## 🚀 CACHED WEBSEARCH - Smart Search with Intelligence
**Claude, execute INTELLIGENT websearch with automatic caching and redundancy elimination.**
### ⚡ **QUICK USAGE FOR ALL 10X COMMANDS**
**REPLACE:** `websearch "your query"`
**WITH:** `/cached_websearch_10x "your query"`
### 🧠 **AUTOMATIC EXECUTION SEQUENCE**
**STEP 1: Cache Analysis**
```sql
-- Check for similar searches (last 30 days)
SELECT
query,
results_file,
timestamp,
relevance_score,
usage_count
FROM search_cache
WHERE
(LOWER(query) LIKE '%keyword1%' OR
LOWER(query) LIKE '%keyword2%' OR
LOWER(keywords) LIKE '%topic%')
AND timestamp > datetime('now', '-30 days')
AND status = 'active'
ORDER BY relevance_score DESC, timestamp DESC
LIMIT 3;
```
**STEP 2: Intelligence Decision**
- **90%+ similarity**: Return cached results + brief summary
- **70-89% similarity**: Return cached + supplement with focused search
- **<70% similarity**: Execute full new search + cache results
**STEP 3: Automatic Execution**
```bash
# Generate search hash
QUERY_CLEAN=$(echo "$1" | tr '[:upper:]' '[:lower:]' | sed 's/[^a-z0-9 ]//g')
SEARCH_HASH=$(echo "$QUERY_CLEAN" | sha256sum | cut -d' ' -f1)
# Check cache
CACHE_MATCH=$(sqlite3 Knowledge/intelligence/search_cache/index/cache.db \
"SELECT query, results_file, relevance_score FROM search_cache
WHERE query_hash = '$SEARCH_HASH' OR
(keywords LIKE '%$(echo $QUERY_CLEAN | cut -d' ' -f1)%' AND
keywords LIKE '%$(echo $QUERY_CLEAN | cut -d' ' -f2)%')
ORDER BY relevance_score DESC LIMIT 1;")
if [ ! -z "$CACHE_MATCH" ]; then
RELEVANCE=$(echo "$CACHE_MATCH" | cut -d'|' -f3)
if [ "$RELEVANCE" -gt 80 ]; then
echo "📋 Using cached search (${RELEVANCE}% match):"
RESULTS_FILE=$(echo "$CACHE_MATCH" | cut -d'|' -f2)
cat "Knowledge/intelligence/search_cache/$RESULTS_FILE"
# Update usage statistics
sqlite3 Knowledge/intelligence/search_cache/index/cache.db \
"UPDATE search_cache SET usage_count = usage_count + 1,
last_accessed = datetime('now') WHERE results_file = '$RESULTS_FILE';"
exit 0
fi
fi
# Execute new search
echo "🔍 Executing new search: $1"
SEARCH_RESULTS=$(websearch "$1")
# Cache the results
TIMESTAMP=$(date +%Y%m%d_%H%M%S)
CACHE_FILE="by_date/$(date +%Y-%m)/${TIMESTAMP}_${SEARCH_HASH:0:8}.md"
FULL_PATH="Knowledge/intelligence/search_cache/$CACHE_FILE"
# Create directory if needed
mkdir -p "$(dirname "$FULL_PATH")"
# Store results
cat > "$FULL_PATH" << EOF
# Search Results: $1
**Date**: $(date)
**Keywords**: $QUERY_CLEAN
**Hash**: $SEARCH_HASH
**Type**: websearch
## Results
$SEARCH_RESULTS
---
*Cached by 10X Agentic Intelligence System*
EOF
# Update database
sqlite3 Knowledge/intelligence/search_cache/index/cache.db << EOF
INSERT INTO search_cache
(query, keywords, results_file, query_hash, file_size, domain)
VALUES
('$1', '$QUERY_CLEAN', '$CACHE_FILE', '$SEARCH_HASH',
$(wc -c < "$FULL_PATH"),
'$(echo "$QUERY_CLEAN" | grep -o -E "(development|security|performance|architecture|ai|tools)" | head -1)');
UPDATE search_metrics
SET total_searches = total_searches + 1
WHERE date = date('now');
EOF
echo "$SEARCH_RESULTS"
```
### 📊 **INTEGRATION WITH 10X COMMANDS**
**Update ALL 10X commands to replace:**
```markdown
# OLD WAY:
- **websearch**: "react performance optimization 2025"
# NEW WAY:
- **cached_websearch_10x**: "react performance optimization 2025"
```
### 🎯 **INTELLIGENT SEARCH ENHANCEMENT**
**Context-Aware Searching:**
- Automatically adds project context to searches
- Includes technology stack in search terms
- Filters results by recency and relevance
**Example Enhancement:**
```bash
# Original query: "testing best practices"
# Enhanced query: "testing best practices javascript react 2025 enterprise"
# Based on: Project detection + current year + enterprise context
```
