Fuzzy-Searching

PharmAlchemy Fuzzy Semantic Search Module

This tool enables fuzzy semantic search over curated datasets of genes, drugs, and diseases. It was developed to enhance user accessibility within the PharmAlchemy knowledge platform.

🔍 Features

• Fuzzy matching using SequenceMatcher with adjustable thresholds
• Real-time search with confidence scoring
• Synonym-aware query expansion
• Modular design for adding new datasets
• Tkinter-based GUI

📂 Datasets

• g_final.csv: Gene symbols, synonyms, UniProt IDs
• DrugBank Structure Links.csv: Names, formulas, InChI/SMILES, CIDs
• HSDN-Symptoms-DO.tsv: Symptoms, diseases, DOIDs

🚀 Usage

python PhAlSemantic.py

Select a dataset, choose the search field, and begin typing to receive live-ranked suggestions.

---

🧩 Adding a New Dataset

To add a new searchable dataset (e.g., pathways, enzymes):

1. Prepare Your CSV/TSV File

   • Ensure the file has no null values in searchable columns.
   • Format all string fields as lowercase, whitespace-trimmed.
   • Each row should represent a single entry.

2. Create a Loader Function Define a new function (e.g., load_pathway_data) in PhAlSemantic.py modeled after load_gene_data.

3. Register the Dataset Add a new entry to the datasets dictionary:

   'Pathways': {
       'csv_filename': 'pathway_data.csv',
       'search_options': ['pathway_name', 'pathway_id'],
       'search_labels': ['Pathway Name', 'Pathway ID'],
       'load_function': load_pathway_data
   }

---

➕ Adding New Search Parameters

To allow searching by a new column (e.g., EC Number):

1. Ensure the Column Exists The field must be present and complete in the dataset.

2. Add to Index In the dataset loader function:

   ec_index = {}
   for entry in candidate_data:
       ec = entry.get('EC_Number', '').lower()
       if ec:
           ec_index[ec] = entry
   indexes['EC_Number'] = ec_index

3. Expose in UI Add the column to search_options and search_labels in your dataset config.

---

📜 License

MIT License for code. Attribution required for datasets (e.g., DrugBank, DOID, etc.).

📘 Citation

If you use this tool, please cite the accompanying report: "Enhancing Biomedical Data Accessibility via Modular Semantic Search: A Fuzzy Matching System for PharmAlchemy"