No project description provided
Project description
Comment Recommendation Framework
The Comment Recommendation Framework is a modular approach to support scientists in the development of prototypes for comment recommendation systems that can be used in real-world scenarios. The advantage of such a system is that it relieves the scientist from the majority of the technical code and only prototype-specific components have to be developed. In this way, the researchers can invest more time in the development of recommendation models and less time has to be spent in the development of a prototype while at the same time getting prototypes that can be used in real-world settings.
Setup
Ensure that the following tools are installed:
- Docker
- Docker-Compose
- Pyhon >= 3.10
Documentation
To build the latest version of the documentation, please run in the docs folder:
$ make clean && make html
Then you find the latest documentation here
Environment Variables
The framework need some environment variables to be set for running properly. Please ensure that you have an .env
file with the following variables:
- NEO4J_PASSWORD
- NEO4J_BOLT_URL (Format:
bolt://neo4j:<NEO4J_PASSWORD>@neo4j:7687
)
Run the Comment Recommendation Framework
Install the package locally
To install the package locally, you have to build it first. For this, run in the folder with setup.py
:
python3 -m build
Please make sure that the build library is installed. Otherwise, you cannot build the package.
This creates a dist
folder at your current location with two files Comment-Recommendation-Framework-X.X.X.tar.gz
and
Comment_Recommendation_Framework-X.X.X-py3-none-any.whl
. The tar.gz
file is the
source distribution and the .whl
is the built distribution.
We recommend to create a virtual environment to isolate your project from the rest of your system to prevent import and version problems.
Then you run inside your virtual environment:
pip install <path_to_the_whl_file>
Create your project
To create the system template you run the following command in your virtual env after you have installed the package:
python3 -m comment_recommendation_framework
Then the package asks you different questions to determine which modules you need for your project. You can answer them
with y
for yes and n
for no.
Run different moduls with docker-compose
We provide you with the following docker-compose
files to run the different components of the recommendation framework.
docker-compose.scraping.yml
: Runs the news agency scraper to retrieve articles and comments from various news agencies.docker-compose.embed.yml
: Starts the embedding process to compute the embeddings for the comments and articles. Should be run directly afterdocker-compose.scraping.yml
.docker-compose.csv.yml
: Imports comments and articles from a csv file into the database.docker-compose.test.yml
: Runs the tests for the system.docker-compose.api.yml
: Runs the comment-recommendation systems.
User-Interface
If you would like to use the default user interface. You have to install the npm packages and build the chrome extension.
For this you have to run in the UI
folder:
$ npm install
and afterwards:
$ npm run build
Then you can import the build
folder in a chromium browser.
Maintainers:
- Jan Steimann
Contributors:
- Jan Steimann
License:
Copyright(c) 2024 - today Jan Steimann
Distributed under the MIT License
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file comment_recommendation_framework-0.16.6-py3-none-any.whl
.
File metadata
- Download URL: comment_recommendation_framework-0.16.6-py3-none-any.whl
- Upload date:
- Size: 444.9 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? Yes
- Uploaded via: twine/4.0.2 CPython/3.11.7
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 8b8a3c1ef429208082ef6f09d352637c1252183cf15346af7817f6dfec9039f8 |
|
MD5 | 411dfb501e7b30b57e232ae664c20599 |
|
BLAKE2b-256 | 040b934eb559f25a03f721af48b137873c3e9aa3acfab2db87cdfeb2efd79de4 |
File details
Details for the file
comment-recommendation-framework-0.16.6.tar.gz
.File metadata
File hashes
da71d73e3704c42825159dfe0e2fa1a7ac7f74d5da41c4459912fe2da18b4a39
a925f3f739dd791341221e02cd79356b
b06ef763feb7e6b9dc86bc7fbb98e1b74fde6353ca8f14744a5b2512a718582c
See more details on using hashes here.