Secure database query utilities for SQL injection prevention.
This module provides utilities for safe database interactions, ensuring that SQL queries are properly parameterized and identifiers are validated. These utilities should be used whenever database functionality is added to TNFR.
These utilities preserve TNFR structural coherence when persisting:
Safe parameterized query
builder = SecureQueryBuilder() query, params = builder.select("nfr_nodes", ["id", "epi", "nu_f"])\ ... .where("nu_f > ?", 0.5)\ ... .order_by("nu_f", "DESC")\ ... .build() print(query) SELECT id, epi, nu_f FROM nfr_nodes WHERE nu_f > ? ORDER BY nu_f DESC print(params) [0.5]
Validate identifiers before use
table_name = validate_identifier("nfr_nodes") column_name = validate_identifier("nu_f")
"""Secure database query utilities for SQL injection prevention.
This module provides utilities for safe database interactions, ensuring that
SQL queries are properly parameterized and identifiers are validated. These
utilities should be used whenever database functionality is added to TNFR.
Security Principles
-------------------
1. **Parameterized Queries**: Always use placeholders (?, :name) for values
2. **Identifier Validation**: Validate table/column names against whitelist
3. **No String Concatenation**: Never build queries with string concatenation
4. **Input Sanitization**: Validate and sanitize all user inputs
TNFR Structural Context
------------------------
These utilities preserve TNFR structural coherence when persisting:
- EPI (Estructura Primaria de Información)
- NFR (Nodo Fractal Resonante) metadata
- Structural frequency (νf) measurements
- Phase relationships (φ) between nodes
- Coherence metrics C(t) and sense index Si
Example
-------
>>> # Safe parameterized query
>>> builder = SecureQueryBuilder()
>>> query, params = builder.select("nfr_nodes", ["id", "epi", "nu_f"])\\
... .where("nu_f > ?", 0.5)\\
... .order_by("nu_f", "DESC")\\
... .build()
>>> print(query)
SELECT id, epi, nu_f FROM nfr_nodes WHERE nu_f > ? ORDER BY nu_f DESC
>>> print(params)
[0.5]
>>> # Validate identifiers before use
>>> table_name = validate_identifier("nfr_nodes")
>>> column_name = validate_identifier("nu_f")
"""
from __future__ import annotations
import re
from typing import Any
# Valid SQL identifier pattern: alphanumeric and underscores only, 1-64 chars
_VALID_IDENTIFIER_PATTERN = re.compile(r"^[a-zA-Z_][a-zA-Z0-9_]{0,63}$")
# Common SQL keywords that should not be used as identifiers
_SQL_KEYWORDS = frozenset(
{
"SELECT",
"INSERT",
"UPDATE",
"DELETE",
"DROP",
"CREATE",
"ALTER",
"TRUNCATE",
"EXEC",
"EXECUTE",
"UNION",
"FROM",
"WHERE",
"AND",
"OR",
"NOT",
"IN",
"EXISTS",
"JOIN",
"TABLE",
"DATABASE",
"GRANT",
"REVOKE",
}
)
class SQLInjectionError(ValueError):
"""Exception raised when potential SQL injection is detected."""
def validate_identifier(identifier: str, *, allow_keywords: bool = False) -> str:
"""Validate a SQL identifier (table or column name).
This function ensures that identifiers are safe to use in SQL queries
by checking against a whitelist pattern and optionally rejecting SQL keywords.
Parameters
----------
identifier : str
The identifier to validate (table name, column name, etc.)
allow_keywords : bool, optional
If False (default), reject SQL keywords as identifiers
Returns
-------
str
The validated identifier
Raises
------
SQLInjectionError
If the identifier is invalid or potentially unsafe
Example
-------
>>> validate_identifier("nfr_nodes")
'nfr_nodes'
>>> validate_identifier("nu_f_measurements")
'nu_f_measurements'
>>> validate_identifier("DROP") # doctest: +SKIP
Traceback (most recent call last):
...
SQLInjectionError: Identifier 'DROP' is a SQL keyword
>>> validate_identifier("invalid-name") # doctest: +SKIP
Traceback (most recent call last):
...
SQLInjectionError: Invalid identifier 'invalid-name'
"""
if not isinstance(identifier, str):
raise SQLInjectionError(
f"Identifier must be a string, got {type(identifier).__name__}"
)
if not identifier:
raise SQLInjectionError("Identifier cannot be empty")
if not _VALID_IDENTIFIER_PATTERN.match(identifier):
raise SQLInjectionError(
f"Invalid identifier '{identifier}': must contain only alphanumeric "
f"characters and underscores, start with letter or underscore, "
f"and be 1-64 characters long"
)
if not allow_keywords and identifier.upper() in _SQL_KEYWORDS:
raise SQLInjectionError(
f"Identifier '{identifier}' is a SQL keyword and cannot be used"
)
return identifier
def sanitize_string_input(value: str, *, max_length: int = 1000) -> str:
"""Sanitize string input for safe database operations.
