xSaherElmAiApi/Providers/Extractors/XFileContentExtractor.cs

using System.IO;
using System.Linq;
using System.Threading;
using xAiModels.Models;
using xCommons.Extensions;
using xExceptions.Constants;
using System.Threading.Tasks;
using System.Collections.Generic;
using xAiApi.Interfaces.Extractors;

namespace xAiApi.Providers.Extractors
{
    /// <summary>
    /// Composite extractor that delegates to appropriate extractor 
    /// based on MIME type ...
    /// </summary>    
    public class XFileContentExtractor : IXFileContentExtractor
    {
        //
        private readonly IList<IXFileContentExtractorBase> _extractors = [];

        public XFileContentExtractor(
            IEnumerable<IXFileContentExtractorBase> availableExtractors
        )
        {
            //
            _extractors = [.. availableExtractors
                .Where(e => e.GetType() != typeof(XFileContentExtractor))];
        }

        /// <summary>
        /// Check if this extractor supports the specified MIME type ...
        /// </summary>
        /// <param name="mimeType"></param>
        /// <param name="extractors"></param>
        /// <returns></returns>
        public bool CanExtract(
            string mimeType,
            string[] extractors = null
        )
        {
            //
            return GetExtractors(extractors)
                .Any(e => e.CanExtract(mimeType));
        }

        /// <summary>
        /// Extract text content from file stream ...
        /// </summary>
        /// <param name="fileStream"></param>
        /// <param name="mimeType"></param>
        /// <param name="extractors"></param>
        /// <param name="cancellationToken"></param>
        /// <returns></returns>
        public async Task<string> ExtractAsync(
            Stream fileStream,
            string mimeType,
            string[] extractors = null,
            CancellationToken cancellationToken = default
        )
        {
            //
            var extractor = GetExtractors(extractors)
                .FirstOrDefault(e => e.CanExtract(mimeType));
            if (extractor.IsNull())
            {
                //
                XException.NotAllowed.Throw(
                    $"Unsupported file type: {mimeType}"
                );
            }

            //
            var result = await extractor
                .ExtractAsync(
                    mimeType: mimeType,
                    fileStream: fileStream,
                    cancellationToken: cancellationToken
                );

            //
            return result;
        }

        /// <summary>
        /// Extract content from stream as Rich Result ...
        /// </summary>
        /// <param name="fileStream"></param>
        /// <param name="fileName"></param>
        /// <param name="mimeType"></param>
        /// <param name="extractors"></param>
        /// <param name="cancellationToken"></param>
        /// <returns></returns>
        public async Task<XFileExtractionResult> ExtractRichAsync(
            Stream fileStream,
            string fileName,
            string mimeType,
            string[] extractors = null,
            CancellationToken cancellationToken = default
        )
        {
            //
            var extractor = GetExtractors(extractors)
                .FirstOrDefault(e => e.CanExtract(mimeType));
            if (extractor.IsNull())
            {
                //
                XException.NotAllowed.Throw(
                    $"Unsupported file type: {mimeType}"
                );
            }

            //
            var result = await extractor
                .ExtractRichAsync(
                    mimeType: mimeType,
                    fileName: fileName,
                    fileStream: fileStream,
                    cancellationToken: cancellationToken
                );

            //
            return result;
        }

        /// <summary>
        /// Extract Provided Extractors ...
        /// </summary>
        /// <param name="extractors"></param>
        /// <returns></returns>
        private IList<IXFileContentExtractorBase> GetExtractors(string[] extractors = null)
        {
            //
            var result = new List<IXFileContentExtractorBase>();

            //
            if (extractors.IsNull() || !extractors.HasChild())
            {
                //
                result = [.. _extractors];
            }
            else
            {
                //
                result = [.. _extractors
                    .Where(e => extractors
                        .Any(ex => e.GetType().FullName.Contains(ex)))];
            }

            //
            return result;
        }
    }
}