diff options
| -rw-r--r-- | Dockerfile | 2 | ||||
| -rw-r--r-- | Server.csproj | 4 | ||||
| -rw-r--r-- | Services/OcrService.cs | 6 |
3 files changed, 8 insertions, 4 deletions
@@ -7,7 +7,7 @@ RUN dotnet publish -o out FROM mcr.microsoft.com/dotnet/aspnet:10.0@sha256:ccdca44cd4f256d50187f920dc8ccc2a9ea7a8a4597ac1d51e08fddb2e3b3205 RUN apt update -RUN apt install -y imagemagick +RUN apt install -y imagemagick tesseract-ocr tesseract-ocr-eng RUN apt clean RUN rm -rf /var/lib/apt/lists/* WORKDIR /App diff --git a/Server.csproj b/Server.csproj index f733581..e435946 100644 --- a/Server.csproj +++ b/Server.csproj @@ -6,9 +6,9 @@ <ImplicitUsings>enable</ImplicitUsings> <AssemblyName>HyperBooru</AssemblyName> <RootNamespace>HyperBooru</RootNamespace> - <AssemblyVersion>0.10.0.0</AssemblyVersion> + <AssemblyVersion>0.11.0.0</AssemblyVersion> <FileVersion>$(AssemblyVersion)</FileVersion> - <Version>0.10-alpha</Version> + <Version>0.11-alpha</Version> <UserSecretsId>2907567f-4640-4581-8f4d-0977952d26bd</UserSecretsId> </PropertyGroup> diff --git a/Services/OcrService.cs b/Services/OcrService.cs index 40905aa..d43db2e 100644 --- a/Services/OcrService.cs +++ b/Services/OcrService.cs @@ -8,6 +8,8 @@ using Tesseract; namespace HyperBooru.Services; public class OcrService : IHostedService { + private readonly string[] InvalidMimeTypes = [ "image/heic", "image/webp" ]; + private readonly TimeSpan ProcessInterval = TimeSpan.FromMinutes(30); private readonly TimeSpan StartupDelay = TimeSpan.FromSeconds(30); @@ -65,10 +67,12 @@ public class OcrService : IHostedService { using var db = dbFactory.CreateDbContext(); Guid[] guids = db.Media + .AsNoTracking() .Include(m => m.CurrentUploadedFile) .Include(m => m.OcrData) .Where(m => m.OcrData == null) - .Where(m => m.CurrentUploadedFile.MimeType.Contains("image/")) + .Where(m => m.CurrentUploadedFile!.MimeType.Contains("image/")) + .Where(m => !InvalidMimeTypes.Contains(m.CurrentUploadedFile!.MimeType)) .Select(m => m.Guid) .ToArray(); db.Dispose(); |