v2025.7.8.5 - Customizable Fuzzy Search

[bendlikeabamboo]

🚀 v2025.7.8.5 Released - Customizable Fuzzy Search Now Live!

Hi everyone 👋

I'm thrilled to announce the latest release of barangay, a python package for exploring and matching the different administrative regions, province, municipalities, cities, and barangays of the Philippines

🔍 What's New in v2025.7.8.5?

This release introduces a robust fuzzy search mechanism purpose-built for enriching datasets with PSGC data, enabling:

• Typo-tolerant matching across barangay names, even with inconsistent spellings or formatting
• Configurable similarity thresholds and what administrative level to include for precision vs. recall tradeoffs
• Performant scoring, reaching sub-25ms on tuned match-hooks levels

You can explore the fuzzy search in action in our sample usage notebook, which walks through:

• Matching barangay names with partial or noisy input
• Ranking candidates by similarity score
• Profiling latency and performance for scalable deployments

🧪 Why It Matters

Civic datasets are notoriously messy. Barangay names may be abbreviated, misspelled, or inconsistently formatted across sources. This fuzzy search module is designed to handle that complexity while remaining:

• Reproducible: deterministic scoring and matching
• Transparent: every match is explainable and auditable
• Scalable: optimized for sub-30ms matching

🛠️ Under the Hood

The fuzzy search engine is built with:

• rapidfuzz for fast, accurate similarity scoring
• pandas for preloading partial functions, locked-and-loaded for the next string to match

📣 Call for Feedback

If you're working with civic data, barangay-level analytics, or just love noisy data handling, I’d love your feedback. Try it out, raise issues, or share your use cases!