How mstabc/cc_node2vec works
Project type and purpose
• This is a command-line, research‐oriented Python pipeline for conformance checking of process models. It ingests Petri net definitions and event logs (.pnml and .xes), constructs graph representations, derives trace embeddings using random walks (Node2Vec), computes distance metrics, performs PCA for visualization, and writes out results for static and context-aware modes.
• Key indicators: a single “src” folder with Python modules, a “data” folder for input PNML and XES, a “results” folder for outputs, and README instructions on running “main.py case” or “main.py mined.”
File structure
• Top-level directories and files:
• data/ — contains subfolders per dataset, each with multiple .pnml files
• src/ — code modules: main.py, petri_builder.py, graph_encoding.py, random_walks.py, distance_cal.py, utils.py
• results/ — output visualizations and CSVs
• mined_results/ — (presumably intermediate or alternate outputs for static mode)
• Architectural patterns: a linear pipeline architecture. main.py orchestrates reading data → building Petri nets → generating graphs → encoding via random walks → embedding comparison → output.
• Configuration/deployment: no web server or database; invoked purely as a CLI script.
README for additional insights
• Describes two modes: “case” (context-aware) and “mined” (static).
• Highlights output artifacts: PCA plots, CSV distance tables.
• Mentions core functions and reference to Petri net and Node2Vec‐based GCN research.