Technology · Agentic → Evaluations · Open Source · wiki:deep

PM4Py

Taxonomy

PM4Py is a Python library for state-of-the-art process mining: read XES/CSV event logs, discover process models (e.g. inductive Petri nets), check conformance, and analyze processes for academia and industry. Developed by PIS — Process Intelligence Solutions (Fraunhofer FIT spin-off).

The GitHub open-source build is AGPL-3.0; PIS also offers a separate commercial license for closed-source use. Install via pip install -U pm4py (Python 3.9–3.14; extras for polars, ML, LLM connectors, etc.).

RepoDocsStack: agentic-evaluations

Why it matters here

Agent and workflow traces become useful when they are event logs, not chat dumps. PM4Py is the consumer side of the ABM/XES path: once agent behavior is exported to XES/CSV, PM4Py discovers models and measures conformance — the Evaluations cell’s process-mining engine, not another LLM judge.

Keeps “what agents actually did” measurable with classical process-mining methods alongside LLM eval suites.

How it works

  1. Ingest an event log (pm4py.read_xes / CSV paths).
  2. Discover a model (e.g. discover_petri_net_inductive).
  3. Visualize / check conformance / run further analytics.
  4. Optionally combine with ABM-style exporters that produce XES from agent runs.

Related: abm-agent-behavior-mining · agent-evaluation · deepeval · topics/32-eval-harnesses

Flow

Scroll inside the canvas to pan

When to reach for it

  • Use when: you have (or will produce) event logs and need discovery/conformance/analytics in Python.
  • Skip when: you need LLM-as-judge CI (deepeval / promptfoo / ragas) without a log format, or only need the ABM export demo without mining algorithms.
  • Prefer instead: ProM for GUI-heavy mining; keep ABM as the multi-agent → XES bridge, pm4py as the library.

Limits

  • AGPL-3.0 on the public GitHub tree — closed-source product embeds need the commercial license path.
  • Garbage-in: mining quality tracks case IDs, timestamps, and activity labels — chat traces without structure fail.
  • Not an agent framework: no orchestration; pair with ABM/LangGraph/etc. for generation of logs.
  • License ≠ “Open Source means free for SaaS” — AGPL obligations matter for hosted OS builds.

Linked from

Catalog backlinks — what points here (wiki “what links here”).

Features

Stacks

Peers

What we checked

Claims below are backed by science sources on disk.

Founding paper / PM4Py library

Process Mining for Python (PM4Py): Bridging the Gap Between Process- and Data Science · STRONG

“In this paper, we present a novel process mining library, i.e. Process Mining for Python (PM4Py) that aims to bridge this gap, providing integration with state-of-the-art data science libraries, e.g., pandas, numpy, scipy and scikit-learn.”

Library purpose / process mining

pm4py/pm4py · MODERATE

“PM4Py is a python library that supports state-of-the-art process mining algorithms in Python.”

Docs / API surface

PM4Py documentation · MODERATE

“Open-source Python library that implements state-of-the-art process mining algorithms for building custom applications”

Pairing with ABM / XES

ABM contrast · MODERATE

“ABM (Agent Behavior Mining) supplementary materials for mining agent behavior into process-mining formats (XES-oriented) so logs can be analyzed with standard process-mining tools.”

In this research stack

Features and peers linked from the catalog map — not a second product surface.

Features it supports

Same cell

Stack

Document management

No DM vendor crosswalk edges yet.

Source

Primary repo github.com/process-intelligence-solutions/pm4py · Open Source

technologies/pm4py/README.md

Research inventory

9 tags · 18 out · 19 in · 3 artifacts · 0 gaps · 35 corpus docs

Catalog tags

landscape.layer
Agentic
landscape.subcategory
Evaluations
license_tag
Open Source
maps.dm
present
maps.features
11
one_liner
Evaluations
review.depth
science
slug
pm4py
title
PM4Py

Artifacts

  • dm_map · present · technologies/pm4py/document-management.md
  • features_map · present · technologies/pm4py/features.md
  • readme · present · technologies/pm4py/README.md

Out · alternative_to

Out · maps_to

In · alternative_to

In · in_stack

In · maps_to

Corpus tags

category
Agentic → Evaluations
dedication
open-source
feature
observability-auditability
outcome-verification
r2-14-process-mining
r2-30-evaluator-critic
r2-32-eval-harnesses
r2-38-abm-span-mapping
r2-42-intervention-preassessment
r2-44-oced-case-events
r2-47-trace-maturity-gate
testing-debugging-optimization
wt-test-debug-evaluate
kind
map_edge
tech_features
tech_quote
tech_readme
tech_science_source
tech_section
needs_deepen
false
quality
ok
section
Evidence
Features map
GitHub map
How it works
Limits & failure modes
Links
Scientific notes
What it is
When to use / skip
Why it matters here
slug
pm4py
source_id
abm-contrast
pm4py-docs
pm4py-github
technology
pm4py

Corpus documents (35)

map_edge · 11

  • pm4py → observability-auditability
  • pm4py → outcome-verification
  • pm4py → r2-14-process-mining
  • pm4py → r2-30-evaluator-critic
  • pm4py → r2-32-eval-harnesses
  • pm4py → r2-38-abm-span-mapping
  • pm4py → r2-42-intervention-preassessment
  • pm4py → r2-44-oced-case-events
  • … +3 more

tech_features · 1

  • PM4Py · features

tech_quote · 9

  • PM4Py · abm-contrast
  • PM4Py · abm-contrast
  • PM4Py · abm-contrast
  • PM4Py · pm4py-docs
  • PM4Py · pm4py-docs
  • PM4Py · pm4py-docs
  • PM4Py · pm4py-github
  • PM4Py · pm4py-github
  • … +1 more

tech_readme · 1

  • PM4Py

tech_science_source · 3

  • PM4Py · abm-contrast
  • PM4Py · pm4py-docs
  • PM4Py · pm4py-github

tech_section · 10

  • PM4Py · Evidence
  • PM4Py · Features map
  • PM4Py · GitHub map
  • PM4Py · How it works
  • PM4Py · Limits & failure modes
  • PM4Py · Links
  • PM4Py · Scientific notes
  • PM4Py · What it is
  • … +2 more