### 📈 **CACHE PERFORMANCE MONITORING**
**View Cache Statistics:**
```sql
SELECT
date,
total_searches,
cache_hits,
ROUND(CAST(cache_hits AS FLOAT) / total_searches * 100, 2) as hit_rate_percent,
api_calls_saved
FROM search_metrics
WHERE date >= date('now', '-7 days')
ORDER BY date DESC;
```
**Top Cached Searches:**
```sql
SELECT
query,
usage_count,
last_accessed,
ROUND(relevance_score, 2) as relevance
FROM search_cache
WHERE status = 'active'
ORDER BY usage_count DESC
LIMIT 10;
```
### 🔧 **MAINTENANCE COMMANDS**
**Clean Old Cache (30+ days):**
```sql
UPDATE search_cache
SET status = 'archived'
WHERE timestamp < datetime('now', '-30 days')
AND usage_count < 2;
```
**Optimize Cache Database:**
```sql
VACUUM;
REINDEX;
ANALYZE;
```
### ✅ **SUCCESS METRICS**
**Expected Performance:**
- **70% cache hit rate** for repeated searches
- **80% reduction** in redundant API calls
- **3x faster** response time for cached queries
- **Cumulative intelligence** builds over time
**Daily Usage Report:**
```bash
echo "📊 Cache Performance Today:"
echo "Total Searches: $(sqlite3 cache.db "SELECT total_searches FROM search_metrics WHERE date = date('now');")"
echo "Cache Hits: $(sqlite3 cache.db "SELECT cache_hits FROM search_metrics WHERE date = date('now');")"
echo "API Calls Saved: $(sqlite3 cache.db "SELECT api_calls_saved FROM search_metrics WHERE date = date('now');")"
```
### 🔄 **AUTOMATIC PATTERN LEARNING**
The system automatically:
- **Identifies** common search patterns
- **Suggests** related searches
- **Pre-caches** likely follow-up queries
- **Optimizes** search terms based on successful results
### 📊 **STRUCTURED DATA OUTPUT WITH ML TRAINING PREPARATION**
**Multi-System Storage Architecture:**
```yaml
# Cached Search Report
filename: Knowledge/intelligence/cached_search_$(date +%Y-%m-%d_%H-%M-%S).md
frontmatter:
type: cached_search
timestamp: $(date -Iseconds)
classification: search_intelligence
ml_labels: [cache_efficiency, search_optimization, knowledge_reuse]
success_metrics: [cache_hit_rate, response_time, api_savings]
cross_references: [search_patterns, cache_strategies, intelligence_systems]
```
**Redundant Storage with Intelligent Classification:**
- **Primary**: `smart_memory_unified` - Unified orchestration with automatic content classification
- **Secondary**: `chroma-rag` - Vector embeddings for search pattern matching and query similarity
- **Tertiary**: `sqlite` - Structured metrics with ML training labels and effectiveness scoring
- **Quaternary**: `Knowledge/` files - Persistent markdown with consistent frontmatter metadata
**ML Training Data Structure:**
```json
{
"cached_search_session": {
"timestamp": "$(date -Iseconds)",
"features": {
"cache_hit_rate": 0.78,
"search_optimization_score": 0.85,
"query_complexity": 0.72,
"knowledge_reuse_effectiveness": 0.88
},
"outcomes": {
"search_efficiency": 0.87,
"api_call_reduction": 0.80,
"response_time_improvement": 0.75
},
"classification_labels": ["efficient_cache", "optimized_search", "intelligent_reuse"],
"success_probability": 0.85
}
}
```
**Cross-System Synchronization:**
- **chroma-rag**: Create semantic embeddings for search patterns and query optimization
- **smart_memory_unified**: Store search methodologies with automatic classification routing
- **sqlite**: Track search metrics and cache correlations for ML model training
- **Knowledge/patterns/**: Archive successful search patterns with effectiveness scoring
**EXECUTE FOR ANY SEARCH:** Use `/cached_websearch_10x` instead of `websearch` in ALL 10X commands for intelligent caching, dramatically improved performance, and ML training preparation.
