pythonでTFIDFの処理を並列処理するには

pythonについてです

タイトル通り、TFIDFの処理を並列処理したいと考えています
普通にTFIDFをするぶんにはできるのですが、TFIDFしたいと考えているファイルが大きくて
並列化しないととてもじゃないけど終わらないということでやっています

コードを書いているところで、エラーが出たので投稿しました

現在以下のコードを書いていて
エラーがでます

python
1import nltk
2import numpy as np
3import json
4#import nltk_exa as nl
5import time
6import multiprocessing as mp
7
8def subcalc(p): #p = 0~8
9    word = []
10    proc = 8
11    f = open("word0_a.txt")
12    line = f.readline()
13    while line:
14        line = f.readline()
15        word.append(line.replace("[","").replace("]","").replace(",","").replace("\"","").split())
16    f.close()
17    word.pop()
18    print(word)
19
20    ini = len(word) * p / proc
21    fin = len(word) * (p+1) / proc
22
23    subdoc = []
24    lists = []
25    proc = 8
26    collection = nltk.TextCollection(word) 
27    t1 = time.time()
28
29    for do in range(ini, fin):
30        wo=[]
31        for term in set(do):
32            if(collection.tf_idf(term, do) > 0):
33                wo.append([term,collection.tf_idf(term, do)]) #ここも上のサイトにのってる
34                #print(wo)
35        wo.sort(key=lambda x:x[1]) 
36        wo.reverse()
37        slice1 = np.array(wo[:20]) 
38        lists = slice1[:,0] 
39        subdoc.append(list(lists)) 
40        #print(doc[0])
41        t2 = time.time()
42        print('processing time1(一回のfor文): ' + str(t2 - t1) + '(sec)')
43        del wo
44
45    return subdoc
46
47def tfidf():
48   
49    proc = 8
50    pool = mp.Pool(proc)
51    doc = pool.map(subcalc, range(8))
52    t3 = time.time()
53    print(doc)
54    print('processing time2(終わり): ' + str(t3 - t1) + '(sec)')
55
56if __name__ == "__main__":
57    tfidf()
58
59-------------------------------------------------------------
60実行結果
61
62doc = pool.map(subcalc, range(8))
63TypeError: 'float' object cannot be interpreted as an integer
64

word0_a.txt(一部抜粋)

["portions", "Sandwich", "top", "Rick", "come", "parents", "ONLY", "Brother", "advice", "My", "lunch", "text", "Sundays", "fish", "people", "friendly", "bit", "Fish", "giant", "Mom", "drinks", "Dad", "Sebak", "Pittsburgh", "CASH", "Open-Faced", "late", "afternoon", "staff", "ATM", "accurate", "several", "dishes", "recommendations", "Reuben", "early", "massive", "Mondays", "times", "affordable", "sandwich", "place", "Saturday", "prepared", "DELICIOUS", "Things", "today", "large", "time", "Steak", "sure", "recommendation", "good", "meal", "pretty", "Oh"]
["busy", "table", "portions", "taste", "ages", "service", "happy", "Rib", "Sandwiich", "years", "more", "while", "menu", "text", "people", "friendly", "fresh", "-Flounder", "little", "huge", "value", "lot", "Special", "delicious", "everything", "style", "nothing", "Pittsburgh", "girls", "movie", "reviews", "St", "few", "ways", "-Fish", "ambiance", "glad", "times", "slow", "Worth", "food", "Norfolk", "place", "miss", "prepared", "prices", "complaints", "-Salmon", "kitchen", "AWESOME", "hit", "Everything", "low", "large", "gem", "sure", "different", "-Prime", "good", "Most", "busier"]

どのように改善すればよろしいですか
よろしくお願いします

ちなみに並列化しないバージョンはこれです

python
1import nltk
2import numpy as np
3import json
4import nltk_exa as nl
5import time
6
7#def tfidf(word):
8def tfidf():
9   
10    word = []
11    f = open("word0_a.txt")
12    line = f.readline()
13    while line:
14        line = f.readline()
15        word.append(line.replace("[","").replace("]","").replace(",","").replace("\"","").split())
16    f.close()
17    word.pop()
18    #tfidfの計算
19    doc = []
20    lists = []
21    collection = nltk.TextCollection(word) #サイトにのっていた
22    t1 = time.time()
23    for do in word:
24        wo=[]
25        for term in set(do):
26            if(collection.tf_idf(term, do) > 0):
27                wo.append([term,collection.tf_idf(term, do)]) #ここも上のサイトにのってる
28                #print(wo)
29        wo.sort(key=lambda x:x[1]) 
30        wo.reverse()
31        #print(wo)
32        slice1 = np.array(wo[:20]) #先頭の文字から終了インデックスまでが抽出
33        lists = slice1[:,0] #[:]は戦闘から終了のインデックスまで抽出と、slice1の0番目を格納
34        doc.append(list(lists)) #listsが文字列だから、リストに格納
35        #print(doc[0])
36        t2 = time.time()
37        #print('processing time1(一回のfor文): ' + str(t2 - t1) + '(sec)')
38        del wo
39
40    print(doc)
41    t3 = time.time()
42    print('processing time2(終わり): ' + str(t3 - t1) + '(sec)')
43   
44if __name__ == "__main__":
45        tfidf()
46