git ssb

1
#!/usr/bin/env python3
2
3
from pprint import pprint
4
import re
5
import json
6
import subprocess as sp
7
from collections import defaultdict
8
from graphviz import Digraph
9
import pickle
10
11
NEEDS_CALL = "rg --json -t js --multiline exports.needs[^\)]*\)"
12
GIVES_CALL = "rg --json -t js --multiline exports.gives[^\)]*\)"
13
14
PARENTHESIS_EXTRACTION_RE = re.compile(r'\(([^)]+)')
15
16
def parse_ripgrep(lines):
17
  for line in lines:
18
    if not line.strip():
19
      continue
20
    line = json.loads(line)
21
    if line['type'] != 'match':
22
      continue
23
    file_name = line['data']['path']['text']
24
    matched_text = line['data']['lines']['text']
25
    assert len(line['data']['submatches']) == 1, json.dumps(line, indent=2)
26
    yield file_name, matched_text
27
28
29
JS_TEMPLATE = '''
30
const nest = require('depnest')
31
32
console.log(JSON.stringify(nest({obj_str})))
33
'''
34
35
def flatten(d):
36
  for k, v in d.items():
37
    if type(v) == bool:
38
      assert v, d
39
      yield k
40
      continue
41
    if type(v) == str:
42
      # print(v)
43
      yield k
44
      continue
45
    v_postfixes = flatten(v)
46
    for postfix in v_postfixes:
47
      yield f'{k}.{postfix}'
48
49
def edges(rg_matches):
50
  rg_matches = list(parse_ripgrep(rg_matches.split('\n')))
51
  # print(len(rg_matches))
52
  result = dict()
53
  for fn, text in rg_matches:
54
    # print('='*80)
55
    # print(f'{fn}\n{text}')
56
    # print()
57
    if 'nest(' in text:
58
      obj_str = text[text.find('(')+1:text.find(')')]
59
    else:
60
      obj_str = text[text.find('{'):text.rfind('}')+1]
61
    obj_str = sp.check_output(['node', '-e', JS_TEMPLATE.format(obj_str=obj_str)])
62
    obj_str = obj_str.decode()
63
    obj = json.loads(obj_str)
64
    # print(json.dumps(obj, indent=2))
65
    # print()
66
    # print('\n'.join(list(flatten(obj))))
67
    result[fn] = list(flatten(obj))
68
  return result
69
70
if __name__ == '__main__':
71
  gives_rg = sp.check_output(GIVES_CALL.split()).decode()
72
  gives = edges(gives_rg)
73
74
  needs_rg = sp.check_output(NEEDS_CALL.split()).decode()
75
  needs = edges(needs_rg)
76
  
77
  # for every method: which files provide this method?
78
  providers = defaultdict(list)
79
  for filename, methods in gives.items():
80
    for method in methods:
81
      providers[method].append(filename)
82
  providers = dict(providers)
83
84
  pprint(providers)
85
86
  # for every file: which other files does this depend on, and for which method?
87
  dependencies = defaultdict(list)
88
  # for every file: which other files depend on this, and for which method?
89
  dependents = defaultdict(list)
90
  for filename, methods in needs.items():
91
    for method in methods:
92
      method_providers = providers.get(method, ['UNKNOWN'])
93
      dependencies[filename] += [(mp, method) for mp in method_providers]
94
      for mp in method_providers:
95
        dependents[mp].append((filename, method))
96
  dependencies = dict(dependencies)
97
98
  # for p in providers.keys():
99
  #   _ = dependents[p]
100
  dependents = dict(dependents)
101
  print('='*80)
102
  pprint(dependencies)
103
  print('='*80)
104
  pprint(dependents)
105
  print('='*80)
106
107
  sortedProviders = list(sorted(dependents.keys(), key=lambda k: len(dependents[k])))
108
  counts = {k: len(dependents[k]) for k in sortedProviders}
109
110
  max_count = len(str(max(*counts.values())))
111
  max_len = len(max(*counts.keys(), key=len))
112
  print(max_count, max_len)
113
114
  print('-'*80)
115
  for k in sortedProviders:
116
    print(f'{counts[k]:{max_count}}\t{k:{max_len}}')
117
118
  with open('results.pickle', 'wb') as f:
119
    pickle.dump({
120
      'gives': gives,
121
      'needs': needs,
122
      'counts': counts,
123
      'providers': providers,
124
      'dependents': dependents,
125
      'dependencies': dependencies,
126
    }, f)
127
128
  # dot = Digraph(comment='Dependencies')
129
  #
130
  # for filename in dependencies.keys():
131
  #   dot.node(filename)
132
  #
133
  # for filename, deps in dependencies.items():
134
  #   for dep, api in deps:
135
  #     dot.edge(filename, dep, label=api)
136
  #
137
  # print(dot.source)
138
  #
139
  # with open('dependencies.dot', 'w') as f:
140
  #   f.write(dot.source)
141
  #
142
143