Spaces:
Running
Running
Update parser.py
Browse files
parser.py
CHANGED
@@ -13,13 +13,38 @@ def get_category(node):
|
|
13 |
return 'async_function'
|
14 |
elif isinstance(node, ast.ClassDef):
|
15 |
return 'class'
|
|
|
|
|
|
|
|
|
|
|
|
|
16 |
elif isinstance(node, ast.Expr):
|
17 |
return 'expression'
|
18 |
else:
|
19 |
return 'other'
|
20 |
|
21 |
-
def
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
22 |
"""Recursively parse an AST node and its children, assigning hierarchy levels."""
|
|
|
|
|
|
|
23 |
parts = []
|
24 |
start_line = getattr(node, 'lineno', prev_end + 1)
|
25 |
end_line = getattr(node, 'end_lineno', start_line)
|
@@ -27,56 +52,71 @@ def parse_node(node, lines, prev_end, level=0):
|
|
27 |
# Handle spacers before the node
|
28 |
if start_line > prev_end + 1:
|
29 |
spacer_lines = lines[prev_end:start_line - 1]
|
|
|
30 |
parts.append({
|
31 |
'category': 'spacer',
|
32 |
'source': ''.join(spacer_lines),
|
33 |
'location': (prev_end + 1, start_line - 1),
|
34 |
-
'level': level
|
|
|
35 |
})
|
36 |
|
37 |
# Capture the node's source
|
38 |
stmt_lines = lines[start_line - 1:end_line]
|
|
|
39 |
parts.append({
|
40 |
'category': get_category(node),
|
41 |
'source': ''.join(stmt_lines),
|
42 |
'location': (start_line, end_line),
|
43 |
-
'level': level
|
|
|
44 |
})
|
45 |
|
46 |
-
# Process nested nodes (e.g., class
|
47 |
if hasattr(node, 'body'):
|
48 |
nested_prev_end = end_line - 1
|
49 |
for child in node.body:
|
50 |
-
child_parts = parse_node(child, lines, nested_prev_end, level + 1)
|
51 |
parts.extend(child_parts)
|
52 |
nested_prev_end = child_parts[-1]['location'][1]
|
53 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
54 |
return parts
|
55 |
|
56 |
def parse_python_code(code):
|
57 |
-
"""Parse Python code string and return parts with hierarchy."""
|
58 |
lines = code.splitlines(keepends=True)
|
|
|
59 |
try:
|
60 |
tree = ast.parse(code)
|
61 |
except SyntaxError:
|
62 |
-
return [{'category': 'error', 'source': 'Invalid Python code', 'location': (1, 1), 'level': 0}]
|
63 |
|
64 |
parts = []
|
65 |
prev_end = 0
|
66 |
|
67 |
for stmt in tree.body:
|
68 |
-
stmt_parts = parse_node(stmt, lines, prev_end)
|
69 |
parts.extend(stmt_parts)
|
70 |
prev_end = stmt_parts[-1]['location'][1]
|
71 |
|
72 |
# Capture trailing spacers
|
73 |
-
if prev_end <
|
74 |
remaining_lines = lines[prev_end:]
|
|
|
75 |
parts.append({
|
76 |
'category': 'spacer',
|
77 |
'source': ''.join(remaining_lines),
|
78 |
-
'location': (prev_end + 1,
|
79 |
-
'level': 0
|
|
|
80 |
})
|
81 |
|
82 |
return parts
|
|
|
13 |
return 'async_function'
|
14 |
elif isinstance(node, ast.ClassDef):
|
15 |
return 'class'
|
16 |
+
elif isinstance(node, ast.If):
|
17 |
+
return 'if'
|
18 |
+
elif isinstance(node, ast.While):
|
19 |
+
return 'while'
|
20 |
+
elif isinstance(node, ast.For):
|
21 |
+
return 'for'
|
22 |
elif isinstance(node, ast.Expr):
|
23 |
return 'expression'
|
24 |
else:
|
25 |
return 'other'
|
26 |
|
27 |
+
def create_vector(category, level, location, total_lines):
|
28 |
+
"""Create a vector representation for a code part."""
