Run Generation Pipeline

Runs the full question generation pipeline locally using run_generation.py.

Usage

code

/run-generation [--type <type>] [--difficulty <difficulty>] [--count <n>] [--dry-run]

Arguments

Argument	Required	Default	Description
`--type`	No	all	Question type: `math`, `logic`, `pattern`, `spatial`, `verbal`, `memory`
`--difficulty`	No	all	Difficulty level: `easy`, `medium`, `hard`
`--count`	No	50	Number of questions to generate
`--dry-run`	No	false	Generate and evaluate but don't insert to database

Implementation

Step 1: Parse Arguments

Extract the optional arguments from the user's input:

•--type <type> or -t <type> - maps to --types <type>
•--difficulty <difficulty> or -d <difficulty> - maps to --difficulties <difficulty>
•--count <n> or -c <n> - maps to --count <n>
•--dry-run - adds --dry-run flag

Validate:

•Type must be one of: math, logic, pattern, spatial, verbal, memory
•Difficulty must be one of: easy, medium, hard
•Count must be a positive integer

Step 2: Build Command

Construct the command with only the specified arguments:

bash

cd /Users/mattgioe/aiq/question-service && source venv/bin/activate && export $(grep -v '^#' .env | xargs) && python run_generation.py --async --async-judge --verbose [additional args based on input]

Examples:

No arguments (generate all types/difficulties):

bash

cd /Users/mattgioe/aiq/question-service && source venv/bin/activate && export $(grep -v '^#' .env | xargs) && python run_generation.py --count 50 --async --async-judge --verbose

With type only:

bash

cd /Users/mattgioe/aiq/question-service && source venv/bin/activate && export $(grep -v '^#' .env | xargs) && python run_generation.py --types math --count 50 --async --async-judge --verbose

With difficulty only:

bash

cd /Users/mattgioe/aiq/question-service && source venv/bin/activate && export $(grep -v '^#' .env | xargs) && python run_generation.py --difficulties hard --count 50 --async --async-judge --verbose

With both:

bash

cd /Users/mattgioe/aiq/question-service && source venv/bin/activate && export $(grep -v '^#' .env | xargs) && python run_generation.py --types spatial --difficulties easy --count 50 --async --async-judge --verbose

Dry run:

bash

cd /Users/mattgioe/aiq/question-service && source venv/bin/activate && export $(grep -v '^#' .env | xargs) && python run_generation.py --types math --count 10 --async --async-judge --verbose --dry-run

Step 3: Run and Report

Execute the command and report results including:

•Number of questions generated
•Number approved by judge
•Number unique after deduplication
•Number inserted to database
•Approval rate

Exit Codes

Code	Meaning	Action
0	Success	All questions generated and inserted
1	Partial failure	Some questions failed - check logs
2	Complete failure	No questions generated - check API keys/quotas
3	Config error	Missing or invalid environment variables
4	Database error	Cannot connect to database

Examples

code

/run-generation

Generate 50 questions of all types and difficulties.

code

/run-generation --type math

Generate 50 math questions at all difficulty levels.

code

/run-generation --difficulty hard

Generate 50 hard questions of all types.

code

/run-generation --type pattern --difficulty easy --count 20

Generate 20 easy pattern questions.

code

/run-generation --type spatial --count 30 --dry-run

Generate 30 spatial questions without inserting to database (for testing).

Requirements

•question-service/.env must contain valid API keys (OPENAI_API_KEY, ANTHROPIC_API_KEY, etc.)
•question-service/venv must be set up with dependencies installed
•DATABASE_URL must be configured in .env