LawGraph

A mapping of the Israeli legalisation.

This repository is the implementation of a final project in Digital Humanities course taken by Guy Freund & David Ziegler from The department of Computer Science, Ben-Gurion University, Israel. This project is jointly guided by the Ben-Gurion University and the Israeli Ministry of Justice. The project's goal is to create a mapping of all the laws in the Israeli legalisation. The data being used is given by the Israeli Ministry of Justice.

Database Format

Each law is structured as an XML file in AKN format and contains several ref elements. Each ref element stores a string reference that points to a specific law and to a specific element nested in the this law's XML file. Each such ref element represents a connection between law, such as:

• Amendment of a section in another law
• Reference to a definition contained in another law
• Reliance on a provision appearing in another law

This program creates the graph from the Israeli Ministry of Justice's database that contains all the laws. Each law's path is of the $LAW_GRAPH_REPO_ROOT/akn/il/$LEG_TYPE/$DATE/$FRBRWORKURI/he@/main.xml format, where:

• LAW_GRAPH_REPO_ROOT: The root of the LawGraph repository.
• LEG_TYPE: The type of legalisation. Currently: PrimaryLegislation, SecondaryLegislation or PrimaryOrSecondaryLegislation.
• DATE: A date of format YYYY-MM-DD.
• FRBRWORKURI: The element's value stored in each law in path akomaNtoso/act/meta/identification/FRBRWork/FRBRthis/FRBRuri.

Graph Creation Logic:

For each reference: from_law, from_vertex, to_law & to_vertex vertexes are created. from_law => from_vertex, from_vertex => to_vertex, to_law => to_vertex edges are created.

Neo4j Implementation:

Spinning up the database: Method 1

For automatic upload to neo4j, use the following command options: "-auto url:port" Url and port are optional, if not specified, "bolt://localhost:7687" will be used.

Spinning up the database: Method 2

Create Nodes.csv and Edges.csv files, copy them into the DB 'Import' folder, and run 'create_all.cypher' to load them into the DB. Notice that nodes have ID used for connecting relationships, than the IDs are removed. To use this method execute using the following command options: "-csv"

Spinning up the database: Method 3

Create Nodes.csv and Edges.csv files, copy them into a newly created DB 'Import' folder. Run the command in the 'command' file in the DB's Terminal. To use this method execute using the following command options: "--csv-import"

Repository's Files:

• generate_graph.py: Generates the Pythonic graph.

• classes.py: Contains all Python Classes used across the program.

• utils.py: Gathers all util functions used to build the Pythonic graph.

• constans.py: Gathers all constants used to build the Pythonic graph.

• error.py: Contains all error handling functions.

• errors.json: The errors file, contains 5 different types errors and is being built during runtime.

• law_graph_logs.log: The log file. The program logs to this file during runtime.

• requirements.txt: The requirements file.

• validate_class_uniqueness.py: Validates that all class's elements can be 1to1 identified.

• validate_graph.py: Validates the graph correctness.

• neo4j : Folder with Cyphers and import command.

  • command : Import command to be used in a DB terminal to import csv files.
  • create_all.cypher : Run delete_all, create_nodes, create_edges, remove_id in a row.
  • create_edges.cypher : Create relationships from csv file.
  • create_nodes.cypher : Create nodes from csv file.
  • delete_all.cypher : Delete all nodes and relationships from DB.
  • merge_all.cypher : Merge data from Nodes.csv and Edges.csv into DB.
  • merge_edges.cypher : Merge Edges.csv into DB.
  • merge_nodes.cypher : Merge Nodes.csv into DB.
  • remove_id.cypher : Remove ID property from all objects in DB.
  • example1_preambles_with_references_to_other_laws.cypher : Example Cypher query. Find all Preambles that reference laws other than the one the belong to.
  • example2_nodes_with_no_in_edges.cypher : Example Cypher query. Find all nodes that has no incoming relationships.

Running the Program:

1. Install Python (version>=3.9.0)
2. Download neo4j https://neo4j.com/download/?ref=try-neo4j-lp
3. Run: pip install -r requirements.txt
4. Follow preferred method from above.