Search Commands¶

./rag-cli search query COLLECTION_ID QUERY [OPTIONS]

./rag-cli search query col_123abc "What are the main benefits of machine learning?"

./rag-cli search query col_123abc "Explain deep learning architectures" \
  --max-chunks 8 \
  --similarity-threshold 0.6 \
  --temperature 0.2 \
  --include-chunks

./rag-cli search query col_123abc "summarize recent research findings" \
  --format json \
  --max-tokens 256

./rag-cli search query col_research "What were the accuracy results in the CNN study?" \
  --similarity-threshold 0.8 \
  --include-sources

./rag-cli search similar COLLECTION_ID TEXT [OPTIONS]

./rag-cli search similar col_123abc "neural network architectures"

./rag-cli search similar col_123abc "deep learning applications" \
  --search-type documents \
  --max-results 5

./rag-cli search similar col_123abc "convolutional neural networks" \
  --include-content \
  --similarity-threshold 0.7 \
  --format json

./rag-cli search similar col_research "machine learning algorithms" \
  --group-by-document \
  --max-results 15

./rag-cli search history [OPTIONS]

./rag-cli search history

./rag-cli search history --collection col_123abc --limit 10

./rag-cli search history --since 2024-01-10 --format json

./rag-cli search history --filter "machine learning"

./rag-cli search replay QUERY_ID [OPTIONS]

./rag-cli search replay query_abc123

./rag-cli search replay query_abc123 --compare

./rag-cli search replay query_abc123 --update-settings

./rag-cli search explain QUERY_ID [OPTIONS]

./rag-cli search explain query_abc123

./rag-cli search explain query_abc123 \
  --include-embeddings \
  --include-ranking \
  --format json

$ ./rag-cli search query col_123abc "test query"

❌ Error: No LLM provider available for user.
   Please configure an LLM provider before searching.

   Fix: ./rag-cli providers create watsonx --api-key YOUR_KEY --project-id YOUR_PROJECT

$ ./rag-cli search query col_invalid "test query"

❌ Error: Collection not found or access denied.
   Collection ID: col_invalid

   Fix: Check collection ID or verify access permissions

$ ./rag-cli search query col_123abc "test query"

❌ Error: Failed to create default pipeline.
   Reason: Invalid LLM provider configuration

   Fix: Update your LLM provider settings

#!/bin/bash
query="artificial intelligence applications"
collections=("col_research" "col_technical" "col_business")

echo "🔍 Multi-Collection Search: '$query'"
echo "========================================="

for collection in "${collections[@]}"; do
    collection_name=$(./rag-cli collections get "$collection" --format json | jq -r '.name')
    echo ""
    echo "📚 Searching: $collection_name"
    echo "--------------------------------"

    ./rag-cli search query "$collection" "$query" \
      --max-chunks 3 \
      --format text \
      --include-sources
done

#!/bin/bash
collection_id="col_123abc"
queries_file="queries.txt"
results_dir="./search_results"

mkdir -p "$results_dir"

echo "📝 Processing batch queries..."

while IFS= read -r query; do
    if [ -n "$query" ] && [[ ! "$query" =~ ^[[:space:]]*# ]]; then
        echo "Querying: $query"

        # Create safe filename
        filename=$(echo "$query" | tr ' ' '_' | tr -cd '[:alnum:]_' | cut -c1-50)

        ./rag-cli search query "$collection_id" "$query" \
          --format json \
          --include-chunks \
          > "$results_dir/${filename}.json"
    fi
done < "$queries_file"

echo "✅ Batch processing completed. Results in: $results_dir"

#!/bin/bash
echo "📊 Search Analytics Dashboard"
echo "============================"

# Recent search activity
recent_searches=$(./rag-cli search history --limit 100 --format json)

# Total queries
total=$(echo "$recent_searches" | jq '.total')
echo "Recent Queries: $total"

# Average response time
avg_time=$(echo "$recent_searches" | jq '[.queries[].response_time_ms] | add / length')
printf "Average Response Time: %.1fs\n" $(echo "$avg_time / 1000" | bc -l)

