CLI Reference

LSJI includes a command-line interface for training and playing without writing code.

Installation

# Global install
npm install -g lsji

# Or use npx
npx lsji --help

Commands

lsji train

Train the agent.

lsji train [options]

Options:

Option	Description	Default
--episodes <n>	Number of training episodes	200
--pattern <0-3>	Training pattern	0
--batch-size <n>	Database batch size	200
--opponent <type>	Opponent strategy	random
--storage <type>	Storage backend	sqlite
--db-path <path>	Database file path	./lsji.db
--alpha <n>	Learning rate	0.1
--gamma <n>	Discount factor	0.9
--epsilon <n>	Exploration rate	0.1
--json	Output as JSON	false

Training Patterns:

• 0 — Random actions
• 1 — Always Rock
• 2 — Counter previous action
• 3 — Sequential (0,1,2,0,1,2...)

Opponent Strategies:

• random — Random actions
• always_rock — Always plays Rock
• counter — Counters agent's previous action
• sequential — Cycles through actions

Examples:

# Default training
lsji train --episodes 500

# Train against always-rock opponent
lsji train --episodes 100 --pattern 1 --opponent always_rock

# Train with custom hyperparameters
lsji train --episodes 1000 --alpha 0.05 --gamma 0.95 --epsilon 0.2

# Use memory storage (ephemeral)
lsji train --episodes 100 --storage memory

lsji play

Play a single game against the agent.

lsji play --hand <0|1|2> [options]

Options:

Option	Description
--hand <0|1|2>	Your hand: 0=Rock, 1=Scissors, 2=Paper
--opponent <type>	Opponent strategy
--storage <type>	Storage backend
--db-path <path>	Database file path
--json	Output as JSON

Examples:

lsji play --hand 0        # Play Rock
lsji play --hand 1        # Play Scissors
lsji play --hand 2 --json # Play Paper, JSON output

lsji status

Show system status and statistics.

lsji status [options]

Options:

Option	Description
--storage <type>	Storage backend
--db-path <path>	Database file path
--json	Output as JSON

Example:

lsji status --json

Output:

{
  "status": "running",
  "todayTotal": 42,
  "limit": 90000,
  "performance": [
    { "mode": "train", "total": 1000, "win_rate": 65.5 },
    { "mode": "test", "total": 50, "win_rate": 72.0 }
  ],
  "aiBrain": [
    { "state": "0", "action": 0, "q_value": 0.45 },
    { "state": "0", "action": 1, "q_value": 0.12 }
  ]
}

lsji start

Enable training and play.

lsji start [options]

lsji stop

Disable training and play (system paused).

lsji stop [options]

lsji help

Show help message.

lsji help
lsji --help
lsji -h

Environment Variables

Variable	Description	Default
LSJI_STORAGE	Default storage backend	sqlite
LSJI_DB_PATH	Default database path	./lsji.db

Examples

Full Training Session

# Start fresh
rm -f lsji.db

# Train against random opponent
lsji train --episodes 500 --opponent random

# Train against counter opponent
lsji train --episodes 500 --opponent counter

# Check progress
lsji status --json

# Play a few games
lsji play --hand 0
lsji play --hand 1
lsji play --hand 2

Using Memory Storage (CI/Testing)

lsji train --episodes 100 --storage memory
lsji play --hand 0 --storage memory
lsji status --storage memory

Custom Hyperparameters

lsji train \
  --episodes 2000 \
  --alpha 0.05 \
  --gamma 0.95 \
  --epsilon 0.2 \
  --opponent random \
  --storage sqlite \
  --db-path ./custom.db