DGLError: Expect number of features to match number of nodes (len(u)). Got 14 and 15 instead

[melnimr]

This is driving me insane. Here is the code I have:

import pandas as pd
import numpy as np
import re
import torch
import glob
import dgl
from dgl.data import DGLDataset
class InductorDataset(DGLDataset):
    def __init__(self):
        super().__init__(name='inductor_dataset')
    def process(self):
        n_node = []
        n_edge = []
        self.graphs = []
        self.labels = []
        for i,file in enumerate(glob.iglob(r'./fh_spiral_*.inp')):
            self.node_pos = []
            self.node_attr = []
            src = []
            dst = []
            n_count = 0 
            e_count = 0  
            self.edge_attr = []
            self.targets = []
            array = []
            df = pd.read_csv(file,header =None)
            mat_df = pd.read_csv('../mat_files/L0_cell_LNA_SGM_'+ str(i) + '.mat',skiprows = 1,header=None,delim_whitespace=True)
            for index, row in df.iterrows():
                if row[0][0] == 'N':
                    n_count = n_count + 1
                    array = re.findall(r'[\d\.\d]+', row[0])
                    n = array[0]
                    x = array[1]
                    y = array[2]
                    z = 0.00 
                    self.node_pos.append(np.array((x,y,z),dtype= np.float32))
                    #node_attr.append(np.array([1,2,3]))
                if row[0][0] == 'E':
                    e_count = e_count + 1
                    self.edge_attr.append([5.0,3.2])  # w= 5.0, T = 3.2
                    src.append(e_count)
                    dst.append(e_count + 1)
            src = np.array(src).astype(np.int64)
            dst = np.array(dst).astype(np.int64)
            n_node.append(n_count)
            n_edge.append(e_count)
            label = float(mat_df[1].to_list()[0]) #nH
            #Create a graph and add it to the list of graphs and labels.
            g = dgl.graph((src, dst))
            #self.node_pos = np.concatenate(self.node_pos, axis = 0)
            #g.ndata['pos'] = torch.tensor(self.node_pos, dtype=torch.float32)
            #g.ndata['attr'] = F.tensor(self.node_attr[self.n_cumsum[idx]:self.n_cumsum[idx+1]], dtype=F.data_type_dict['float32'])
            #g.edata['edge_attr'] = torch.tensor(self.edge_attr[self.ne_cumsum[idx]:self.ne_cumsum[idx+1]], dtype=torch.float32)
            self.graphs.append(g)
            self.labels.append(label)
        self.node_pos = np.concatenate(self.node_pos, axis = 0)
        self.n_cumsum = np.concatenate([[0], np.cumsum(n_node)])
        self.ne_cumsum = np.concatenate([[0], np.cumsum(n_edge)])
        self.labels = torch.tensor(self.labels)
    def __getitem__(self, idx):
        pos = self.node_pos[self.n_cumsum[idx]:self.n_cumsum[idx+1]]
        self.graphs[idx].ndata['pos'] = torch.tensor(pos, dtype=torch.float32)
        return self.graphs[idx], self.labels[idx]
    def __len__(self):
        return len(self.graphs)

But when create an instance:

inductor_graph = InductorDataset()

and call:

inductor_graph[0]

I get:

DGLError: Expect number of features to match number of nodes (len(u)). Got 14 and 15 instead.

[melnimr]

Figured it out. The src and dst arrays index has to start at 0.