Smart Internal Link Strategy Tool

A Flask web application that generates an internal-linking strategy for e-commerce category pages using rule-based semantic scoring and tier-based filtering.

Features

• CSV and Excel (.xlsx) file upload with drag-and-drop support
• Semantic similarity scoring via root-word matching, adjective sharing, and contextual grouping
• Three-tier category ranking: RICH (high product count), PRIORITY (user-defined keywords), DEFAULT
• Assigns up to 5 outgoing internal links per category
• Starving-category balancing — redistributes link equity to under-linked pages
• Downloadable result file in the same format as the input (CSV or Excel)
• 16 MB upload limit with binary-content rejection for CSV uploads

Requirements

• Python 3.10+
• pip

Installation

git clone https://github.com/sungurkeremgezergun/internal-link-strategy-tool.git
cd internal-link-strategy-tool
python -m venv .venv
source .venv/bin/activate        # Windows: .venv\Scripts\activate
pip install -r requirements.txt

Running

# Development
python app.py

# Using Flask CLI
flask --app app run --port 5001

# Production (set FLASK_DEBUG=False before deploying behind a real WSGI server)
FLASK_PORT=5001 python app.py

Open http://127.0.0.1:5001 in your browser.

Configuration

Environment Variable	Default	Description
FLASK_DEBUG	False	Enable Flask debug mode (true/false)
FLASK_PORT	5001	Port the development server listens on

Input CSV Format

The uploaded file must contain at minimum the column below. Column names are matched case-insensitively.

Column	Required	Description
Category Name	Yes	Human-readable category label
Product Count	No	Number of products in the category (integer)
Link	No	URL slug or full URL for the category page

See example_input.csv for a ready-to-use sample.

Output Format

The result file contains all original columns plus five new columns appended to the right:

Column	Description
Internal Link 1	Category name of the first recommended link
Internal Link 2	...
Internal Link 3	...
Internal Link 4	...
Internal Link 5	Category name of the fifth recommended link

Architecture

app.py               Flask routes, file upload/download, request validation
link_algorithm.py    Category class, tier assignment, candidate scoring, process_data()
templates/index.html Single-page React UI (Babel in-browser transpile, Tailwind CSS)

The algorithm runs entirely in memory — no database required.

Tier system:

Tier	Condition	Role in linking
RICH	product count > 5	Preferred candidate for semantic links
PRIORITY	name contains a user-supplied keyword	Safe-harbour fallback target
DEFAULT	everything else	Used only when higher tiers are absent

Candidate selection order per category: root match → sibling match → parent backfill → contextual match → PRIORITY fallback → starving-category fill → RICH fill.

License

MIT License. Copyright (c) 2026 sungurkeremgezergun. See LICENSE for details.