QA/mapping1/onehot_generate.py

import json
import numpy as np
import os
path = os.path.abspath(os.path.dirname(__file__))
def onehot_generate():
    # 从文件读取映射关系
    with open(path + "/mapping.json", "r", encoding="utf-8") as json_file:
        loaded_data = json.load(json_file)

    # 提取关键词和故障现象
    keywords_list = []
    phenomenon_list = []

    for mapping in loaded_data:
        keywords_list.extend(mapping["keywords"])
        phenomenon_list.append(mapping["phenomenon"])

    # 生成关键词词表
    vocabulary = sorted(set(keywords_list))

    # 生成one-hot编码
    one_hot_vectors = []
    for word in vocabulary:
        one_hot = [1 if word == w else 0 for w in vocabulary]
        one_hot_vectors.append(one_hot)

    # 转换为NumPy数组
    one_hot_array = np.array(one_hot_vectors, dtype=np.int32)

    # 生成关键词与one-hot编码的字典
    keyword_one_hot_dict = {}
    for i, word in enumerate(vocabulary):
        keyword_one_hot_dict[word] = one_hot_array[i].tolist()

    # 生成故障现象词的向量
    phenomenon_vectors = []
    for phenomenon in phenomenon_list:
        vector_sum = np.zeros(len(vocabulary), dtype=np.int32)
        for keyword in keywords_list:
            if keyword in phenomenon:
                vector_sum = np.logical_or(vector_sum, np.array(keyword_one_hot_dict[keyword], dtype=np.int32))
        phenomenon_vectors.append(vector_sum.tolist())

    # 创建故障现象词的向量字典
    phenomenon_vector_dict = {}
    for i, phenomenon in enumerate(phenomenon_list):
        phenomenon_vector_dict[phenomenon] = [int(val) for val in phenomenon_vectors[i]]

    # 保存关键词与one-hot编码的字典为JSON文件
    with open(path + "/keyword_one_hot_dict.json", "w", encoding="utf-8") as json_file:
        json.dump(keyword_one_hot_dict, json_file, ensure_ascii=False, indent=4)

    # 保存故障现象词的向量字典为JSON文件
    with open(path + "/phenomenon_vector_dict.json", "w", encoding="utf-8") as json_file:
        json.dump(phenomenon_vector_dict, json_file, ensure_ascii=False, indent=4)