Trash Can

One man's trash is another man's related work.

trashcan -- web based service for indexing scientific papers (or any other documents)

Features:

• Single lightweight backend service written in C++ using SQLite database.
• Single page web interface. PDF files can be preview in the same browser window.
• Documents metadata can be fetched using DOI, ArXiv or ISBN (using Google Books) identifiers.
• Documents can be categorized using user-defined tags.
• Full-text search support. Documents contents is extracted with Apache Tika and indexed using SQLite FTS.

Screenshots

Search page:

Edit page:

Mobile version:

Deploying with Docker

Build container:

docker build -t trashcan .

If you are going to run the service on non-root URL prefix, for example /trash/, you should specify the PREFIX argument:

docker build --build-arg PREFIX=/trash/ -t trashcan .

Start container:

docker run \
    -p 8000:8000 \
    -v $HOME/trashcan/:/var/lib/trashcan/ \
    -t trashcan

trashcan service would listen on 8000 port (http://localhost:8000/) address and store its files in your $HOME/trashcan/ directory.

Compiling from source

Requirements:

• meson build system
• sqlite3
• libmagic (libmagic-dev or file-devel packages)
• Apache Tika App
• nodejs, npm (optional, required only for frontend compilation)

You can download compiled frontend assets archive (frontend.tar.gz) from GitLab Release page and extract it to the project root. Or, instead, you may compile them manually:

cd frontend
npm install 
npm run build 
cd ..

Configure backend project:

mkdir build && cd build
meson ..

By default, your files and database would be stored in /var/lib/trashcan/ directory, to change you can use command line arguments later or compile it with the corresponding option:

mkdir build && cd build
meson .. -Ddefault-storage-dir=/path/to/your/trash/

Compile and install it:

meson compile
sudo meson install

By default, frontend files and content extraction script would be placed in $PREFIX/usr/share and backend executable would be in $PREFIX/bin.

Usage

You'll need to set path to tika-app jar file to TIKA_JAR_PATH environment variable, or you may create a link path to tika-app.jar in the current workdir (check script/extract.sh).

export TIKA_JAR_PATH=/home/user/tika/tika-app-1.26.jar
# or 
# ln -s /home/user/tika/tika-app-1.26.jar tika-app.jar

And then you may start backend service:

./build/trashcan

Run ./build/trashcan -h to see more configuration options.

Systemd service

Make sure TIKA jar files is installed to /usr/share/java/tika-app.jar, or change the path in the service file.

loginctl enable-linger $USER

mkdir -p ~/.config/systemd/user/
cp trashcan.service ~/.config/systemd/user/
systemctl --user enable trashcan
systemctl --user start trashcan

Nginx integration

Nginx can be used as a reverse proxy and for dispatching static content. In the following configuration nginx would serve static files from /trash/ prefix and redirect request coming to /trash/api/ to the backend service.

To change prefixes for static assets and backend API in the frontend, you'd need to specify them in .env.local file and recompile the frontend:

cd frontend
cat > .env.local <<EOF
VUE_APP_PUBLIC_URL=/trash/
VUE_APP_BACKEND_URL=/trash/api/
EOF

npm install 
npm run build 
cd ..

To specify backend API prefix, you would need to compile it with api-prefix option:

meson .. -Dapi-prefix=/trash/api/ 
meson compile
sudo meson install

Then, in nginx configuration file, you'd need to set /trash/ location as an alias to your static content and /trash/api as a proxy pass to the backend. Make sure to increase request size in order for file uploads to work.

server {
	listen 80;
	location /trash/ {
		autoindex off;
		alias /usr/local/share/trashcan/dist/;
	}
	location /trash/api/ {
		client_max_body_size 300M;
		proxy_pass http://127.0.0.1:8000;
	}
}