This function validates string inputs to prevent SQL injection and ensure
reasonable length constraints. Note: This is NOT a replacement for
parameterized queries, but an additional validation layer.
Parameters
----------
value : str
The string value to sanitize
max_length : int, optional
Maximum allowed length (default: 1000)
Returns
-------
str
The validated string value
Raises
------
SQLInjectionError
If the value is too long or contains suspicious patterns
Example
-------
>>> sanitize_string_input("valid string")
'valid string'
>>> sanitize_string_input("a" * 10000) # doctest: +SKIP
Traceback (most recent call last):
...
SQLInjectionError: Input exceeds maximum length...
"""
if not isinstance(value, str):
raise SQLInjectionError(f"Value must be a string, got {type(value).__name__}")
if len(value) > max_length:
raise SQLInjectionError(
f"Input exceeds maximum length of {max_length} characters "
f"(got {len(value)})"
)
# Check for null bytes which can truncate strings in some contexts
if "\x00" in value:
raise SQLInjectionError("Input contains null bytes which are not allowed")
return value
class SecureQueryBuilder:
"""Builder for constructing safe, parameterized SQL queries.
This class provides a fluent interface for building SQL queries with
proper parameterization to prevent SQL injection. All values are
automatically parameterized, and identifiers are validated.
TNFR Context
------------
Use this builder when persisting TNFR structural data:
- NFR node states (EPI, νf, phase)
- Network topology and couplings
- Coherence measurements C(t)
- Sense index Si values
- Operator application history
Example
-------
>>> builder = SecureQueryBuilder()
>>> query, params = builder.select("nfr_nodes", ["id", "epi"])\\
... .where("nu_f > ?", 0.5)\\
... .where("phase BETWEEN ? AND ?", 0.0, 3.14)\\
... .order_by("nu_f", "DESC")\\
... .limit(10)\\
... .build()
>>> print(query) # doctest: +NORMALIZE_WHITESPACE
SELECT id, epi FROM nfr_nodes WHERE nu_f > ? AND phase BETWEEN ? AND ?
ORDER BY nu_f DESC LIMIT 10
>>> print(params)
[0.5, 0.0, 3.14]
"""
def __init__(self) -> None:
"""Initialize a new query builder."""
self._query_parts: list[str] = []
self._params: list[Any] = []
self._operation: str | None = None
def select(
self, table: str, columns: list[str] | None = None
) -> SecureQueryBuilder:
"""Start a SELECT query.
Parameters
----------
table : str
Table name (will be validated)
columns : list[str], optional
list of column names to select (default: all columns with *)
Returns
-------
SecureQueryBuilder
Self for method chaining
"""
self._operation = "SELECT"
table = validate_identifier(table)
if columns is None:
cols = "*"
else:
validated_cols = [validate_identifier(col) for col in columns]
cols = ", ".join(validated_cols)
self._query_parts.append(f"SELECT {cols} FROM {table}")
return self
def insert(self, table: str, columns: list[str]) -> SecureQueryBuilder:
"""Start an INSERT query.
Parameters
----------
table : str
Table name (will be validated)
columns : list[str]
list of column names
Returns
-------
SecureQueryBuilder
Self for method chaining
"""
self._operation = "INSERT"
table = validate_identifier(table)
validated_cols = [validate_identifier(col) for col in columns]
cols_str = ", ".join(validated_cols)
placeholders = ", ".join(["?"] * len(columns))
self._query_parts.append(
f"INSERT INTO {table} ({cols_str}) VALUES ({placeholders})"
)
return self
def update(self, table: str) -> SecureQueryBuilder:
"""Start an UPDATE query.
Parameters
----------
table : str
Table name (will be validated)
Returns
-------
SecureQueryBuilder
Self for method chaining
"""
self._operation = "UPDATE"
table = validate_identifier(table)
self._query_parts.append(f"UPDATE {table}")
return self
def set(self, **columns: Any) -> SecureQueryBuilder:
"""Add SET clause for UPDATE query.
Parameters
----------
**columns
Column-value pairs to update
Returns
-------
SecureQueryBuilder
Self for method chaining
"""
if self._operation != "UPDATE":
raise SQLInjectionError("SET clause can only be used with UPDATE")
validated_cols = [validate_identifier(col) for col in columns.keys()]
set_parts = [f"{col} = ?" for col in validated_cols]
self._params.extend(columns.values())
self._query_parts.append("SET " + ", ".join(set_parts))
return self
def delete(self, table: str) -> SecureQueryBuilder:
"""Start a DELETE query.
