Skip to content

Preprocessing4MatrixFilter.md

Latest commit

 

History

History
53 lines (37 loc) · 1.78 KB

Preprocessing4MatrixFilter.md

File metadata and controls

53 lines (37 loc) · 1.78 KB

Preprocessing for Matrix Filter

Overview

Python scripts for preprocessing graphs and training the ML filter used by the CPU subgraph isomorphism (SubIso) application. The pipeline converts graph formats, generates embeddings, and trains similarity models.

Functionality

  1. Rapids format → PyTorch Geometric (.pt)
  2. PyTorch .pt → MatrixGraph embedding (binary)
  3. Ground truth generation (for training)
  4. Model training (similarity + GNN)

Tool Chain

Step Script Input Output
1 graph_reader.py Rapids-format graph PyTorch .pt
2 data.py or embedding.py PyTorch .pt MatrixGraph embedding (binary)
3 (custom) — Ground truth (uint64 array)
4 train.py Config + embeddings + ground truth Trained model

Parameters

Script Parameters Description
graph_reader.py [input_path] [output_path] Convert Rapids graph to PyTorch
data.py [input_path] [output_path] Convert .pt to embedding binary
train.py (config) Train similarity model

Source

tools/python/graph_reader.py
tools/python/data.py
tools/python/embedding.py
tools/python/train.py

Examples

# Step 1: Rapids → PyTorch
python tools/python/graph_reader.py [path-to-rapids-graph] [output.pt]

# Step 2: PyTorch → embedding
python tools/python/data.py [path-to-binary-torch] [output-embedding]

# Step 4: Train (configure config.yaml first)
python tools/python/train.py

Ground Truth

Ground truth must be a 1D binary array of uint64_t storing candidate vertex IDs for the first query vertex of the pattern. See SubIsoTraining.md for details.