[pattern discovery - apriori] feedback'll be appreciated.

solution.py

import itertools

abs_min_support = 771
freq_items = [None, {}]

# part1
counting = {}
with open('categories.txt', 'r') as f:
  while True:
    line = f.readline()
    if not line:
      break

    for c in line.split(';'):
      c = c.strip()
      if c in counting:
        counting[c] += 1
      else:
        counting[c] = 1

with open('patterns.txt', 'w') as f:
  for k, v in counting.items():
    if v > abs_min_support:
      freq_items[1][k] = v
      f.write(f'{v}:{k}\n')

# part2
tdb = []
indexed_tdb = {}
max_count = 0
with open('categories.txt', 'r') as f:
  while True:
    line = f.readline()
    if not line:
      break
    items = list(map(lambda x: x.strip(), line.split(';')))
    max_count = max(max_count, len(items))
    tdb.append(items)
    for item in items:
      if item in indexed_tdb:
        indexed_tdb[item].append(len(tdb) - 1)
      else:
        indexed_tdb[item] = [len(tdb) - 1]

prev_freq = freq_items[1].keys()
for count in range(2, max_count + 1):
  counting = {}
  combs = list(itertools.combinations(prev_freq, count))
  for comb in combs:
    get_related_tdb = []
    flatten_comb = list(map(lambda x: x.split(';'), comb))
    flatten_comb = list(itertools.chain.from_iterable(flatten_comb))
    comb = set(flatten_comb)
    if len(comb) != count:
      continue

    for c in comb:
      get_related_tdb += indexed_tdb[c]
    for index in set(get_related_tdb):
      line = tdb[index]
      for c in comb:
        if c not in line:
          break
      else:
        comb = list(comb)
        comb.sort()
        key = ';'.join(comb)
        if key in counting:
          counting[key] += 1
        else:
          counting[key] = 1

  freq_items.append({})
  for k, v in counting.items():
    if v > abs_min_support:
      freq_items[count][k] = v
  prev_freq = freq_items[count].keys()

with open('patterns2.txt', 'w') as f:
  for freq in freq_items[1:]:
    for k, v in freq.items():
      f.write(f'{v}:{k}\n')