Comprehensive Branched Declare Constraints Mining extension module

Overview

This branch contains an extension module for the SIESTA framework, designed to extend the incremental mining procedure of Declare constraints. The module introduces support for branching mechanisms on both source and target events within the constraint discovery process.

Purpose

The primary objective of this module is to enhance the standard Declare constraints mining by enabling branched relationships:

• Source Branching: Allowing multiple source events to jointly activate a constraint.
• Target Branching: Allowing multiple target events to be linked under a single activation event.

This is achieved by building on top of SIESTA’s original incremental mining strategy, maintaining scalability while supporting more expressive process models.

Features

• Implementation of AND, OR, and XOR branching policies for both sources and targets.
• Compatible with batch event log processing.
• Optimized to maintain the efficiency of the incremental mining procedure, suitable for big data scenarios.

Run with Docker Compose (v2+)

1. From docker-compose.yml, run the following S3 service:

  docker compose up minio

2. Preprocess a new log file (if not already done) saved locally in input/:

  docker compose up preprocess
  curl -X 'POST' 'http://localhost:8000/preprocess' \
  -H 'Content-Type: application/json' \
  -d '{
    "spark_master": "local[*]",
    "file": "test.xes",
    "logname": "test"
  }'

3. Configure the CBDeclare service by modifying it in the docker-compose.yml file:

   • Replace test in the command section with the logname of your log file.
   • Optionally, add any arguments in the command section as needed separated by commas. See the next section for all available arguments.
   • If your RAM is less than 10g, consider changing the entrypoint driver memory configuration as well.

   E.g. command:["-l", "test", "--support", "0.5", "-p", "AND"]

4. Run the CBDeclare service:

  docker compose up cbdeclare

Note: if network issues arise, consider uncommenting the external: true configuration in the yml file.

Arguments

Required Argument

• -l, --logname <logname>
  Type: String
  Required: Yes
  Description: Name of the log file. This refers to the corresponding log stored in S3.

Optional Arguments

• -s, --support <value>
  Type: Double
  Default: 0
  Description: Minimum support threshold for constraint discovery. Only constraints with support greater than this value will be considered. If set to 0, all constraints are considered.

• -p, --branchingPolicy <AND|OR|XOR>
  Type: String
  Default: null
  Description: Branching policy to apply during constraint discovery.

• -t, --branchingType <SOURCE|TARGET> Type: String
  Default: "TARGET"
  Description: Specifies the direction of the branching. Use TARGET to group multiple targets under a single source.

• -b, --branchingBound <value>
  Type: Int
  Default: 0
  Description: Maximum number of activities allowed in a branched group. If 0, no branching bound is enforced.

• -d, --dropFactor <value>
  Type: Double
  Default: 1.5
  Description: Drop factor used to reduce candidate branching sets. A lower value increases the likelihood of early stopping during greedy expansion.

• -r, --filterRare <true|false>
  Type: Boolean
  Default: false
  Description: If set to true, filters out events that are considered rare (e.g., based on frequency thresholds).

• -u, --filterUnderBound <true|false>
  Type: Boolean
  Default: false
  Description: If enabled, removes templates that do not satisfy the required branching bound.

• -h, --hardRediscover <true|false>
  Type: Boolean
  Default: false
  Description: Forces a full rediscovery from scratch, ignoring any previously cached intermediate results.

---