Local Setup
Developers can run Web-BAMT manually on local machine. In order to do this, follow these steps:
Install Web-BAMT from here :
git clone https://github.com/ITMO-NSS-team/Web-BAMT
We highly recommend using virtual enviroment.
Install Python >=3.9, <3.11. Then install everything from requirements.txt:
pip install -r requirements.txt
Install BAMT (special version for web) from here to ../site-packages/bamt directory of virtual
enviroment:
git clone https://github.com/Roman223/bamt_special
Take care of BAMT's config: selbst.ini. Nodes must be a directory where models of logit nodes store (it can be any directory), Log must be absolute path to log file of BAMT. Default log is inside bamt_special repo, called logging.conf.
If npm is not installed, follow instructions .
Move to "frontend" folder and run:
npm config set fetch-retry-mintimeout 2000000
npm config set fetch-retry-maxtimeout 12000000
npm set timeout=10000000
npm install --legacy-peer-deps --loglevel verbose
set FLASK_APP=main.py
Move to Web-BAMT folder and run:
flask run
This run a local backend API. You can either run each server in terminal or run it as back process using nohup:
nohup flask run > log.txt 2>&1 &\.
Note that frontend and backend must be on different ports. You can use flask run --port=xxxx to set port for API.
Open frontend/.env.development and passed the adress of API, it can be found in output of flask run (usually http://localhost:5000).
Move to 'frontend' folder and run:
yarn start
This run a local frontend. You can either run each server in terminal or run it as back process using nohup the same as API.
Go to adress from output of frontend server. There should be an authorization page.
Create and run test.py with query to API:
import requests
BASE = '<YOUR_API>' # usually http://127.0.0.1:5000
assert requests.get(f"{BASE}/api/htest/health").status_code == 200
print("OK")
To use your dataset in the service, you need to submit it in a specific .csv format. Each column must correspond to the features being modeled, and each row must correspond to a sample instance. The dataset should not contain the values nan and infinity. The dataset can have both discrete features (these should be either strings or int values) and continuous features (these should be float values).
Go to app/api/experiment/controller.py and find snippet (52-55):
if case_id == 0:
directory = r"data/hack_processed_with_rf.csv"
elif case_id == 1:
directory = r"data/vk_data.csv"
turn it into something like (don't forget about conditions dataset must meet):
directory = r"data/<YOUR_DATASET.csv>"
Each module has the following files (.py), their content has described in API docs:
- Controller (endpoints, routes, queries)
- Service (core functions)
- Schema (docs schemas)
- Models (table descriptions for database)
