lizard 1.17.31__py2.py3-none-any.whl → 1.19.0__py2.py3-none-any.whl

lizard_languages/plsql.py

@@ -0,0 +1,419 @@
+"""
+Language parser for PL/SQL (Oracle's Procedural Language extension to SQL)
+
+This module implements complexity analysis for PL/SQL code, supporting:
+- Procedures, Functions, and Triggers
+- Package Bodies (not specifications - they only contain signatures)
+- Nested procedures and functions
+- Anonymous blocks with nested functions (blocks themselves aren't counted)
+- Control structures: IF/ELSIF/ELSE, CASE/WHEN, LOOP/WHILE/FOR
+- Exception handlers
+- Cursor declarations and cursor FOR loops
+
+Design Decisions:
+- EXIT WHEN: The WHEN keyword is filtered out by the preprocessor because
+  "EXIT WHEN condition" is not a branching construct - it's a conditional
+  exit that doesn't create alternate execution paths. The LOOP itself adds
+  complexity.
+
+- CONTINUE WHEN: Similar to EXIT WHEN, the WHEN is counted as it does create
+  a branch in the loop execution.
+
+- GOTO: Does not add to cyclomatic complexity as it's just an unconditional
+  jump, not a decision point.
+
+- Standalone LOOP: Adds +1 complexity as it creates a repeating path.
+
+- FOR/WHILE LOOP: The FOR/WHILE keyword adds complexity; the following LOOP
+  keyword is part of the same construct and doesn't add additional complexity.
+
+- Parameter Counting: Currently counts all non-whitespace tokens and commas
+  in parameter lists. This approach works but differs from some other language
+  implementations.
+"""
+
+from .code_reader import CodeReader, CodeStateMachine
+from .clike import CCppCommentsMixin
+
+
+class PLSQLReader(CodeReader, CCppCommentsMixin):
+    """
+    Reader for PL/SQL language supporting procedures, functions, packages,
+    and core control structures.
+    """
+
+    ext = ["sql", "pks", "pkb", "pls", "plb", "pck"]
+    language_names = ["plsql", "pl/sql"]
+
+    # PL/SQL conditions for cyclomatic complexity
+    # Note: 'loop' is NOT in this set because LOOP has special handling:
+    # - standalone LOOP adds +1
+    # - LOOP after WHILE/FOR should not add (it's part of the compound statement)
+    _conditions = {"if", "elsif", "when", "while", "for", "and", "or"}
+
+    def __init__(self, context):
+        super(PLSQLReader, self).__init__(context)
+        self.parallel_states = [PLSQLStates(context)]
+        # PL/SQL is case-insensitive, so add both lowercase and uppercase versions
+        # of keywords to the conditions set for the automatic condition counter
+        self.conditions = self.conditions | {c.upper() for c in self.conditions}
+
+    def preprocess(self, tokens):
+        """
+        Preprocess tokens to handle PL/SQL-specific constructs.
+        Merge compound keywords to prevent the condition counter from double-counting:
+        - "END IF", "END LOOP", "END CASE", "END WHILE", "END FOR" -> single tokens
+        - "EXIT WHEN" -> remove the WHEN keyword (EXIT doesn't create a branch)
+        """
+        last_nonwhitespace_token = None
+        pending_tokens = []
+
+        for token in tokens:
+            if not token.isspace() or token == "\n":
+                token_upper = token.upper()
+
+                # Handle "END IF", "END LOOP", etc.
+                if (
+                    last_nonwhitespace_token
+                    and last_nonwhitespace_token.upper() == "END"
+                ):
+                    if token_upper in ("IF", "LOOP", "CASE", "WHILE", "FOR"):
+                        # Merge into "END_IF", "END_LOOP", etc.
+                        yield "END_" + token_upper
+                        last_nonwhitespace_token = None
+                        pending_tokens = []
+                        continue
+
+                # Handle "EXIT WHEN" - skip the WHEN keyword
+                if (
+                    last_nonwhitespace_token
+                    and last_nonwhitespace_token.upper() == "EXIT"
+                    and token_upper == "WHEN"
+                ):
+                    # Skip this WHEN keyword
+                    pending_tokens = []
+                    continue
+
+                # Yield any pending tokens
+                if last_nonwhitespace_token:
+                    yield last_nonwhitespace_token
+                for pending in pending_tokens:
+                    yield pending
+                pending_tokens = []
+
+                # Update tracking
+                last_nonwhitespace_token = token
+            else:
+                # Accumulate whitespace
+                pending_tokens.append(token)
+
+        # Don't forget the last tokens
+        if last_nonwhitespace_token:
+            yield last_nonwhitespace_token
+        for pending in pending_tokens:
+            yield pending
+
+    @staticmethod
+    def generate_tokens(source_code, addition="", token_class=None):
+        """
+        Generate tokens for PL/SQL code.
