PyTorch Geometric (PyG) Adapter

The ArangoDB-PyG Adapter exports Graphs from ArangoDB into PyTorch Geometric (PyG), a PyTorch-based Graph Neural Network library, and vice-versa

PyTorch Geometric (PyG) is a library built upon PyTorch  to easily write and train Graph Neural Networks (GNNs) for a wide range of applications related to structured data.

It consists of various methods for deep learning on graphs and other irregular structures, also known as geometric deep learning , from a variety of published papers. In addition, it consists of easy-to-use mini-batch loaders for operating on many small and single giant graphs, multi GPU-support , DataPipe support , distributed graph learning via Quiver , a large number of common benchmark datasets (based on simple interfaces to create your own), the GraphGym  experiment manager, and helpful transforms, both for learning on arbitrary graphs as well as on 3D meshes or point clouds.


The ArangoDB-PyG Adapter repository  is available on Github. Check it out!


To install the latest release of the ArangoDB-PyG Adapter, run the following command:

pip install torch
pip install adbpyg-adapter


The following examples show how to get started with ArangoDB-PyG Adapter. Check also the interactive tutorial .


import torch
import pandas
from torch_geometric.datasets import FakeHeteroDataset

from arango import ArangoClient  # Python-Arango driver

from adbpyg_adapter import ADBPyG_Adapter, ADBPyG_Controller
from adbpyg_adapter.encoders import IdentityEncoder, CategoricalEncoder

# Load some fake PyG data for demo purposes
data = FakeHeteroDataset(
    avg_num_channels=3,  # avg number of features per node
    edge_dim=2,  # number of features per edge
    num_classes=3,  # number of unique label values

# Let's assume that the ArangoDB "IMDB" dataset is imported to this endpoint
db = ArangoClient(hosts="http://localhost:8529").db("_system", username="root", password="")

adbpyg_adapter = ADBPyG_Adapter(db)

PyG to ArangoDB

# 1.1: PyG to ArangoDB
adb_g = adbpyg_adapter.pyg_to_arangodb("FakeData", data)

# 1.2: PyG to ArangoDB with a (completely optional) metagraph for customized adapter behaviour
def y_tensor_to_2_column_dataframe(pyg_tensor):
    A user-defined function to create two
    ArangoDB attributes out of the 'y' label tensor

    NOTE: user-defined functions must return a Pandas Dataframe
    label_map = {0: "Kiwi", 1: "Blueberry", 2: "Avocado"}

    df = pandas.DataFrame(columns=["label_num", "label_str"])
    df["label_num"] = pyg_tensor.tolist()
    df["label_str"] = df["label_num"].map(label_map)

    return df

metagraph = {
    "nodeTypes": {
        "v0": {
            "x": "features",  # 1) you can specify a string value for attribute renaming
            "y": y_tensor_to_2_column_dataframe,  # 2) you can specify a function for user-defined handling, as long as the function returns a Pandas DataFrame
    "edgeTypes": {
        ("v0", "e0", "v0"): {
            # 3) you can specify a list of strings for tensor disassembly (if you know the number of node/edge features in advance)
            "edge_attr": [ "a", "b"]  

adb_g = adbpyg_adapter.pyg_to_arangodb("FakeData", data, metagraph, explicit_metagraph=False)

# 1.3: PyG to ArangoDB with the same (optional) metagraph, but with `explicit_metagraph=True`
# With `explicit_metagraph=True`, the node & edge types omitted from the metagraph will NOT be converted to ArangoDB.
# Only 'v0' and ('v0', 'e0', 'v0') will be brought over (i.e 'v1', ('v0', 'e0', 'v1'), ... are ignored)
adb_g = adbpyg_adapter.pyg_to_arangodb("FakeData", data, metagraph, explicit_metagraph=True)

# 1.4: PyG to ArangoDB with a Custom Controller  (more user-defined behavior)
class Custom_ADBPyG_Controller(ADBPyG_Controller):
    def _prepare_pyg_node(self, pyg_node: dict, node_type: str) -> dict:
        """Optionally modify a PyG node object before it gets inserted into its designated ArangoDB collection.

