CodeAnalizer/CodeBase/Services/ScipParser.cs

using CodeBase.Models;
using CodeBase.Orchestrators;
using CodeBase.Services;
using Google.Protobuf;
using System;
using System.Collections.Generic;
using System.IO;
using System.Linq;
using System.Threading.Tasks;

namespace CodeBase.Parsers
{
    public class ScipParser(ScipOrchestrator scipOrchestrator)
    {
        public async Task<List<CodeChunk>> ParseProjectAsync(string projectRootPath, string language, string name)
        {
            var scipFilePath = Path.Combine(projectRootPath, "index.scip");
            if (!File.Exists(scipFilePath))
            {
                await scipOrchestrator.GenerateScipAsync(projectRootPath, language);
            }

            var filter = new UniversalFileFilter();

            using var stream = File.OpenRead(scipFilePath);
            var scipIndex = Scip.Index.Parser.ParseFrom(stream);

            var extractedChunks = new List<CodeChunk>();

            foreach (var document in scipIndex.Documents)
            {
                if (!filter.IsValidCodeFile(document.RelativePath)) continue;

                var absoluteFilePath = Path.Combine(projectRootPath, document.RelativePath);
                if (!File.Exists(absoluteFilePath)) continue;

                var fileLines = await File.ReadAllLinesAsync(absoluteFilePath);
                var declarations = document.Occurrences.Where(o => (o.SymbolRoles & 1) == 1);

                foreach (var occ in declarations)
                {
                    int startLine = occ.Range[0];
                    int endLine = occ.Range.Count == 3 ? occ.Range[0] : occ.Range[2];

                    if (startLine < 0 || endLine >= fileLines.Length) continue;

                    var codeSnippet = string.Join(
                        Environment.NewLine,
                        fileLines.Skip(startLine).Take(endLine - startLine + 1)
                    );

                    // Возвращаем чанки пока БЕЗ векторов
                    extractedChunks.Add(new CodeChunk
                    {
                        Id = occ.Symbol,
                        EntityName = ExtractSimpleName(occ.Symbol),
                        FilePath = document.RelativePath,
                        Content = codeSnippet,
                        Language = document.Language,
                        ProjectName = name
                    });
                }
            }

            return extractedChunks;
        }

        private string ExtractSimpleName(string scipSymbol)
        {
            if (string.IsNullOrEmpty(scipSymbol)) return "Unknown";
            var parts = scipSymbol.Split(new[] { '#', '.', '(', ')' }, StringSplitOptions.RemoveEmptyEntries);
            return parts.LastOrDefault() ?? scipSymbol;
        }
    }
}