+        PL/SQL uses:
+        - Single-line comments: --
+        - Multi-line comments: /* */
+        - String literals: 'text' (with '' for escaping)
+        - Assignment operator: :=
+        """
+        # Add PL/SQL-specific patterns
+        addition = r"|--[^\n]*" + addition  # Single-line comment starting with --
+        return CodeReader.generate_tokens(source_code, addition, token_class)
+
+    def get_comment_from_token(self, token):
+        """
+        Override to recognize PL/SQL's -- line comments in addition to /* */ block comments.
+        PL/SQL uses -- for single-line comments (like SQL standard).
+
+        Note: This method correctly identifies -- comments, but due to a limitation in
+        the NLOC calculation, these comments may still be counted in NLOC.
+        """
+        if token.startswith("--"):
+            return token  # Return full comment token (like Lua does)
+        # Delegate to parent for /* */ and // comments
+        return super().get_comment_from_token(token)
+
+
+class PLSQLStates(CodeStateMachine):
+    """
+    State machine for parsing PL/SQL code structure.
+    """
+
+    def __init__(self, context):
+        super(PLSQLStates, self).__init__(context)
+        self.in_parameter_list = False
+        self.last_control_keyword = None  # Track FOR/WHILE to avoid counting their LOOP
+        self.declaring_nested_function = (
+            False  # Track if we're declaring a nested function
+        )
+
+    def _state_global(self, token):
+        """Global state - looking for function/procedure/trigger declarations."""
+        token_lower = token.lower()
+
+        if token_lower == "procedure":
+            self.next(self._procedure_name)
+        elif token_lower == "function":
+            self.next(self._function_name)
+        elif token_lower == "trigger":
+            self.next(self._trigger_name)
+
+    def _procedure_name(self, token):
+        """Read procedure name."""
+        if token.isspace() or token == "\n":
+            return
+        if token == "(":
+            self.in_parameter_list = True
+            self.next(self._parameters, "(")
+        elif token.lower() in ("is", "as"):
+            self.context.confirm_new_function()
+            self.next(self._state_before_begin)
+        else:
+            # Check if this is a nested function
+            if self.declaring_nested_function:
+                self.context.push_new_function(token)
+                self.declaring_nested_function = False
+            else:
+                self.context.try_new_function(token)
+            self.next(self._procedure_after_name)
+
+    def _procedure_after_name(self, token):
+        """After procedure name, look for parameters or IS/AS."""
+        if token == ".":
+            # Schema-qualified name: the previous token was the schema,
+            # next non-whitespace token will be the actual procedure name
+            self.next(self._procedure_name_after_dot)
+        elif token == "(":
+            self.in_parameter_list = True
+            self.next(self._parameters, "(")
+        elif token.lower() in ("is", "as"):
+            self.context.confirm_new_function()
+            self.next(self._state_before_begin)
+        # Skip whitespace and other tokens
+
+    def _procedure_name_after_dot(self, token):
+        """Read the actual procedure name after schema.dot."""
+        if token.isspace() or token == "\n":
+            return
+        # Replace the previous (schema) name with the actual procedure name
+        self.context.current_function.name = token
+        self.next(self._procedure_after_name)
+
+    def _function_name(self, token):
+        """Read function name."""
+        if token.isspace() or token == "\n":
+            return
+        if token == "(":
+            self.in_parameter_list = True
+            self.next(self._parameters, "(")
+        elif token.lower() == "return":
+            self.next(self._return_type)
+        elif token.lower() in ("is", "as"):
+            self.context.confirm_new_function()
+            self.next(self._state_before_begin)
+        else:
+            # Check if this is a nested function
+            if self.declaring_nested_function:
+                self.context.push_new_function(token)
+                self.declaring_nested_function = False
+            else:
+                self.context.try_new_function(token)
+            self.next(self._function_after_name)
+
+    def _function_after_name(self, token):
+        """After function name, look for parameters, RETURN, or IS/AS."""
