From 063aa5a470c4f24d6cecc85dbc38df571fb407c4 Mon Sep 17 00:00:00 2001
From: Alex Kriuchykhin <oleksii@scinote.net>
Date: Thu, 20 Feb 2025 09:41:41 +0100
Subject: [PATCH] Install tika text extractor directly from Dockerfiles, remove
 old yomu gem [SCI-11589] (#8257)

---
 Dockerfile                                              | 5 +++++
 Dockerfile.production                                   | 5 +++++
 Gemfile                                                 | 1 -
 Gemfile.lock                                            | 1 -
 app/models/concerns/tiny_mce_images.rb                  | 2 ++
 lib/active_storage/analyzer/text_extraction_analyzer.rb | 5 ++++-
 6 files changed, 16 insertions(+), 3 deletions(-)

diff --git a/Dockerfile b/Dockerfile
index da11bb47f..c8c236fd2 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -1,6 +1,9 @@
 FROM ruby:3.2-bookworm
 MAINTAINER SciNote <info@scinote.net>
 
+ARG TIKA_DIST_URL="https://dlcdn.apache.org/tika/2.9.3/tika-app-2.9.3.jar"
+ENV TIKA_PATH=/usr/local/bin/tika-app.jar
+
 # additional dependecies
 # libreoffice for file preview generation
 RUN apt-get update -qq && \
@@ -23,6 +26,8 @@ RUN apt-get update -qq && \
   chromium \
   chromium-sandbox \
   yarnpkg && \
+  wget -O $TIKA_PATH $TIKA_DIST_URL && \
+  chmod +x $TIKA_PATH && \
   ln -s /usr/lib/x86_64-linux-gnu/libvips.so.42 /usr/lib/x86_64-linux-gnu/libvips.so && \
   rm -rf /var/lib/apt/lists/*
 
diff --git a/Dockerfile.production b/Dockerfile.production
index 2a8af22f0..aedacaf84 100644
--- a/Dockerfile.production
+++ b/Dockerfile.production
@@ -44,6 +44,9 @@ RUN \
 FROM ruby:3.2-bookworm AS runner
 MAINTAINER SciNote <info@scinote.net>
 
+ARG TIKA_DIST_URL="https://dlcdn.apache.org/tika/2.9.3/tika-app-2.9.3.jar"
+ENV TIKA_PATH=/usr/local/bin/tika-app.jar
+
 RUN rm -f /etc/apt/apt.conf.d/docker-clean; echo 'Binary::apt::APT::Keep-Downloaded-Packages "true";' > /etc/apt/apt.conf.d/keep-cache
 RUN \
   --mount=type=cache,target=/var/cache/apt,sharing=locked \
@@ -79,6 +82,8 @@ RUN \
   chromium-sandbox \
   libfile-mimeinfo-perl \
   yarnpkg && \
+  wget -O $TIKA_PATH $TIKA_DIST_URL && \
+  chmod +x $TIKA_PATH && \
   /usr/share/nodejs/yarn/bin/yarn add puppeteer@npm:puppeteer-core@^22.15.0 && \
   apt-get install -y libreoffice && \
   ln -s /usr/lib/x86_64-linux-gnu/libvips.so.42 /usr/lib/x86_64-linux-gnu/libvips.so
diff --git a/Gemfile b/Gemfile
index af5807835..62b9e7530 100644
--- a/Gemfile
+++ b/Gemfile
@@ -17,7 +17,6 @@ gem 'recaptcha'
 gem 'sanitize'
 gem 'sprockets-rails'
 gem 'view_component'
-gem 'yomu', git: 'https://github.com/scinote-eln/yomu', branch: 'master'
 
 # Gems for OAuth2 subsystem
 gem 'doorkeeper', '>= 4.6'
diff --git a/Gemfile.lock b/Gemfile.lock
index b08be085f..f0aef81a6 100644
--- a/Gemfile.lock
+++ b/Gemfile.lock
@@ -893,7 +893,6 @@ DEPENDENCIES
   webmock
   whacamole
   wicked_pdf
-  yomu!
   zip-zip
 
 RUBY VERSION
diff --git a/app/models/concerns/tiny_mce_images.rb b/app/models/concerns/tiny_mce_images.rb
index cc0bcaa88..eae246af5 100644
--- a/app/models/concerns/tiny_mce_images.rb
+++ b/app/models/concerns/tiny_mce_images.rb
@@ -1,5 +1,7 @@
 # frozen_string_literal: true
 
+require 'mime/types'
+
 module TinyMceImages
   extend ActiveSupport::Concern
 
diff --git a/lib/active_storage/analyzer/text_extraction_analyzer.rb b/lib/active_storage/analyzer/text_extraction_analyzer.rb
index 5925f042b..41d22f54e 100644
--- a/lib/active_storage/analyzer/text_extraction_analyzer.rb
+++ b/lib/active_storage/analyzer/text_extraction_analyzer.rb
@@ -2,6 +2,8 @@
 
 module ActiveStorage
   class Analyzer::TextExtractionAnalyzer < Analyzer
+    DEFAULT_TIKA_PATH = 'tika-app.jar'
+
     def self.accept?(blob)
       blob.content_type.in?(Constants::TEXT_EXTRACT_FILE_TYPES) && blob.attachments.where(record_type: 'Asset').any?
     end
@@ -40,7 +42,8 @@ module ActiveStorage
     end
 
     def process_other(file)
-      text_data = Yomu.new(file.path).text
+      tika_path = ENV['TIKA_PATH'] || DEFAULT_TIKA_PATH
+      text_data = IO.popen(['java', '-Djava.awt.headless=true', '-jar', tika_path, '-t', file.path], 'r').read
       create_or_update_text_data(text_data)
     end