viz

visualise/flatten.py

@@ -0,0 +1,37 @@
+'''
+- parse the AST for every line of code
+- add this to a graph of all possibly trees with frequency counts
+- plot this tree
+'''
+import json
+
+
+with open('visualise/trie.json', 'r') as f:
+    trie = json.load(f)
+
+
+def flatten_children(root):
+    children = []
+    for name, vals in root.items():
+        if name[0] == ' ':
+            continue
+        rr = {'name': name, 'value': vals[' count']}
+        rr['children'] = flatten_children(vals)
+        if not rr['children']:
+            del rr['children']
+        children.append(rr)
+
+    tot = sum(child['value'] for child in children)
+    for child in children:
+        child['value'] = child['value'] / tot
+
+    return children
+
+
+trie = {
+    "name": "Module",
+    "children": flatten_children(trie)
+}
+
+with open('visualise/trie.flat.json', 'w') as f:
+    f.write(json.dumps(trie))

visualise/main.py

@@ -0,0 +1,37 @@
+'''
+- parse the AST for every line of code
+- add this to a graph of all possibly trees with frequency counts
+- plot this tree
+'''
+import ast
+import json
+from tqdm import tqdm
+
+
+trie = {}
+
+
+class Reader:
+    def __init__(self, trie, code):
+        self.trie = trie
+        self.root = ast.parse(code)
+
+    def run(self):
+        return self.add_code(self.trie, self.root)
+
+    def add_code(self, cursor, node):
+        for child in ast.iter_child_nodes(node):
+            name = type(child).__name__
+            cursor[name] = cursor.get(name, {' count': 0})
+            cursor[name][' count'] += 1            
+            cursor[name] = self.add_code(cursor[name], child)
+        return cursor
+
+
+with open('data.jsonl', 'r', encoding="utf-8") as f:
+    for id_, line in tqdm(enumerate(f), total=8968897):
+        trie = Reader(trie, json.loads(line)['code']).run()
+
+
+with open('visualise/trie.json', 'w') as f:
+    f.write(json.dumps(trie))