+        if token == ".":
+            # Schema-qualified name: the previous token was the schema,
+            # next non-whitespace token will be the actual function name
+            self.next(self._function_name_after_dot)
+        elif token == "(":
+            self.in_parameter_list = True
+            self.next(self._parameters, "(")
+        elif token.lower() == "return":
+            self.next(self._return_type)
+        elif token.lower() in ("is", "as"):
+            self.context.confirm_new_function()
+            self.next(self._state_before_begin)
+        # Skip whitespace and other tokens
+
+    def _function_name_after_dot(self, token):
+        """Read the actual function name after schema.dot."""
+        if token.isspace() or token == "\n":
+            return
+        # Replace the previous (schema) name with the actual function name
+        self.context.current_function.name = token
+        self.next(self._function_after_name)
+
+    def _return_type(self, token):
+        """Skip return type declaration."""
+        if token.lower() in ("is", "as"):
+            self.context.confirm_new_function()
+            self.next(self._state_before_begin)
+        # Skip everything else (return type tokens)
+
+    def _parameters(self, token):
+        """Read parameters."""
+        if token == ")":
+            self.in_parameter_list = False
+            self.next(self._after_parameters)
+        elif token == ",":
+            # Each comma separates parameters
+            self.context.parameter(token)
+        elif not token.isspace() and token != "\n":
+            # Track non-whitespace tokens as potential parameters
+            self.context.parameter(token)
+
+    def _after_parameters(self, token):
+        """After parameters, look for IS/AS or RETURN."""
+        if token.lower() == "return":
+            self.next(self._return_type)
+        elif token.lower() in ("is", "as"):
+            self.context.confirm_new_function()
+            self.next(self._state_before_begin)
+        # Skip whitespace and other tokens
+
+    def _trigger_name(self, token):
+        """Read trigger name."""
+        if token.isspace() or token == "\n":
+            return
+        # Trigger name found
+        self.context.try_new_function(token)
+        self.seen_trigger_name_token = False  # Track if we've seen non-whitespace after name
+        self.next(self._trigger_after_name)
+
+    def _trigger_after_name(self, token):
+        """After trigger name, skip until DECLARE or BEGIN."""
+        token_lower = token.lower()
+
+        # Only check for dot immediately after trigger name (before any other tokens)
+        if token == "." and not self.seen_trigger_name_token:
+            # Schema-qualified name: the previous token was the schema,
+            # next non-whitespace token will be the actual trigger name
+            self.next(self._trigger_name_after_dot)
+            return
+
+        # Mark that we've seen a non-whitespace token after the trigger name
+        if not token.isspace() and token != "\n":
+            self.seen_trigger_name_token = True
+
+        if token_lower == "declare":
+            self.context.confirm_new_function()
+            self.next(self._state_before_begin)
+        elif token_lower == "begin":
+            self.context.confirm_new_function()
+            self.br_count = 1
+            self.next(self._state_body)
+        # Skip everything else (BEFORE/AFTER, INSERT/UPDATE/DELETE, ON table_name, FOR EACH ROW, etc.)
+
+    def _trigger_name_after_dot(self, token):
+        """Read the actual trigger name after schema.dot."""
+        if token.isspace() or token == "\n":
+            return
+        # Replace the previous (schema) name with the actual trigger name
+        self.context.current_function.name = token
+        self.seen_trigger_name_token = False  # Reset for the real trigger name
+        self.next(self._trigger_after_name)
+
+    def _state_before_begin(self, token):
+        """
+        State between IS/AS and BEGIN - this is the declaration section.
+        Watch for nested procedures/functions and the BEGIN keyword.
+        """
+        token_lower = token.lower()
+
+        # Check for nested procedure/function declarations
+        if token_lower == "procedure":
+            self.declaring_nested_function = True
+            # Store current br_count level to know when nested function ends
+            if not hasattr(self, "nested_br_level"):
+                self.nested_br_level = 0
+            self.next(self._procedure_name)
+            return
+        elif token_lower == "function":
+            self.declaring_nested_function = True
+            # Store current br_count level to know when nested function ends
+            if not hasattr(self, "nested_br_level"):
+                self.nested_br_level = 0
+            self.next(self._function_name)
+            return
+        elif token_lower == "begin":
+            # Start of the implementation body
+            # Check if we had nested functions and need to reset br_count tracking
+            if hasattr(self, "nested_br_level"):
+                self.br_count = self.nested_br_level + 1
+                delattr(self, "nested_br_level")
+            else:
+                self.br_count = 1  # Initialize counter for the first BEGIN
+            self.next(self._state_body)
+
+    def _state_body(self, token):
+        """
+        Process function/procedure body.
+        Track control structures for cyclomatic complexity.
+        Manually track BEGIN/END blocks.
