str='ligand')->pd.DataFrame:
    r"""
    Returns a subset of the means DataFrame for a given cell type and cell name.

    Arguments:
        means: DataFrame containing the means for each interaction.
        cell_names: Name of the cell type.
        cell_type: Whether the cell type is a ligand or a receptor.

    Returns:
        means: Subset of the means DataFrame for a given cell type and cell name.

    """
    if cell_type=='ligand':
        means_columns=means.columns[:11].tolist()+means.columns[means.columns.str.contains('{}\|'.format(cell_names))].tolist()
    elif cell_type=='receptor':
        means_columns=means.columns[:11].tolist()+means.columns[means.columns.str.contains('\|{}'.format(cell_names))].tolist()
    else:
        raise ValueError('cell_type must be ligand or receptor')
    return means.loc[:,means_columns]