# Top collections
echo ""
echo "Top Collections by Query Volume:"
echo "$recent_searches" | \
jq -r '.queries | group_by(.collection_id) | sort_by(length) | reverse | .[0:5] | .[] | [.[0].collection_name, length] | @tsv' | \
while IFS=$'\t' read -r collection count; do
    echo "  - $collection: $count queries"
done

# Query patterns
echo ""
echo "Common Query Terms:"
echo "$recent_searches" | \
jq -r '.queries[].query' | \
tr '[:upper:]' '[:lower:]' | \
tr -s '[:punct:][:space:]' '\n' | \
sort | uniq -c | sort -rn | head -10 | \
while read count term; do
    echo "  - $term: $count occurrences"
done

# Performance trends
echo ""
echo "Performance by Hour (last 24h):"
echo "$recent_searches" | \
jq -r '.queries[] | [(.timestamp | strftime("%H")), .response_time_ms] | @tsv' | \
awk '{sum[$1]+=$2; count[$1]++} END {for(h in sum) printf "  %02d:00 - Avg: %.1fs (%d queries)\n", h, sum[h]/(count[h]*1000), count[h]}' | \
sort

#!/bin/bash
collection_id="col_123abc"
query="machine learning algorithms"

echo "🎯 Similarity Threshold Optimization"
echo "==================================="

for threshold in 0.5 0.6 0.7 0.8 0.9; do
    echo "Testing threshold: $threshold"

    result=$(./rag-cli search query "$collection_id" "$query" \
      --similarity-threshold "$threshold" \
      --format json \
      --max-chunks 5)

    chunks=$(echo "$result" | jq '.retrieved_chunks | length')
    avg_score=$(echo "$result" | jq '[.retrieved_chunks[].similarity_score] | add / length')
    time=$(echo "$result" | jq '.response.response_time_ms')

    printf "  Chunks: %d, Avg Score: %.3f, Time: %dms\n" "$chunks" "$avg_score" "$time"
done

#!/bin/bash
collection_id="col_123abc"

echo "🏥 Search Index Health Check"
echo "============================"

# Test query performance
test_queries=("machine learning" "neural networks" "deep learning" "AI applications")

for query in "${test_queries[@]}"; do
    echo "Testing: $query"

    result=$(./rag-cli search similar "$collection_id" "$query" \
      --max-results 10 \
      --format json)

    results_count=$(echo "$result" | jq '.results | length')
    avg_similarity=$(echo "$result" | jq '[.results[].similarity_score] | add / length')
    search_time=$(echo "$result" | jq '.search_time_ms')

    printf "  Results: %d, Avg Similarity: %.3f, Time: %dms\n" "$results_count" "$avg_similarity" "$search_time"
done

# Check for performance degradation
echo ""
echo "Performance Recommendations:"
echo "- Results < 5: Consider lowering similarity threshold"
echo "- Avg Similarity < 0.7: May need better query formulation"
echo "- Time > 3000ms: Consider index optimization"

#!/bin/bash
# Simple Slack webhook integration for search

slack_webhook="$SLACK_WEBHOOK_URL"
collection_id="col_knowledge"

query="$1"
if [ -z "$query" ]; then
    echo "Usage: $0 'your search query'"
    exit 1
fi

echo "🔍 Searching knowledge base..."

# Perform search
result=$(./rag-cli search query "$collection_id" "$query" \
  --max-chunks 3 \
  --format json \
  --max-tokens 200)

# Extract answer and sources
answer=$(echo "$result" | jq -r '.response.answer')
sources=$(echo "$result" | jq -r '.sources[] | "• " + .document_title')

# Format for Slack
slack_message=$(cat <<EOF
{
  "text": "Knowledge Base Search Results",
  "blocks": [
    {
      "type": "section",
      "text": {
        "type": "mrkdwn",
        "text": "*Query:* $query\n\n*Answer:*\n$answer"
      }
    },
    {
      "type": "section",
      "text": {
        "type": "mrkdwn",
        "text": "*Sources:*\n$sources"
      }
    }
  ]
}
EOF
)

# Send to Slack
curl -X POST -H 'Content-type: application/json' \
  --data "$slack_message" \
  "$slack_webhook"

echo "✅ Results sent to Slack"