        :param pyg_node: The PyG node object to (optionally) modify.
        :param node_type: The PyG Node Type of the node.
        :return: The PyG Node object
        pyg_node["foo"] = "bar"
        return pyg_node

    def _prepare_pyg_edge(self, pyg_edge: dict, edge_type: tuple) -> dict:
        """Optionally modify a PyG edge object before it gets inserted into its designated ArangoDB collection.

        :param pyg_edge: The PyG edge object to (optionally) modify.
        :param edge_type: The Edge Type of the PyG edge. Formatted
            as (from_collection, edge_collection, to_collection)
        :return: The PyG Edge object
        pyg_edge["bar"] = "foo"
        return pyg_edge

adb_g = ADBPyG_Adapter(db, Custom_ADBPyG_Controller()).pyg_to_arangodb("FakeData", data)

ArangoDB to PyG

# Start from scratch!
db.delete_graph("FakeData", drop_collections=True, ignore_missing=True)
adbpyg_adapter.pyg_to_arangodb("FakeData", data)

# 2.1: ArangoDB to PyG via Graph name (does not transfer attributes)
pyg_g = adbpyg_adapter.arangodb_graph_to_pyg("FakeData")

# 2.2: ArangoDB to PyG via Collection names (does not transfer attributes)
pyg_g = adbpyg_adapter.arangodb_collections_to_pyg("FakeData", v_cols={"v0", "v1"}, e_cols={"e0"})

# 2.3: ArangoDB to PyG via Metagraph v1 (transfer attributes "as is", meaning they are already formatted to PyG data standards)
metagraph_v1 = {
    "vertexCollections": {
        # we instruct the adapter to create the "x" and "y" tensor data from the "x" and "y" ArangoDB attributes
        "v0": { "x": "x", "y": "y"},  
        "v1": {"x": "x"},
    "edgeCollections": {
        "e0": {"edge_attr": "edge_attr"},
pyg_g = adbpyg_adapter.arangodb_to_pyg("FakeData", metagraph_v1)

# 2.4: ArangoDB to PyG via Metagraph v2 (transfer attributes via user-defined encoders)
# For more info on user-defined encoders in PyG, see
metagraph_v2 = {
    "vertexCollections": {
        "Movies": {
            "x": {  # Build a feature matrix from the "Action" & "Drama" document attributes
                "Action": IdentityEncoder(dtype=torch.long),
                "Drama": IdentityEncoder(dtype=torch.long),
            "y": "Comedy",
        "Users": {
            "x": {
                "Gender": CategoricalEncoder(mapping={"M": 0, "F": 1}),
                "Age": IdentityEncoder(dtype=torch.long),
    "edgeCollections": {
        "Ratings": {
            "edge_weight": "Rating"
pyg_g = adbpyg_adapter.arangodb_to_pyg("IMDB", metagraph_v2)

# 2.5: ArangoDB to PyG via Metagraph v3 (transfer attributes via user-defined functions)
def udf_v0_x(v0_df):
    # process v0_df here to return v0 "x" feature matrix
    # v0_df["x"] = ...
    return torch.tensor(v0_df["x"].to_list())

def udf_v1_x(v1_df):
    # process v1_df here to return v1 "x" feature matrix
    # v1_df["x"] = ...
    return torch.tensor(v1_df["x"].to_list())

metagraph_v3 = {
    "vertexCollections": {
        "v0": {
            "x": udf_v0_x,  # supports named functions
            "y": lambda df: torch.tensor(df["y"].to_list()),  # also supports lambda functions
        "v1": {"x": udf_v1_x},
    "edgeCollections": {
        "e0": {"edge_attr": (lambda df: torch.tensor(df["edge_attr"].to_list()))},
pyg_g = adbpyg_adapter.arangodb_to_pyg("FakeData", metagraph_v3)