InsightNet

InsightNet is a Python-based tool for analyzing local routing models by inferring the global state of a network and listing all possible alive/failed edge combinations for a given local state. It includes both a CLI and an API, supports multiple output formats (CSV, JSON, JSONL, DOT), and is containerized with Docker for easy deployment and reproducibility.

Getting Started

Prerequisites

• Python 3.10+
• Poetry
• Docker

Setup

1. Clone the repository

   git clone https://github.com/zachathanasiadis/InsightNet.git
   cd InsightNet

2. Install dependencies

   poetry install

Sample Files

The repository includes example files to help you get started:

• sample_graphs/ – Contains example input graph JSON files for testing the CLI.
• sample_requests/ – Contains example API request JSON files for testing the API.

Usage (CLI)

Run the CLI tool with:

python -m insightnet \
  --input-json path/to/input.json \
  --current-state edge,node \
  --output path/to/output.csv

Required Arguments:

• --input-json (-i): Path to the input JSON file containing the routing model and graph data.
• --current-state (-s): The edge and node of the state to be queried, separated by a comma.
• --output (-o): Output file path with extension (.csv, .json, .jsonl, .dot).

Required CLI Input JSON Fields:

• routing_model (string): The name of the routing model.
• nodes (list of strings): List of node names.
• edge_to_node_mapping (list): Each item specifies an edge and the nodes it connects:
  • edge (int): Edge identifier.
  • nodes (list of strings): Nodes connected by this edge.
• routing_table (list of dictionaries): Each item specifies the routing rules for a node.

Example

Example Command:

python -m insightnet -i sample_graphs/graph4.json -s 2,A -o results/output.jsonl

Example Output (results/output.jsonl):

{"path": [[2, "A"], [null, "B"]], "alive_edges": [2], "failed_edges": [3]}
{"path": [[2, "A"], [2, "B"], [null, "A"]], "alive_edges": [2], "failed_edges": [1, 3]}
{"path": [[2, "A"], [3, "B"], [null, "C"]], "alive_edges": [2, 3], "failed_edges": [1]}
{"path": [[2, "A"], [3, "B"], [3, "C"], [null, "B"]], "alive_edges": [2, 3], "failed_edges": [1]}
{"path": [[2, "A"], [3, "B"], [3, "C"], [2, "B"], [null, "A"]], "alive_edges": [2, 3], "failed_edges": [1]}

Usage (API)

Run the API with Docker

1. Build the Docker image:

   docker build -t insightnet-api .

2. Run the API container:

   docker run -p 8000:8000 insightnet-api

The API will be available at http://localhost:8000.

Endpoint:

• POST /infer

Required Request Body Fields:

• routing_model (string): The name of the routing model.
• nodes (list of strings): List of node names.
• edge_to_node_mapping (list): Each item specifies an edge and the nodes it connects:
  • edge (int): Edge identifier.
  • nodes (list of strings): Nodes connected by this edge.
• routing_table (list of dictionaries): Each item specifies the routing rules for a node.
• current_state (string): State to be queried, formatted as "edge,node".

Example

Example Request:

curl -X POST "http://localhost:8000/infer" \
  -H "Content-Type: application/json" \
  -d @sample_requests/request1.json

Example Response:

{
  "results": [
    {
      "path": [[1,"v1"],[null,"v2"]],
      "alive_edges": [1],
      "failed_edges": [2]
    },
    {
      "path": [[1,"v1"],[1,"v2"],[null,"v1"]],
      "alive_edges": [1],
      "failed_edges": [2]
    },
    {
      "path": [[1,"v1"],[2,"v2"],[null,"v3"]],
      "alive_edges": [1,2],
      "failed_edges": []
    },
    {
      "path": [[1,"v1"],[2,"v2"],[2,"v3"],[null,"v2"]],
      "alive_edges": [1,2],
      "failed_edges": []
    },
    {
      "path": [[1,"v1"],[2,"v2"],[2,"v3"],[1,"v2"],[null,"v1"]],
      "alive_edges": [1,2],
      "failed_edges": []
    }
  ]
}

Authors

• Zacharias Athanasiadis (Github)

Acknowledgements

• Csaba Györgyi (Github) – Project advisor, for guidance and feedback throughout the development of InsightNet.