public IEnumerable<MedStandardInfo> Find(string keywords, IProgress<string> progress)
{
using (var directory = GetDirectory())
using (var searcher = new IndexSearcher(directory))
{
var query = GetQuery(keywords);
var sort = GetSort();
var docs = searcher.Search(query, null, 1000, sort);
var result = new List<MedStandardInfo>();
foreach (var scoreDoc in docs.ScoreDocs)
{
var doc = searcher.Doc(scoreDoc.Doc);
var product = new MedStandardInfo()
{
StandardName = doc.Get("StandardName"),
FileName = doc.Get("FileName"),
Mkb = doc.Get("Mkb"),
OrderNum = doc.Get("OrderNum"),
};
result.Add(product);
}
return result;
}
}
public async Task<MedStandardInfo> ReadPdfAsync(string fileName, CancellationToken cancellationToken, IProgress<string> progress = null)
{
if (File.Exists(fileName))
{
progress = progress ?? new Progress<string>();
progress.Report($@"Чтение текста файла {System.IO.Path.GetFileName(fileName)} ...");
var text = new StringBuilder();
await Task.Run(() =>{
using (PdfReader reader = new PdfReader(fileName))
for (int i = 1; i <= reader.NumberOfPages; i++)
text.Append(PdfTextExtractor.GetTextFromPage(reader, i));
}, cancellationToken);
var result = new MedStandardInfo();
result.FileName = System.IO.Path.GetFileName(fileName);
result.Text = text.ToString();
var lines = result.Text.ToLower().Split(new [] { '\n'}).Select(x => x.Trim()).ToList();
var regNum = new Regex(@"№\s*\d+\s*(н)?");
var ordNumLineIndex = lines.FindIndex(x => regNum.IsMatch(x));
if (ordNumLineIndex >= 0)
{
var num = regNum.Match(lines[ordNumLineIndex]).Value;
result.OrderNum = num.Replace(@"№", "").Trim();
}
var lineIndex = lines.FindIndex(x => x.Contains(@"зарегистрировано"));
if (lineIndex >= 0)
{
lineIndex = lines.FindIndex(lineIndex, x => x.Contains(@"стандарт"));
if (lineIndex >= 0)
{
do
{
result.StandardName += " " + lines[lineIndex++];
} while (!lines[lineIndex].Contains(":") && !lines[lineIndex].Contains("мероприят"));
result.StandardName = result.StandardName.Trim().Replace(" ", " ");
lineIndex = lines.FindIndex(0, x => x.Contains("код по мкб") || x.Contains("нозолог"));
if (lineIndex >= 0)
{
var reg = new Regex(@"[a-z,а-я]\d{1,2}(\.\d{1,2})?");
for (int i = lineIndex; i < lines.Count; i++)
{
foreach (var match in reg.Matches(lines[i]).OfType<Match>())
{
result.Mkb += " " + match.Value;
}
if(lines[i].Contains("мероприят") || lines[i].Contains("услуги"))
break;
}
if (!string.IsNullOrEmpty(result.Mkb))
{
result.Mkb = Translit(result.Mkb.Trim().Replace(" ", " ").ToUpper());
result.Text = string.Join(" ", lines.GetRange(0, lineIndex));
}
}
else
progress.Report(@"Не найдена секция диагнозов");
}
else
{
progress.Report(@"Не найдено слово СТАНДАРТ");
}
}
return result;
}
else
throw new FileNotFoundException($@"Файл {fileName} не найден", fileName);
}
private Document MapMedStandard(MedStandardInfo medStandard)
{
var document = new Document();
document.Add(new Field("StandardName", medStandard.StandardName, Field.Store.YES, Field.Index.ANALYZED));
document.Add(new Field("Text", medStandard.Text, Field.Store.YES, Field.Index.ANALYZED));
document.Add(new Field("Mkb", medStandard.Mkb, Field.Store.YES, Field.Index.ANALYZED));
document.Add(new Field("OrderNum", medStandard.OrderNum, Field.Store.YES, Field.Index.NOT_ANALYZED_NO_NORMS));
document.Add(new Field("FileName", medStandard.FileName, Field.Store.YES, Field.Index.NOT_ANALYZED_NO_NORMS));
return document;
}