https://github.com/zeionara/cold
An experimental tool for improving recommender systems by considering more data about source annotations
https://github.com/zeionara/cold
Last synced: 4 months ago
JSON representation
An experimental tool for improving recommender systems by considering more data about source annotations
- Host: GitHub
- URL: https://github.com/zeionara/cold
- Owner: zeionara
- License: apache-2.0
- Created: 2023-04-18T17:42:48.000Z (about 2 years ago)
- Default Branch: master
- Last Pushed: 2023-06-19T21:51:14.000Z (almost 2 years ago)
- Last Synced: 2025-01-02T07:46:10.039Z (6 months ago)
- Language: Python
- Size: 183 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Cold
**Co**mmunity **l**abelling **d**river - a tool for inferring labels for data elements using provided corpus of annotations. In particular, the tools attempts to solve the *cold*-start problem of the recommender systems.
## Installation
To install the project dependencies, create a conda environment using the provided config file:
```sh
conda env create -f environment.yml
```
The following env variables are required for working with `vk api`:
- `COLD_APP_ID` - api app id, see at the [vk api web page](https://vk.com/apps?act=manage);
- `COLD_VK_API_KEY` - api access token, use [this script](tools/make-token-generation-url.sh) to obtain it;
- `COLD_USER_ID` - user id, from whose profile automated queries will be sent, see property `user_id` filled after running the [token generation script](tools/make-token-generation-url.sh).
## Launch
To run the project in a manual mode, execute the following command:
```sh
python -m cold probe assets/default/train.cld assets/default/spec.yml
```
## Test
To run tests, use the following command:
```sh
python -m unittest discover test
```