Skip to main content
Version: 3.2.X

Dictionary & Spell Checker ("Did You Mean?")

The ABCD OPAC features a mechanism to assist users when a search returns zero results. It compares the user's input against a static list of terms and suggests similar existing words using the Levenshtein distance algorithm.

Example: User types 'Enrgy' $\rightarrow$ System suggests: "Did you mean: Energy?"

Script: view_dic.php File Created: bases/[db]/opac/[db].dic

1. The Static Dictionary Concept​

Unlike the A-Z Browser (which queries the live index), the spell checker uses a static text file (.dic) stored in the database's OPAC folder.

  • Performance: Searching a static text file is faster and safer than running fuzzy logic queries on the main database index for every typo.
  • Maintenance: Because it is static, you must regenerate this file periodically to include newly cataloged terms.

2. Managing the Dictionary​

Access: Menu Databases > Manage Dictionary (or via the link in the A-Z Browsing module).

The interface allows three main actions:

A. View Terms​

Allows you to inspect the current contents of the .dic file. Useful to verify if the file exists and if the terms are clean (readable).

B. Update/Regenerate (The Core Function)​

This process reads the database's Inverted File, extracts terms, cleans them, and saves the .dic file.

Generation Modes:

  1. Based on A-Z Configuration (Recommended):
    • It reads the [db].ix file.
    • It extracts only terms belonging to the configured prefixes (e.g., TI_, AU_).
    • Result: A clean, relevant dictionary without "noise" (like IDs, dates, or internal control numbers).
  2. Complete Database Dump:
    • It extracts all keys from the Inverted File.
    • Result: Maximum coverage, but may include technical data that you don't want to suggest to users.

C. Upload​

Allows you to upload a pre-compiled list of terms (e.g., a standard Thesaurus list) instead of generating it from the database.

3. The Generation Process (Technical)​

When you click Regenerate, the system performs a sophisticated batch process to handle large databases without timing out.

  1. Initialization: The system calculates the total number of keys in the database.
  2. Batch Processing (AJAX):
    • The script view_dic.php loops through the index in chunks (e.g., 5,000 keys at a time).
    • It filters terms based on the .ix configuration.
    • It cleans the text (removes punctuation, lowercases terms depending on settings).
    • A Progress Bar shows the status in real-time.
  3. Finalization:
    • Once all keys are processed, the system removes duplicates.
    • The final list is sorted alphabetically.
    • The file is saved to bases/[db]/opac/[db].dic.
Best Practice

Always configure your A-Z Browsing (Indices) first. Then, use the "Based on A-Z Configuration" mode to generate the dictionary. This ensures that the system only suggests terms that the user can actually find in specific fields.