📖 Custom Search

What is it?

A unified search interface that lets you choose between different search methods.

How to use:

Enter your search query in natural language
Select search type: Semantic, Hybrid, or Metadata
Adjust the result limit (1-100)
Click Search

Expected Results:

Returns dev requests matching your query with relevance scores, showing ID, text, type, module, status, and keywords.

Example:

workflow status errors with SQL triggers

📊 Statistics

What is it?

Overview of all dev requests in the database with aggregated metrics.

How to use:

Click "Load Statistics" button
View charts and metrics

Expected Results:

Displays:

Total number of requests
Distribution by type (Bug, Enhancement, etc.)
Top modules
Top keywords
Status breakdown

🔍 Metadata Query

What is it?

Filter dev requests by specific metadata fields like type, module, and status.

How to use:

Select Type from dropdown (e.g., Bug, Enhancement)
Select Module from dropdown (e.g., SQL*LIMS)
Select Status from dropdown (e.g., Open, Closed)
Set result limit
Click Query

Expected Results:

Returns all dev requests matching the selected filters. Leave filters empty to see all requests.

Tip:

Combine multiple filters to narrow down results.

🏷️ Keyword Search

What is it?

Search for dev requests containing specific keywords extracted from their descriptions.

How to use:

Enter keywords separated by commas
Example: SQL trigger, status transition, workflow
Set result limit
Click Search

Expected Results:

Returns dev requests that contain ANY of the specified keywords in their excerpt_keywords field.

Note:

Keywords are case-insensitive and matched exactly.

🧠 Semantic Search

What is it?

AI-powered search that understands the meaning of your query, not just keywords.

How to use:

Enter a natural language query describing what you're looking for
Example: issues with database triggers causing workflow problems
Set result limit
Click Search

Expected Results:

Returns semantically similar dev requests ranked by relevance score. Results may not contain exact keywords but are conceptually related.

Best for:

Finding related issues when you don't know exact keywords or want conceptually similar requests.

⚡ Hybrid Search

What is it?

Combines semantic search with keyword matching for more accurate results.

How to use:

Enter your search query
Adjust Alpha slider (0=keyword only, 1=semantic only, 0.5=balanced)
Set result limit
Click Search

Expected Results:

Returns results that balance semantic similarity and keyword matching based on the alpha value.

Alpha Guide:

0.0 - Pure keyword search
0.5 - Balanced (recommended)
1.0 - Pure semantic search

🔗 Similar Requests

What is it?

Find dev requests similar to a specific request by its ID.

How to use:

Enter a DevRequest ID (e.g., DR-12345)
Set result limit
Click Find Similar

Expected Results:

Returns dev requests that are semantically similar to the specified request, ranked by similarity score.

Use case:

Great for finding duplicate issues, related bugs, or similar enhancement requests.

🐛 Advanced Grouping

What is it?

Groups items by similarity using multiple methods: keyword similarity, exact keyword matching, business domain filtering, or NLP semantic clustering.

Metadata Filters:

Type: Filter by Bug, Enhancement, etc.
Module: Filter by specific module
Status: Filter by Open, Closed, etc.
Exclude Modules: Exclude multiple modules from results (comma-separated)

Grouping Methods:

1. Similarity (Jaccard): Groups items with similar keyword sets

Uses Jaccard similarity coefficient
Threshold: 0.5-0.6 (loose), 0.7-0.8 (balanced), 0.9+ (strict)

2. Keyword Match: Groups items sharing exact keywords

Requires minimum number of common keywords
Good for finding items with specific shared terms

3. Domain-Based: Groups items by business domain

Define domain with comma-separated keywords
Returns items matching any domain keyword
Use "Browse Top Keywords" to discover relevant terms

4. NLP Semantic Clustering: Groups items by semantic meaning

Uses vector embeddings and cosine similarity
Groups conceptually similar items even with different keywords
Threshold: 0.6-0.7 (loose), 0.75-0.85 (balanced), 0.9+ (strict)
Best for finding semantically related issues

Use case:

Identify patterns, find duplicates, prioritize common issues, and understand trends across your dev requests.

🔍 DevRequest Query Engine

Custom Search

Overall Statistics

Node Stats Visualization

Node Details

Query by Metadata

Keyword Search

Semantic Search

Hybrid Search

Find Similar Requests

Advanced Grouping

Custom Search

Overall Statistics

Node Stats Visualization

Node Details

Query by Metadata

Keyword Search

Semantic Search

Hybrid Search

Find Similar Requests

Advanced Grouping

Top 100 Keywords