Social Network Graph Explorer

A dependency-free Python command-line application for searching, ranking, and exploring directed social-network graphs.

The project is intended for students and developers interested in practical graph algorithms and indexing data structures. It loads pipe-delimited user, relationship, and block data; builds efficient in-memory indexes; and exposes the results through an interactive terminal menu.

Key features

• Iterative PageRank with dangling-node handling and warm starts
• Personalized PageRank combined with Jaccard biography similarity for recommendations
• Username and biography search backed by hash maps and an inverted index
• Trie-based, case-insensitive username autocomplete
• Levenshtein-distance spelling suggestions
• Breadth-first graph traversal by level
• Block-aware follow relationships and session interaction history
• A synthetic English demo dataset with no personal information

Technology and architecture

The application uses Python 3.10 or newer and only the standard library. SocialGraph owns adjacency sets, reverse adjacency sets, block indexes, PageRank state, a biography inverted index, and a username trie. The CLI delegates graph operations to that domain class, while small utility modules contain text algorithms and data models.

Getting started

Prerequisites

• Python 3.10+

No third-party packages are required.

Prepare the data

Serialized graph files are generated locally and intentionally excluded from version control:

python prepare_data.py

Run this command again whenever an input text file changes.

Run the application

python main.py

Choose the included dataset interactively, or pass it directly:

python main.py dataset/small

Run the tests

python -m unittest discover -s tests -v

Input format

Each dataset directory contains three pipe-delimited files:

users.txt:       id|username|bio
connections.txt: from_id|to_id
blocked.txt:     blocker_id|blocked_id

Malformed references, duplicate users and relationships, self-links, and unknown user IDs are ignored during loading.

Example workflows

With the small dataset, try username search for al, biography search for python graph, autocomplete for gr*, or a did-you-mean query for alce. Menu option 11 displays graph size and the latest PageRank timings.

Project structure

main.py          Interactive command-line interface
social_graph.py  Graph storage, loading, ranking, search, and recommendations
models.py        User, interaction, and load-report data classes
trie.py          Prefix-search data structure
text_utils.py    Tokenization, Jaccard similarity, and Levenshtein distance
prepare_data.py  Local serialized-data generator
dataset/         Pipe-delimited sample datasets
tests/           Standard-library unit tests

Known limitations

• Changes made in the CLI are held in memory and are not written back to the source files.
• Interaction history includes only relationships added during the current session.
• Personalized PageRank and spelling suggestions scan the loaded graph, so performance depends on dataset size.
• Pickle files must only be generated from and loaded within a trusted local checkout.

Data and license

The included dataset is synthetic and exists only to demonstrate the application. The project is available under the MIT License.

Contributing and security

See CONTRIBUTING.md for the development workflow and SECURITY.md for responsible vulnerability reporting.