sqlfuzz is a random SQLAlchemy query generator for DBMS fuzzing toolchains. sqlfuzz improves upon other random query generators, such as sqlsmith, by leveraging SQLAlchemy's ability to compile the same query for different DBMS dialects. For example, sqlfuzz allows us to easily test the queries that triggers bugs in Postgres to other DBMSs such as MySQL, MariaDB, Sqlite, etc. sqlfuzz can target any dialects that SQLAlchemy supports.
$ pip3 install sqlfuzz
import json
from sqlfuzz.fuzz import Fuzz
f = open("./path/to/probability/conf.json")
prob_conf = json.load(f)
connection_string = "postgresql+psycopg2://{user}:{password}@{host}:{port}/{dbname}"
fuzz = Fuzz(prob_conf, connection_string)
queries = fuzz.gen_orm_queries(count=10)
Probability of the fuzzer chosing to generate a query with following clauses (range: 0-1000)
{
'order': 500,
'limit': 500,
'group': 500,
'left': 500,
'inner': 500,
'full': 500,
'scalar': 500,
'true': 500,
'func_expr': 500,
'literal_column': 500,
'distinct': 500,
'set': 500,
'offset': 500,
'simple': 500,
'window': 500,
'extractyear': 500,
'extractmonth': 500,
'subquery': 500,
'nested': 500
}