Is it possible to use heterogeneous pyg object to find anaomalies?

[monk1337]

Hi, Thank you for this awesome package. I am working with heterogeneous and knowledge graphs. For example, if I use the famous MovieLens dataset and construct a heterogeneous graph, Can I feed it to model.fit(data) ?

[monk1337]

I tried to experiment with hetro graph but getting errors.

My data looks like this:

import pandas as pd
from torch_geometric.data import HeteroData
import torch
from torch_geometric.data import download_url, extract_zip

url = 'https://files.grouplens.org/datasets/movielens/ml-latest-small.zip'
extract_zip(download_url(url, '.'), '.')

movie_path = './ml-latest-small/movies.csv'
rating_path = './ml-latest-small/ratings.csv'

print(pd.read_csv(movie_path).head())
print(pd.read_csv(rating_path).head())

def load_node_csv(path, index_col, encoders=None, **kwargs):
    df = pd.read_csv(path, index_col=index_col, **kwargs)
    mapping = {index: i for i, index in enumerate(df.index.unique())}

    x = None
    if encoders is not None:
        xs = [encoder(df[col]) for col, encoder in encoders.items()]
        x = torch.cat(xs, dim=-1)

    return x, mapping



class SequenceEncoder(object):
    def __init__(self, model_name='all-MiniLM-L6-v2', device=None):
        pass

    @torch.no_grad()
    def __call__(self, df):
        x = torch.rand(df.shape[0], 384)
        # for demo purpose no loading actual model
        print(x.shape)
        return x.cpu()
    
    
class GenresEncoder(object):
    def __init__(self, sep='|'):
        self.sep = sep

    def __call__(self, df):
        genres = set(g for col in df.values for g in col.split(self.sep))
        mapping = {genre: i for i, genre in enumerate(genres)}

        x = torch.zeros(len(df), len(mapping))
        for i, col in enumerate(df.values):
            for genre in col.split(self.sep):
                x[i, mapping[genre]] = 1
        return x
    
    
movie_x, movie_mapping = load_node_csv(
    movie_path, index_col='movieId', encoders={
        'title': SequenceEncoder(),
        'genres': GenresEncoder()
    })


_, user_mapping = load_node_csv(rating_path, index_col='userId')



data = HeteroData()

data['user'].num_nodes = len(user_mapping)  # Users do not have any features.
data['movie'].x = movie_x

print(data)

Now feeding this data to pygod model

# train a dominant detector
from pygod.models import DOMINANT
model = DOMINANT(num_layers=4, epoch=2)  # hyperparameters can be set here

model.fit(data)  # data is a Pytorch Geometric data object

giving error

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
Cell In [39], line 1
----> 1 model.fit(data)  # data is a Pytorch Geometric data object

File ~/miniforge3/envs/M1Max/lib/python3.8/site-packages/pygod/models/dominant.py:129, in DOMINANT.fit(self, G, y_true)
    111 def fit(self, G, y_true=None):
    112     """
    113     Fit detector with input data.
    114 
   (...)
    127         Fitted estimator.
    128     """
--> 129     G.node_idx = torch.arange(G.x.shape[0])
    130     G.s = to_dense_adj(G.edge_index)[0]
    132     # automated balancing by std

File ~/miniforge3/envs/M1Max/lib/python3.8/site-packages/torch_geometric/data/hetero_data.py:133, in HeteroData.__getattr__(self, key)
    131 elif bool(re.search('_dict$', key)):
    132     return self.collect(key[:-5])
--> 133 raise AttributeError(f"'{self.__class__.__name__}' has no "
    134                      f"attribute '{key}'")

AttributeError: 'HeteroData' has no attribute 'x'

[kayzliu]

PyGOD only works with homogeneous graphs by far. One of the potential solutions is converting the heterogeneous graph to a homogeneous one with to_homogeneous. Hope this helps.