lookyloo/README.md

![Lookyloo icon](website/web/static/lookyloo.jpeg)

*Lookyloo* is a web interface allowing to scrape a website and then displays a
tree of domains calling each other.

Thank you very much [Tech Blog @ willshouse.com](https://techblog.willshouse.com/2012/01/03/most-common-user-agents/)
for the up-to-date list of UserAgents.

# What is that name?!


```
1. People who just come to look.
2. People who go out of their way to look at people or something often causing crowds and more disruption.
3. People who enjoy staring at watching other peoples misfortune. Oftentimes car onlookers to car accidents.
Same as Looky Lou; often spelled as Looky-loo (hyphen) or lookylou
In L.A. usually the lookyloo's cause more accidents by not paying full attention to what is ahead of them.
```

Source: [Urban Dictionary](https://www.urbandictionary.com/define.php?term=lookyloo)

# Screenshot

![Screenshot of Lookyloo](doc/example.png)

# Implementation details

This code is very heavily inspired by [webplugin](https://github.com/etetoolkit/webplugin) and adapted to use flask as backend.

The two core dependencies of this project are the following:

* [ETE Toolkit](http://etetoolkit.org/): A Python framework for the analysis and visualization of trees.
* [Splash](https://splash.readthedocs.io/en/stable/): Lightweight, scriptable browser as a service with an HTTP API

# Cookies

If you want to scrape a website as if you were loggged in, you need to pass your sessions cookies.
You can do it the the folloing way:

1. Install [Cookie Quick Manager](https://addons.mozilla.org/en-US/firefox/addon/cookie-quick-manager/)
2. Click onthe icon in the top right of your browser > Manage all cookies
3. Search for a domain, tick the Sub-domain box if needed
4. Right clock on the domain you want to export > save to file > $LOOKYLOO_HOME/cookies.json

Then, you need to restart the webserver and from now on, every cookies you have in that file will be available for the browser used by Splash

# Python client

You can use `pylookyloo` as a standalone script, or as a library, [more details here](https://github.com/CIRCL/lookyloo/tree/master/client)

# Installation

**IMPORTANT**: Use [pipenv](https://pipenv.readthedocs.io/en/latest/)

**NOTE**: Yes, it requires python3.6+. No, it will never support anything older.

**NOTE**: If you want to run a public instance, you should set `only_global_lookups=True`
in `website/web/__init__.py` and `bin/async_scrape.py` to disallow scraping of private IPs.

## Installation of Splash

You need a running splash instance, preferably on [docker](https://splash.readthedocs.io/en/stable/install.html)

```bash
sudo apt install docker.io
sudo docker pull scrapinghub/splash
sudo docker run -p 8050:8050 -p 5023:5023 scrapinghub/splash --disable-ui --disable-lua --disable-browser-caches
# On a server with a decent abount of RAM, you may want to run it this way:
# sudo docker run -p 8050:8050 -p 5023:5023 scrapinghub/splash --disable-ui -s 100 --disable-lua -m 50000 --disable-browser-caches
```

## Install redis

```bash
git clone https://github.com/antirez/redis.git
cd redis
git checkout 5.0
make
cd ..
```

## Installation of Lookyloo

```bash
git clone https://github.com/CIRCL/lookyloo.git
cd lookyloo
pipenv install
echo LOOKYLOO_HOME="'`pwd`'" > .env
```

# Run the app

```bash
pipenv run start.py
```

# Run the app in production

## With a reverse proxy (Nginx)

```bash
pip install uwsgi
```

## Config files

You have to configure the two following files:

* `etc/nginx/sites-available/lookyloo`
* `etc/systemd/system/lookyloo.service`

Copy them to the appropriate directories, and run the following command:
```bash
sudo ln -s /etc/nginx/sites-available/lookyloo /etc/nginx/sites-enabled
```

If needed, remove the default site
```bash
sudo rm /etc/nginx/sites-enabled/default
```

Make sure everything is working:

```bash
sudo systemctl start lookyloo
sudo systemctl enable lookyloo
sudo nginx -t
# If it is cool:
sudo service nginx restart
```

And you can open ```http://<IP-or-domain>/```

Now, you should configure [TLS (let's encrypt and so on)](https://www.digitalocean.com/community/tutorials/how-to-secure-nginx-with-let-s-encrypt-on-ubuntu-16-04)

# Use aquarium for a reliable multi-users app

Aquarium is a haproxy + splash bundle that will allow lookyloo to be used by more than one user at once.

The initial version of the project was created by [TeamHG-Memex](https://github.com/TeamHG-Memex/aquarium) but
we have a [dedicated repository](https://github.com/circl/aquarium) that fits our needs better.

Follow [the documentation](https://github.com/CIRCL/aquarium/blob/master/README.rst) if you want to use it.


# Run the app with a simple docker setup

## Dockerfile
The repository includes a [Dockerfile](Dockerfile) for building a containerized instance of the app.

Lookyloo stores the scraped data in /lookyloo/scraped. If you want to persist the scraped data between runs it is sufficient to define a volume for this directory.

## Running a complete setup with Docker Compose
Additionally you can start a complete setup, including the necessary Docker instance of splashy, by using
Docker Compose and the included service definition in [docker-compose.yml](docker-compose.yml) by running

```
docker-compose up
```

After building and startup is complete lookyloo should be available at [http://localhost:5000/](http://localhost:5000/)

If you want to persist the data between different runs uncomment  the "volumes" definition in the last two lines of
[docker-compose.yml](docker-compose.yml) and define a data storage directory in your Docker host system there.