# Claude Log Analyzer - Demo Script
This directory contains an automated demonstration script that showcases all features of the Claude Log Analyzer.
## Quick Start
```bash
# Run the full demo
./demo.sh
```
The script will:
1. Build the release binary (if needed)
2. Create an `examples/` directory for outputs
3. Run through all major features
4. Generate sample outputs in multiple formats
## What Gets Demonstrated
### 1. Basic Tool Analysis
Shows overall tool usage statistics from your Claude sessions.
### 2. Tool-Specific Filtering
Demonstrates filtering by specific tool (e.g., Bash, Read, Edit).
### 3. Export Formats
Generates outputs in:
- **JSON** (`examples/tools.json`) - For programmatic analysis
- **CSV** (`examples/tools.csv`) - For spreadsheet import
- **Markdown** (`examples/tools.md`) - For documentation
### 4. Agent-Tool Correlation Matrix
Shows which AI agents use which tools most frequently.
### 5. Tool Chain Detection
Identifies common sequences of tools used together (e.g., Read → Edit → Write).
### 6. Knowledge Graph Search
Demonstrates semantic search across tool usage (e.g., finding deployment-related tools).
### 7. Sorting Methods
Shows different ways to sort results:
- By frequency (most used)
- By recency (most recent)
- Alphabetically
### 8. Self-Analysis
Analyzes how this very project was built using Claude Code.
### 9. Session Summary
Provides high-level statistics across all sessions.
### 10. Timeline Visualization
Generates an interactive HTML timeline of tool usage.
## Output Directory
All outputs are saved to `examples/`:
```
examples/
├── output-basic.txt # Basic tool analysis
├── output-tool-filter.txt # Tool-specific filtering
├── tools.json # JSON export
├── tools.csv # CSV export
├── tools.md # Markdown export
├── output-correlation.txt # Agent-tool correlation
├── output-chains.txt # Tool chains
├── output-kg-search.txt # Knowledge graph search
├── output-sort-recent.txt # Recent sorting
├── output-self-analysis.txt # Self-analysis
├── output-summary.txt # Summary statistics
└── timeline.html # Interactive timeline
```
## Manual Feature Testing
If you want to test specific features manually:
```bash
# Basic tool analysis
./target/release/cla tools
# Filter by tool
./target/release/cla tools --tool Bash
# Show correlation matrix
./target/release/cla tools --show-correlation
# Show tool chains
./target/release/cla tools --show-chains
# Knowledge graph search
./target/release/cla tools --kg-search "deploy OR test"
# Export to JSON
./target/release/cla tools --format json -o output.json
# Analyze specific session
./target/release/cla tools ~/.claude/projects/YOUR-PROJECT-DIR/
# Generate timeline
./target/release/cla timeline --output timeline.html
# Watch for new sessions
./target/release/cla watch
```
## Requirements
- Rust toolchain (for building)
- Claude Code sessions in `~/.claude/projects/`
- At least one session file to analyze
## Expected Output
The demo takes approximately 10-30 seconds to complete, depending on:
- Number of sessions in your `~/.claude/projects/` directory
- Whether the release binary needs to be built
- System performance
## Troubleshooting
**No sessions found:**
- Ensure you have used Claude Code before
- Check that `~/.claude/projects/` exists and contains session directories
**Build errors:**
- Ensure Rust toolchain is installed: `rustup --version`
- Try cleaning and rebuilding: `cargo clean && cargo build --release --features terraphim`
**Permission errors:**
- Make script executable: `chmod +x demo.sh`
- Check read permissions on `~/.claude/projects/`
## Next Steps
After running the demo:
1. **Explore the JSON output** for programmatic integration:
```bash
jq . examples/tools.json
```
2. **View the timeline** in your browser:
```bash
open examples/timeline.html
```
3. **Analyze specific patterns** using the correlation matrix:
```bash
cat examples/output-correlation.txt
```
4. **Integrate with your workflow** using the CSV export:
```bash
open examples/tools.csv
```
## Demo Script Features
The script includes:
- **Color-coded output** for better readability
- **Timing information** for each step
- **Error handling** with graceful failures
- **Automatic directory creation** for outputs
- **Session count verification** before running
- **Progress indicators** throughout execution