+        """
+        token_lower = token.lower()
+        token_upper = token.upper()
+
+        # Check for merged compound keywords like "END_IF", "END_LOOP", etc.
+        # These are created by the preprocessor
+        if token_lower.startswith("end_"):
+            # This is a compound END keyword, reset tracking
+            self.last_control_keyword = None
+            return
+
+        # Handle nested procedure/function declarations
+        if token_lower == "procedure":
+            self.next(self._procedure_name)
+            return
+        elif token_lower == "function":
+            self.next(self._function_name)
+            return
+
+        # Track FOR and WHILE to know when LOOP follows them
+        if token_upper in ("FOR", "WHILE"):
+            self.last_control_keyword = token_upper
+
+        # Handle LOOP keyword manually
+        # - Standalone LOOP adds +1 complexity
+        # - LOOP after FOR/WHILE does not add complexity (already counted for FOR/WHILE)
+        elif token_upper == "LOOP":
+            if self.last_control_keyword not in ("FOR", "WHILE"):
+                # This is a standalone LOOP, add complexity
+                self.context.add_condition()
+            # Reset tracking after processing LOOP
+            self.last_control_keyword = None
+
+        # PL/SQL uses BEGIN/END instead of {}
+        if token_lower == "begin":
+            self.br_count += 1
+            self.context.add_bare_nesting()
+        elif token_lower == "end":
+            # This is a standalone END (for BEGIN/END block)
+            self.br_count -= 1
+            if self.br_count == 0:
+                # This END closes the function/procedure
+                # Check if we have a parent function BEFORE ending (stack gets popped)
+                has_parent = len(self.context.stacked_functions) > 0
+                self.context.end_of_function()
+                # Return to appropriate state based on whether this was nested
+                if has_parent:
+                    # Return to parent function's declaration section
+                    self.next(self._state_before_begin)
+                else:
+                    # No parent function, return to global
+                    self.next(self._state_global)
+                return
+            else:
+                self.context.pop_nesting()
+
+        # Note: Basic conditions (if, elsif, when, while, for, and, or)
+        # are automatically counted by the condition_counter processor
+        # based on the _conditions set in the Reader class.

lizard_languages/python.py

@@ -37,6 +37,7 @@ class PythonReader(CodeReader, ScriptLanguageMixIn):
     def __init__(self, context):
         super(PythonReader, self).__init__(context)
         self.parallel_states = [PythonStates(context, self)]
+        self._last_meaningful_token = None  # Track the last meaningful token
 
     @staticmethod
     def generate_tokens(source_code, addition='', token_class=None):
@@ -46,6 +47,41 @@ class PythonReader(CodeReader, ScriptLanguageMixIn):
             r"|(?:\'\'\'(?:\\.|[^\']|\'(?!\'\')|\'\'(?!\'))*\'\'\')",
             token_class)
 
+    def process_token(self, token):
+        """Process triple-quoted strings used as comments.
+
+        Triple-quoted strings that are not docstrings (i.e., not immediately
+        after function definitions) should be treated like comments and not
+        counted in NLOC, but only if they appear to be standalone statements
+        rather than part of assignments or other expressions.
+
+        Returns:
+            bool: True if the token was handled specially, False otherwise
+        """
+        if (token.startswith('"""') or token.startswith("'''")) and len(token) >= 6:
+            # Check if this is likely a standalone comment (not a docstring)
+            # Docstrings are handled separately in _state_first_line
+            current_state = self.parallel_states[0]._state
+
+            # If we're not in the first line state, check if this is a standalone string
+            if current_state != current_state.__self__._state_first_line:
+                # Check if the immediate previous meaningful token suggests this is part of an expression
+                assignment_tokens = ['=', '+=', '-=', '*=', '/=', '%=', '//=', '**=', '&=', '|=', '^=',
+                                     '<<=', '>>=', '(', 'return', ',', '[', '+', '-', '*', '/', '%']
+
+                is_part_of_expression = self._last_meaningful_token in assignment_tokens
+
+                # Only treat as comment if it's NOT part of an expression
+                if not is_part_of_expression:
+                    # Subtract the NLOC contribution of this triple-quoted string
+                    self.context.add_nloc(-(token.count('\n') + 1))
+
+        # Update last meaningful token (ignore whitespace and newlines)
+        if token not in ['\n', ' ', '\t'] and not token.isspace():
+            self._last_meaningful_token = token
+
+        return False  # Continue with normal processing
+
     def preprocess(self, tokens):
         indents = PythonIndents(self.context)
         current_leading_spaces = 0