5.3 KiB
5.3 KiB
Core CLI Workflows Test Scenarios
This document outlines test scenarios for the most commonly used, business-critical CLI commands that represent the core user journeys in the AITBC ecosystem.
1. Core Workflow: Client Job Submission Journey
This scenario traces a client's path from generating a job to receiving the computed result.
Scenario 1.1: Submit a Job
- Command:
aitbc client submit --type inference --model "llama3" --data '{"prompt":"Hello AITBC"}' - Description: Submit a new AI inference job to the network.
- Expected Output: Success message containing the
job_idand initial status (e.g., "pending").
Scenario 1.2: Check Job Status
- Command:
aitbc client status <job_id> - Description: Poll the coordinator for the current status of the previously submitted job.
- Expected Output: Status indicating the job is queued, processing, or completed, along with details like assigned miner and timing.
Scenario 1.3: Retrieve Job Result
- Command:
aitbc client result <job_id> - Description: Fetch the final output of a completed job.
- Expected Output: The computed result payload (e.g., the generated text from the LLM) and proof of execution if applicable.
2. Core Workflow: Miner Operations Journey
This scenario traces a miner's path from registering hardware to processing jobs.
Scenario 2.1: Register as a Miner
- Command:
aitbc miner register --gpus "1x RTX 4090" --price-per-hour 0.5 - Description: Register local hardware with the coordinator to start receiving jobs.
- Expected Output: Success message containing the assigned
miner_idand confirmation of registered capabilities.
Scenario 2.2: Poll for a Job
- Command:
aitbc miner poll - Description: Manually check the coordinator for an available job matching the miner's capabilities.
- Expected Output: If a job is available, details of the job (Job ID, type, payload) are returned and the job is marked as "processing" by this miner. If no job is available, a "no jobs in queue" message.
Scenario 2.3: Mine with Local Ollama (Automated)
- Command:
aitbc miner mine-ollama --model llama3 --continuous - Description: Start an automated daemon that polls for jobs, executes them locally using Ollama, submits results, and repeats.
- Expected Output: Continuous log stream showing: polling -> job received -> local inference execution -> result submitted -> waiting.
3. Core Workflow: Wallet & Financial Operations
This scenario covers basic token management required to participate in the network.
Scenario 3.1: Create a New Wallet
- Command:
aitbc wallet create --name test_wallet - Description: Generate a new local keypair and wallet address.
- Expected Output: Success message displaying the new wallet address and instructions to securely backup the seed phrase (which may be displayed once).
Scenario 3.2: Check Wallet Balance
- Command:
aitbc wallet balance - Description: Query the blockchain for the current token balance of the active wallet.
- Expected Output: Display of available balance, staked balance, and total balance.
Scenario 3.3: Client Job Payment
- Command:
aitbc client pay <job_id> --amount 10 - Description: Authorize payment from the active wallet to fund a submitted job.
- Expected Output: Transaction hash confirming the payment, and the job status updating to "funded".
4. Core Workflow: GPU Marketplace
This scenario covers interactions with the decentralized GPU marketplace.
Scenario 4.1: Register GPU on Marketplace
- Command:
aitbc marketplace gpu register --model "RTX 4090" --vram 24 --hourly-rate 0.5 - Description: List a GPU on the open marketplace for direct rental or specific task assignment.
- Expected Output: Success message with a
listing_idand confirmation that the offering is live on the network.
Scenario 4.2: List Available GPU Offers
- Command:
aitbc marketplace offers list --model "RTX 4090" - Description: Browse the marketplace for available GPUs matching specific criteria.
- Expected Output: A table showing available GPUs, their providers, reputation scores, and hourly pricing.
Scenario 4.3: Check Pricing Oracle
- Command:
aitbc marketplace pricing --model "RTX 4090" - Description: Get the current average, median, and suggested market pricing for a specific hardware model.
- Expected Output: Statistical breakdown of current market rates to help providers price competitively and users estimate costs.
5. Advanced Workflow: AI Agent Execution
This scenario covers the deployment of autonomous AI agents.
Scenario 5.1: Create Agent Workflow
- Command:
aitbc agent create --name "data_analyzer" --type "analysis" --config agent_config.json - Description: Define a new agent workflow based on a configuration file.
- Expected Output: Success message with
agent_idindicating the agent is registered and ready.
Scenario 5.2: Execute Agent
- Command:
aitbc agent execute <agent_id> --input "Analyze Q3 financial data" - Description: Trigger the execution of the configured agent with a specific prompt/input.
- Expected Output: Streamed or final output showing the agent's thought process, actions taken (tool use), and final result.