Performance improvements for indexing text.

This commit is contained in:
2023-01-11 12:38:10 +01:00
parent d830c523a9
commit d7170126f0
2 changed files with 36 additions and 35 deletions
@@ -1,5 +1,6 @@
using System; using System;
using System.Collections.Generic; using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks; using System.Threading.Tasks;
using OpenSearch.Client; using OpenSearch.Client;
using Rsdo.Concordancer.Infrastructure.Search.Extensions; using Rsdo.Concordancer.Infrastructure.Search.Extensions;
@@ -27,23 +28,27 @@ public abstract class BaseAddRecordsHandler<TEntity, TElasticEntity> : IAddRecor
// Get index name // Get index name
var indexName = indexProvider.IndexName; var indexName = indexProvider.IndexName;
// Create request // Batch entities by 1000
var request = new BulkRequest(indexName) foreach (var batch in entities.Chunk(1000))
{ {
Operations = new List<IBulkOperation>(), // Create request
Timeout = TimeSpan.FromMinutes(5), var request = new BulkRequest(indexName)
}; {
Operations = new List<IBulkOperation>(),
Timeout = TimeSpan.FromMinutes(5),
};
// Convert entities to elastic entities // Convert entities to elastic entities
foreach (var entity in entities) foreach (var entity in batch)
{ {
var elasticEntity = ConvertEntity(entity); var elasticEntity = ConvertEntity(entity);
request.Operations.Add(new BulkIndexOperation<TElasticEntity>(elasticEntity)); request.Operations.Add(new BulkIndexOperation<TElasticEntity>(elasticEntity));
}
// Get response
var response = await client.BulkAsync(request);
response.ThrowIfInvalid();
} }
// Get response
var response = await client.BulkAsync(request);
response.ThrowIfInvalid();
} }
protected abstract TElasticEntity ConvertEntity(TEntity entity); protected abstract TElasticEntity ConvertEntity(TEntity entity);
@@ -90,39 +90,35 @@ public class IndexTextHandler : IRequestHandler<IndexText, ExecutionResult>
private static List<Concordance> GetConcordances(Text text, Paragraph paragraph, List<Token> tokens) private static List<Concordance> GetConcordances(Text text, Paragraph paragraph, List<Token> tokens)
{ {
// Get window size (max 10)
var windowSize = Math.Min(tokens.Count - 1, 10);
// Create list of tokens with positions
var tokensIdx = tokens.Select((t, i) => new KeyValuePair<int, Token>(i, t)).ToList();
// Get tokens which will appear in current concordance
List<KeyValuePair<int, Token>> concordanceTokens;
var concordances = new List<Concordance>(); var concordances = new List<Concordance>();
while ((concordanceTokens = tokensIdx.Where(x => x.Key >= -windowSize && x.Key <= windowSize).ToList().OrderBy(x => x.Key).ToList()).Any()) for (var i = 0; i < tokens.Count; i++)
{ {
if (concordanceTokens.Last().Key < 0)
{
break;
}
var concordance = new Concordance() var concordance = new Concordance()
{ {
ParagraphId = paragraph.Id, ParagraphId = paragraph.Id,
TextId = text.Id, TextId = text.Id,
}; };
// Loop through tokens and set it in the position // set center token
foreach (var concordanceToken in concordanceTokens) concordance.SetToken(tokens[i], 0);
// set left and right context tokens
for (var c = 1; c <= 10; c++)
{ {
concordance.SetToken(concordanceToken.Value, concordanceToken.Key); // left context
if (i - c >= 0)
{
concordance.SetToken(tokens[i - c], -c);
}
// right context
if (i + c < tokens.Count)
{
concordance.SetToken(tokens[i + c], c);
}
} }
concordances.Add(concordance); concordances.Add(concordance);
// Decrease indexes of tokens (shift tokens to left, relative to window size)
tokensIdx = tokensIdx.Select(x => new KeyValuePair<int, Token>(x.Key - 1, x.Value)).ToList();
} }
return concordances; return concordances;