Python：显示字典单词的匹配键

Question

我想在我的项目中显示字典单词的匹配键。我的代码当前输出的键，但对于您键入的任何单词，都使用相同的键。例如，如果我把返回的密钥将是.如果我把同样的钥匙将被归还。请参阅下面的代码，如果做错了什么，请告诉我'england played well'[737, 736, 735, 734, 733, 732, 731, 730, 729, 728]'Hello'import reimport osimport mathimport heapqdef readfile(path, docid):    files = sorted(os.listdir(path))    f = open(os.path.join(path, files[docid]), 'r',encoding='latin-1')    s = f.read()    f.close()    return sDELIM = '[ \n\t0123456789;:.,/\(\)\"\'-]+'def tokenize(text):    return re.split(DELIM, text.lower())N = len(sorted(os.listdir('docs')))def indextextfiles_RR(path):    postings={}    docLength = {}    term_in_document = {}    for docID in range(N):        s = readfile(path, docID)        words = tokenize(s)        length = 0        for w in words:            if w!='':                length += (math.log10(words.count(w)))**2        docLength[docID] = math.sqrt(length)        for w in words:            if w!='':                doc_length = math.log10(words.count(w))/docLength[docID]                term_in_document.setdefault(doc_length, set()).add(docID)                postings[w] = term_in_document    return postingsdef query_RR(postings, qtext):    words = tokenize(qtext)    doc_scores = {}    for docID in range(N):        score = 0        for w in words:            tf = words.count(w)            df = len(postings[w])            idf = math.log10(N / (df+1))            query_weights = tf * idf        for w in words:            if w in postings:                score = score + query_weights        doc_scores[docID] = score    res = heapq.nlargest(10, doc_scores)    return respostings = indextextfiles_RR('docs')print(query_RR(postings, 'hello'))当我运行帖子时，它应该返回hello和与之关联的键列表。

Answer 1

最有可能的是，您的错误来自您对每个文件中的所有单词使用相同的字典。term_in_document

几条评论

1. len(sorted(...))它浪费资源对不需要排序的东西（排序并不便宜），因为你只得到长度。

2. 按数字读取文件根本没有意义，要做到这一点，您最终会调用文件系统多个时间来读取整个目录的文件名，因为您每次读取一个目录时都会列出文件。

3. 文件应该在处理为我们关闭文件的语句中打开。with

4. 变量和函数应使用，而类应使用 。this_notationThisNotation

5. 您在单词列表上迭代两次只是为了获得十进制对数。

之后的逻辑非常令人困惑，您似乎正在对每个单词出现次数的十进制对数进行RMS（均方根），但您不会将其除以单词数。之后，你又得到了对数。你应该更好地定义你的问题。当我获得新信息时，我将编辑我的答案。

import re

import os

import math

import heapq

def read_file(path):

    with open(path, 'r', encoding='latin-1') as f:

        return f.read()

DELIM = '[ \n\t0123456789;:.,/\(\)\"\'-]+'

def tokenize(text):

    return re.split(DELIM, text.lower())

def index_text_files_rr(path):

    postings = {}

    doc_lengths = {}

    term_in_document = {}

    files = sorted(os.listdir(path))

    for i, file in enumerate(files):

        file_path = os.path.join(path, file)

        s = read_file(file_path)

        words = tokenize(s)

        length = 0

        # We will store pairs of the word with the decimal logarithm of

        # the word count here to use it later

        words_and_logs = []

        for word in words:

            # Discard empty words

            if word != '':

                # Compute the decimal logarithm of the word count

                log = math.log10(words.count(word))

                # Add the square of the decimal logarithm to the length

                length += log**2

                # Store the word and decimal logarithm pair

                words_and_logs.append((word, log))

        # Compute the square root of the sum of the squares

        # of the decimal logarithms of the words count

        doc_lengths[i] = math.sqrt(length)

        # Iterate over our stored pairs where we already have the

        # decimal logarithms computed so we do not have to do it again

        for word, log in words_and_logs:

            # No need to discard empty words here as we discarded them before

            # so words_and_logs will not have the empty word

            term_in_document.setdefault(log / doc_lengths[i], set()).add(i)

            postings[w] = term_in_document

    return postings

def query_rr(postings, qtext):

    words = tokenize(qtext)

    doc_scores = {}

    for i in range(N):

        score = 0

        for w in words:

            tf = words.count(w)

            df = len(postings[w])

            idf = math.log10(N / (df+1))

            query_weights = tf * idf

        for w in words:

            if w in postings:

                score = score + query_weights

        doc_scores[i] = score

    res = heapq.nlargest(10, doc_scores)

    return res

postings = index_text_files_rr('docs')

print(query_rr(postings, 'hello'))