裁判文书关键信息提取

 项目初期的部分代码，比较粗糙，随便看看就行了，效果不太好，且我后面又改了很多。

2021年11月8日

"""单个文件等级地处理""" # 程序实现：1.传入一个路径(doc或docx文件路径，系统会自动判断是doc文件还是docx文件，doc文件会转换为docx文件），在本路径下生成同文件名json文件。2.传入一个路径，返回一个列表（内嵌套固定类型字典）。 import os import docx import json from process_doc import doc2docx_1 # 通过文件路径解析信息（txt文件编码类型为ANSI），获取案件名（文件名）、案由、案件类型、文书类型 def case_name(filepath, anyoutxt, leixingtxt, papertypetxt, casetypetxt): # 获取案件名 (path, filename) = os.path.split(filepath) (casename, ext) = os.path.splitext(filename) # 获取案由 anyou = '' f01 = open(anyoutxt, "r") # 设置文件对象，其中txt文件编码类型为ANSI data01 = f01.readlines() # 直接将文件中按行读到list里，效果与方法2一样 f01.close() # 关闭文件 for ay in data01: ay = ay.replace('\n', '') # 删除回车 if ay in casename: anyou = ay # 获取审理程序 chengxu = '' f02 = open(leixingtxt, "r") # 设置文件对象 data02 = f02.readlines() # 直接将文件中按行读到list里，效果与方法2一样 f02.close() # 关闭文件 for ay in data02: ay = ay.replace('\n', '') # 删除回车 if ay in casename: chengxu = ay # 获取文书类型 papertype = '' f03 = open(papertypetxt, "r") # 设置文件对象 data03 = f03.readlines() # 直接将文件中按行读到list里，效果与方法2一样 f03.close() # 关闭文件 for ay in data03: ay = ay.replace('\n', '') # 删除回车 if ay in casename: papertype = ay # 获取诉讼程序 casetype = '' f04 = open(casetypetxt, "r") # 设置文件对象 data04 = f04.readlines() # 直接将文件中按行读到list里，效果与方法2一样 f04.close() # 关闭文件 for ay in data04: ay = ay.replace('\n', '') # 删除回车 if ay in casename: casetype = ay if casename =='': print('未找到案件名') casename = 'void' if anyou == '': print("未找到案由！") anyou = 'void' if chengxu == '': print("匹配审理程序失败！") chengxu = 'void' if papertype == '': print("解析文书类型失败！") papertype = 'void' if casetype == '': print("解析案件类型失败！") casetype = 'void' return (casename, anyou, chengxu, papertype, casetype) # 后两项仅用于传入case_base_information函数 # 读取docx文档,并将每段存入一个列表 def read_docx(path): if not os.path.exists(path): print('发生错误：\n'+path+'文件不存在！') exit(0) # 无错误结束程序 doc = docx.Document(path) paras_temp = [] # 用来存储段落 for p in doc.paragraphs: p1 = p.text.replace(" ", "") # 去除空格,并以字符串的形式存储在列表中 paras_temp.append(p1) print(path + ": 文件读取成功") return paras_temp # 传入读取的文件段落列表,获取案件基本信息:法院，文书性质，案号,判决日期 def case_base_information(paras, papertype, casetype, papertypetxt, casetypetxt): case_court = paras[0] type = paras[1] case_number = paras[2] # 变量名分配 prosecution_organ = '' # 公诉机关 part_people_index0 = 0 # 参与人第一段索引(初始化 part_people_index1 = 0 # 参与人最后一段索引(初始化 law_index0 = 0 # 相关法律初始段（初始化 defendants = [] # 被告人 law_relate = [] # 获取文书类型 if papertype == 'void': f03 = open(papertypetxt, "r") # 设置文件对象 data03 = f03.readlines() # 直接将文件中按行读到list里，效果与方法2一样 f03.close() # 关闭文件 for ay in data03: ay = ay.replace('\n', '') # 删除回车 if ay in type: papertype = ay # 获取诉讼程序 if casetype == 'void': f04 = open(casetypetxt, "r") # 设置文件对象 data04 = f04.readlines() # 直接将文件中按行读到list里，效果与方法2一样 f04.close() # 关闭文件 for ay in data04: ay = ay.replace('\n', '') # 删除回车 if ay in type: casetype = ay # 遍历检索 judge_day = 'void' n1 = 1 for p in paras: # 提取公诉机关（检察院） if '公诉机关' in p: if len(prosecution_organ) <= 0: prosecution_organ = p.replace('公诉机关','') prosecution_organ = prosecution_organ.replace('。','') part_people_index0 = paras.index(p) + 1 n1 += 1 # 案由段，暂时未提取！！！！！！！ elif '已审理终结' in p: anyou = p if '年' and '月' and '日' in p: if len(p) <= 12: judge_day = p # 相关法律条文起始段索引 if '法律条文' in p: law_index0 = paras.index(p) if n1 ==2 and (prosecution_organ in p): part_people_index1 = paras.index(p) - 1 # 提取被告人信息（包括辩护人） while part_people_index0 <= part_people_index1: defendants.append(paras[part_people_index0]) part_people_index0 += 1 # 相关法律条文 if law_index0 > 0: while law_index0 < len(paras): law_relate.append(paras[law_index0]) law_index0 += 1 if prosecution_organ == '': print('未提取到公诉机关') prosecution_organ = 'void' if defendants == []: print('未提取到被告人信息') defendants.append('void') if law_relate == []: print('文书未附相关法律条文') law_relate.append('void') if judge_day == 'void': print('提取判决日期失败') return (papertype, casetype, case_court, case_number, judge_day, defendants, prosecution_organ, law_relate) # 将信息存入字典（在此设置字典）（暂时不用） def information_dict(casename, anyou, chengxu, papertype, casetype, fayuan, wenshuleixing, anhao, panjueriqi, beigaoren, gongsuren): info_dict_temp = {'案件名': casename, '案号':anhao, '案由':anyou, '程序': chengxu, '文书类型':papertype, '案件类型': casetype, '法院':fayuan, '文书类型':wenshuleixing, '判决日期': panjueriqi, '被告人':beigaoren, '公诉人':gongsuren } return info_dict_temp # 将信息存入列表（内嵌套字典） def information_list(casename, anyou, chengxu, papertype, anjianleixing, fayuan, anhao, panjueriqi, beigaoren, gongsuren, law_ralated): info_list = [ {'案件名': casename}, {'案由': anyou}, {'程序': chengxu}, {'文书类型': papertype}, {'案件类型': anjianleixing}, {'法院': fayuan}, {'案号':anhao}, { '判决日期': panjueriqi}, {'被告人': beigaoren}, { '公诉人':gongsuren}, {'相关法律': law_ralated} ] return info_list # 字典转json文件（savefile_load是文件夹路径） def dict2json(info_dict, savefile_load, casename): b = json.dumps(info_dict, ensure_ascii=False) savefile_load = savefile_load + '/' + casename + '.json' f2 = open(savefile_load, 'w') f2.write(b) f2.close() def list_output(file_path): if os.path.splitext(file_path)[-1] == ".doc": # 判断文件类型是否为doc print('传入文件格式为doc，将转换为docx格式。') doc2docx_1(file_path) file_path = file_path + 'x' # 设置引用文件路径（因为后期可能要更新相应信息，所以采用此种方式更方便一些） anyoutxt = 'D:/projects_pycharm/lawdocs2excel/data_access/data/anyou.txt' leixingtxt = 'D:/projects_pycharm/lawdocs2excel/data_access/data/chengxu.txt' papertypetxt = 'D:/projects_pycharm/lawdocs2excel/data_access/data/paper_type.txt' casetypetxt = 'D:/projects_pycharm/lawdocs2excel/data_access/data/case_type.txt' (casename, anyou, chengxu, papertype0, casetype0) = \ case_name(file_path, anyoutxt, leixingtxt, papertypetxt, casetypetxt) paras = read_docx(file_path) # 读取文书段落 (papertype, casetype, fayuan, anhao, panjueriqi, beigaoren, gongsuren, law_ralated) \ = case_base_information(paras, papertype0, casetype0, papertypetxt, casetypetxt) # 读取基本信息 info_list = information_list(casename, anyou, chengxu, papertype, casetype, fayuan, anhao, panjueriqi, beigaoren, gongsuren, law_ralated) return info_list if __name__ == '__main__': path = 'E:/NLP/doc2docx_test/王云清应先勇等强迫交易罪非法占用农用地罪夏君兵王某29强迫交易罪一审刑事判决书.doc' info_list = list_output(path) for i in info_list: print(i)

以下为文件内容举例。

anyou.txt

paper_type.txt

代码补充：

# 实现单个doc文件转换为docx文件 def doc2docx_1(doc_path): # 打开word并设置 word = wc.Dispatch("Word.Application") word.Visible = 0 # 1程序可见，0不可见 print(doc_path) doc = word.Documents.Open(doc_path) docx_save_path = doc_path + 'x' doc.SaveAs(docx_save_path, 12, False, "", True, "", False, False, False, False) # 转换后的文件,12代表转换后为docx文件 doc.Close() word.Quit()