117 lines
4.7 KiB
C#
117 lines
4.7 KiB
C#
using CloudBuilder.AI.Entity;
|
||
using CloudBuilder.Core.DatabaseAccessor.Entity;
|
||
using CloudBuilder.Core.DependencyInjection.Task;
|
||
using CloudBuilder.Topshelf.Python;
|
||
using CloudBuilder.Topshelf.Utility;
|
||
using static System.Net.Mime.MediaTypeNames;
|
||
|
||
namespace CloudBuilder.Topshelf.Task
|
||
{
|
||
//task:ChineseNameExtractorTask book_id:B000008
|
||
public class ChineseNameExtractorTask : IScheduleTask
|
||
{
|
||
private readonly IApplicationService service;
|
||
|
||
public ChineseNameExtractorTask(IApplicationService service)
|
||
{
|
||
this.service = service;
|
||
}
|
||
|
||
public void Run(Dictionary<string, string> bodyDict)
|
||
{
|
||
if (!bodyDict.ContainsKey("book_id")) return;
|
||
|
||
string guid = bodyDict["book_id"];
|
||
|
||
PythonCommand python = new PythonCommand(service);
|
||
try
|
||
{
|
||
Hanlp2 hanlp = new Hanlp2(python);
|
||
|
||
IRepository<AiSentenceViewEntity> repository = service.GetRepository<IRepository<AiSentenceViewEntity>>();
|
||
|
||
IRepository<AiSentenceEntity> repositoryAiSentenceEntity = service.GetRepository<IRepository<AiSentenceEntity>>();
|
||
|
||
IRepository<AiChapterPersonEntity> repositoryAiChapterPersonEntity = service.GetRepository<IRepository<AiChapterPersonEntity>>();
|
||
|
||
string depRoot;
|
||
|
||
List<NamedEntity> ners;
|
||
|
||
List<List<MeaningRepresentationParsingEntity>> srls;
|
||
|
||
List<HanlpConstituencyNode> cons = null;
|
||
|
||
SpeakerAnalysisHelper speakerAnalysisHelper = new SpeakerAnalysisHelper();
|
||
|
||
AiSentenceViewEntity[] ents = repository.DetachedEntities.Where(x => x.BookId == guid).OrderBy(x => x.ChapterId).ThenBy(x => x.ParagraphId).ThenBy(x => x.SentenceIndex).ToArray();
|
||
if (ents == null || ents.Length == 0) return;
|
||
|
||
AiSentenceEntity aiSentenceEntity;
|
||
string person = null;
|
||
long paragraphIndex = 0;
|
||
|
||
int totalLines = ents.Length;
|
||
int processedLines = 0; // 已处理行数计数器
|
||
AiChapterPersonEntity aiChapterPerson;
|
||
List<AiChapterPersonEntity> cps = new List<AiChapterPersonEntity>();
|
||
foreach (var ent in ents)
|
||
{
|
||
processedLines++;
|
||
|
||
person = speakerAnalysisHelper.GetPerson(hanlp, ent.Content);
|
||
paragraphIndex = ent.ParagraphId;
|
||
|
||
if (!string.IsNullOrEmpty(person))
|
||
{
|
||
string[] ps = person.Split(',');
|
||
int orderIndex = 1;
|
||
foreach (var p in ps)
|
||
{
|
||
if (p.Length == 1) continue;
|
||
aiChapterPerson = new AiChapterPersonEntity();
|
||
aiChapterPerson.PersonName = p;
|
||
aiChapterPerson.BookId = ent.BookId;
|
||
aiChapterPerson.ChapterId = ent.ChapterId;
|
||
aiChapterPerson.ParagraphIndex = ent.ParagraphId;
|
||
aiChapterPerson.SentenceIndex= ent.SentenceIndex;
|
||
aiChapterPerson.OrderIndex = orderIndex++;
|
||
cps.Add(aiChapterPerson);
|
||
repositoryAiChapterPersonEntity.InsertNow(aiChapterPerson);
|
||
}
|
||
}
|
||
|
||
UpdateProgress(processedLines, totalLines);
|
||
}
|
||
|
||
speakerAnalysisHelper.AiChapterPersons = cps.ToArray();
|
||
|
||
speakerAnalysisHelper.MatchPerson(ents, repositoryAiSentenceEntity);
|
||
}
|
||
catch (Exception ex)
|
||
{
|
||
Console.WriteLine(ex.Message);
|
||
}
|
||
finally
|
||
{
|
||
Console.WriteLine("");
|
||
python.Dispose();
|
||
}
|
||
}
|
||
|
||
private static void UpdateProgress(int processed, int total)
|
||
{
|
||
double progressPercent = (double)processed / total * 100;
|
||
int progressBarLength = 50; // 进度条总长度
|
||
int filledLength = (int)(progressPercent / 100 * progressBarLength);
|
||
|
||
// 构建进度条（如：[██████████░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░]）
|
||
string progressBar = "[" + new string('█', filledLength) + new string('░', progressBarLength - filledLength) + "]";
|
||
|
||
string progressInfo = $"{progressBar} {progressPercent:F2}% | 已处理：{processed}/{total} ";
|
||
Console.Write($"\r{progressInfo}");
|
||
Console.Out.Flush();
|
||
}
|
||
}
|
||
}
|