Parameters
----------
table : str
Table name (will be validated)
Returns
-------
SecureQueryBuilder
Self for method chaining
"""
self._operation = "DELETE"
table = validate_identifier(table)
self._query_parts.append(f"DELETE FROM {table}")
return self
def where(self, condition: str, *values: Any) -> SecureQueryBuilder:
"""Add a WHERE clause with parameterized values.
The condition should use '?' placeholders for values.
Parameters
----------
condition : str
WHERE condition with '?' placeholders
*values
Values to bind to placeholders
Returns
-------
SecureQueryBuilder
Self for method chaining
"""
# Validate that condition doesn't contain suspicious patterns
if ";" in condition or "--" in condition or "/*" in condition:
raise SQLInjectionError("WHERE condition contains suspicious SQL patterns")
# Count expected placeholders
expected_params = condition.count("?")
if len(values) != expected_params:
raise SQLInjectionError(
f"WHERE condition expects {expected_params} parameters, "
f"got {len(values)}"
)
prefix = (
"WHERE" if not any("WHERE" in part for part in self._query_parts) else "AND"
)
self._query_parts.append(f"{prefix} {condition}")
self._params.extend(values)
return self
def order_by(self, column: str, direction: str = "ASC") -> SecureQueryBuilder:
"""Add an ORDER BY clause.
Parameters
----------
column : str
Column name to order by (will be validated)
direction : str, optional
Sort direction: 'ASC' or 'DESC' (default: 'ASC')
Returns
-------
SecureQueryBuilder
Self for method chaining
"""
column = validate_identifier(column)
direction = direction.upper()
if direction not in ("ASC", "DESC"):
raise SQLInjectionError(f"Invalid sort direction: {direction}")
self._query_parts.append(f"ORDER BY {column} {direction}")
return self
def limit(self, count: int) -> SecureQueryBuilder:
"""Add a LIMIT clause.
Parameters
----------
count : int
Maximum number of rows to return
Returns
-------
SecureQueryBuilder
Self for method chaining
"""
if not isinstance(count, int) or count < 0:
raise SQLInjectionError("LIMIT count must be a non-negative integer")
self._query_parts.append(f"LIMIT {count}")
return self
def build(self) -> tuple[str, list[Any]]:
"""Build the final query and parameter list.
Returns
-------
tuple[str, list[Any]]
A tuple of (query_string, parameters)
"""
if not self._query_parts:
raise SQLInjectionError("Cannot build empty query")
query = " ".join(self._query_parts)
return query, self._params
def execute_parameterized_query(
query: str, params: list[Any] | tuple[Any, ...] | None = None
) -> None:
"""Execute a parameterized query safely.
This is a placeholder function that demonstrates the pattern for
executing parameterized queries. In actual usage, this would be
implemented with a real database connection.
Parameters
----------
query : str
The SQL query with '?' placeholders
params : list or tuple, optional
Parameters to bind to the query
Raises
------
SQLInjectionError
If the query appears to contain unparameterized values
Warning
-------
This is a demonstration function. Real implementations should:
1. Use database-specific parameterization (e.g., sqlite3.execute)
2. Handle connection pooling and transactions
3. Implement proper error handling and logging
4. Use context managers for resource cleanup
Example
-------
>>> # Safe parameterized query
>>> execute_parameterized_query(
... "SELECT * FROM nfr_nodes WHERE nu_f > ?",
... [0.5]
... ) # doctest: +SKIP
>>> # UNSAFE: Never do this!
>>> # execute_parameterized_query(
>>> # f"SELECT * FROM nfr_nodes WHERE id = {user_input}"
>>> # )
"""
if params is None:
params = []
# Basic validation: check for suspicious patterns
if "'" in query or '"' in query:
# Allow quoted identifiers if query has standard SQL keywords
if not any(
keyword in query.upper()
for keyword in ["SELECT", "INSERT", "UPDATE", "DELETE"]
):
raise SQLInjectionError(
"Query contains quoted strings. Use parameterized queries instead."
)
# Verify that the number of placeholders matches the number of parameters
placeholder_count = query.count("?")
if placeholder_count != len(params):
raise SQLInjectionError(
f"Query has {placeholder_count} placeholders but {len(params)} "
f"parameters provided"
)
# In a real implementation, this would execute the query
# For example, with sqlite3:
# cursor.execute(query, params)
__all__ = (
"SQLInjectionError",
"SecureQueryBuilder",
"execute_parameterized_query",
"sanitize_string_input",
"validate_identifier",
)