|
29 |
+
# Vector: [category_id, level, start_line_normalized, end_line_normalized]
|
30 |
+
category_map = {
|
31 |
+
'import': 1, 'assignment': 2, 'function': 3, 'async_function': 4, 'class': 5,
|
32 |
+
'if': 6, 'while': 7, 'for': 8, 'expression': 9, 'spacer': 10, 'other': 11
|
33 |
+
}
|
34 |
+
category_id = category_map.get(category, 0)
|
35 |
+
start_line, end_line = location
|
36 |
+
return [
|
37 |
+
category_id,
|
38 |
+
level,
|
39 |
+
start_line / total_lines, # Normalized start position
|
40 |
+
end_line / total_lines # Normalized end position
|
41 |
+
]
|
42 |
+
|
43 |
+
def parse_node(node, lines, prev_end, level=0, total_lines=None):
|
44 |
"""Recursively parse an AST node and its children, assigning hierarchy levels."""
|
45 |
+
if total_lines is None:
|
46 |
+
total_lines = len(lines)
|
47 |
+
|
48 |
parts = []
|
49 |
start_line = getattr(node, 'lineno', prev_end + 1)
|
50 |
end_line = getattr(node, 'end_lineno', start_line)
|
|
|
52 |
# Handle spacers before the node
|
53 |
if start_line > prev_end + 1:
|
54 |
spacer_lines = lines[prev_end:start_line - 1]
|
55 |
+
spacer_vector = create_vector('spacer', level, (prev_end + 1, start_line - 1), total_lines)
|
56 |
parts.append({
|
57 |
'category': 'spacer',
|
58 |
'source': ''.join(spacer_lines),
|
59 |
'location': (prev_end + 1, start_line - 1),
|
60 |
+
'level': level,
|
61 |
+
'vector': spacer_vector
|
62 |
})
|
63 |
|
64 |
# Capture the node's source
|
65 |
stmt_lines = lines[start_line - 1:end_line]
|
66 |
+
node_vector = create_vector(get_category(node), level, (start_line, end_line), total_lines)
|
67 |
parts.append({
|
68 |
'category': get_category(node),
|
69 |
'source': ''.join(stmt_lines),
|
70 |
'location': (start_line, end_line),
|
71 |
+
'level': level,
|
72 |
+
'vector': node_vector
|
73 |
})
|
74 |
|
75 |
+
# Process nested nodes (e.g., class/function bodies, control structures)
|
76 |
if hasattr(node, 'body'):
|
77 |
nested_prev_end = end_line - 1
|
78 |
for child in node.body:
|
79 |
+
child_parts = parse_node(child, lines, nested_prev_end, level + 1, total_lines)
|
80 |
parts.extend(child_parts)
|
81 |
nested_prev_end = child_parts[-1]['location'][1]
|
82 |
|
83 |
+
# Handle additional bodies (e.g., else, elif, orelse for loops)
|
84 |
+
if hasattr(node, 'orelse') and node.orelse:
|
85 |
+
orelse_prev_end = parts[-1]['location'][1]
|
86 |
+
for child in node.orelse:
|
87 |
+
child_parts = parse_node(child, lines, orelse_prev_end, level + 1, total_lines)
|
88 |
+
parts.extend(child_parts)
|
89 |
+
orelse_prev_end = child_parts[-1]['location'][1]
|
90 |
+
|
91 |
return parts
|
92 |
|
93 |
def parse_python_code(code):
|
94 |
+
"""Parse Python code string and return parts with hierarchy and vectors."""
|
95 |
lines = code.splitlines(keepends=True)
|
96 |
+
total_lines = len(lines)
|
97 |
try:
|
98 |
tree = ast.parse(code)
|
99 |
except SyntaxError:
|
100 |
+
return [{'category': 'error', 'source': 'Invalid Python code', 'location': (1, 1), 'level': 0, 'vector': [0, 0, 1.0, 1.0]}]
|
101 |
|
102 |
parts = []
|
103 |
prev_end = 0
|
104 |
|
105 |
for stmt in tree.body:
|
106 |
+
stmt_parts = parse_node(stmt, lines, prev_end, total_lines=total_lines)
|
107 |
parts.extend(stmt_parts)
|
108 |
prev_end = stmt_parts[-1]['location'][1]
|
109 |
|
110 |
# Capture trailing spacers
|
111 |
+
if prev_end < total_lines:
|
112 |
remaining_lines = lines[prev_end:]
|
113 |
+
spacer_vector = create_vector('spacer', 0, (prev_end + 1, total_lines + 1), total_lines)
|
114 |
parts.append({
|
115 |
'category': 'spacer',
|
116 |
'source': ''.join(remaining_lines),
|
117 |
+
'location': (prev_end + 1, total_lines + 1),
|
118 |
+
'level': 0,
|
119 |
+
'vector': spacer_vector
|
120 |
})
|
121 |
|
122 |
return parts
|