act-scio 0.0.47

ACT SCIO

act-scio2

Scio v2 is a reimplementation of Scio in Python3.

Scio uses tika to extract text from documents (PDF, HTML, DOC, etc).

The result is sent to the Scio Analyzer that extracts information using a combination of NLP (Natural Language Processing) and pattern matching.

Changelog

0.0.42

SCIO now supports setting TLP on data upload, to annotatet documents with tlp tag. Documents downloaded by feeds will have a default TLP white, but this can be changed in the config for feeds.

Source code

The source code the workers are available on github.

Setup

To setup, first install from PyPi:

sudo pip3 install act-scio

You will also need to install beanstalkd. On debian/ubuntu you can run:

sudo apt install beanstalkd

Configure beanstalk to accept larger payloads with the -z option. For red hat derived setups this can be configured in /etc/sysconfig/beanstalkd:

MAX_JOB_SIZE=-z 524288

You then need to install NLTK data files. A helper utility to do this is included:

scio-nltk-download

You will also need to create a default configuration:

scio-config user

API

To run the api, execute:

scio-api

This will setup the API on 127.0.0.1:3000. Use --port <PORT> and --host <IP> to listen on another port and/or another interface.

Configuration

You can create a default configuration using this command (should be run as the user running scio):

scio-config user

Common configuration can be found under ~/.config/scio/etc/scio.ini

Running Manually

Scio Tika Server

The Scio Tika server reads jobs from the beanstalk tube scio_doc and the extracted text will be sent to the tube scio_analyze.

The first time the server runs, it will download tika using maven. It will use a proxy if $https_proxy is set.

scio-tika-server

scio-tika-server uses tika-python which depends on tika-server.jar. If your server has internet access, this will downloaded automatically. If not or you need proxy to connect to the internet, follow the instructions on "Airagap Environment Setup" here: https://github.com/chrismattmann/tika-python. Currently only tested with tika-server version 1.24.1.

Scio Analyze Server

Scio Analyze Server reads (by default) jobs from the beanstalk tube scio_analyze.

scio-analyze

You can also read directly from stdin like this:

echo "The companies in the Bus; Finanical, Aviation and Automobile industry are large." | scio-analyze --beanstalk=

Running as a service

Systemd compatible service scripts can be found under examples/systemd.

To install:

sudo cp examples/systemd/*.service /usr/lib/systemd/system
sudo systemctl enable scio-tika-server
sudo systemctl enable scio-analyze
sudo service start scio-tika-server
sudo service start scio-analyze

Local development

Use pip to install in local development mode. act-scio uses namespacing, so it is not compatible with using setup.py install or setup.py develop.

In repository, run:

pip3 install --user -e .