In today's digital age, where platforms like YouTube serve as both entertainment hubs and windows into public sentiment, our project delves into the intricacies of immigration in Ireland. Using cutting-edge NLP models, we dissect sentiments from YouTube comments in near real-time, offering nuanced insights beyond the positive vs. negative spectrum. By annotating Irish-specific immigration videos, our approach equips policymakers and analysts with a deeper understanding of evolving narratives, demonstrating its potential in addressing complex societal challenges in our ever-evolving digital landscape.
-
Install WSL extension in VSCode and then open the project in wsl mode(Short cut : CTRL+SHIFT+P)
-
Verify the docker is using wsl and is enabled in settings
Open the terminal for root folder (Irish-Immigrations-Study-with-ML) in VSCode and type the following commands
- docker-compose build
- docker-compose up
- Sample postman Request for Data Service:
Url : http://localhost:8001/comments
Sample Request JSON: { "url": "https://www.youtube.com/watch?v=bsWZF7g2R-Q", "commentcount": 10, "jobid": "4b0ceaf1-704c-465e-bb6a-93e7d1bf8aow", "modelid": 1 } 2. Preprocessing Service:
Url : http://localhost:8002/api/preprocess
Sample Request JSON: { "jobID": "4b0ceaf1-704c-465e-bb6a-93e7d1bf8aow", "model_id": 1 }
- NLP Service:
Url: http://localhost:8003/api/callmodel
Sample Request JSON: { "jobID": "4b0ceaf1-704c-465e-bb6a-93e7d1bf8aow", "model_id": 1 }
Stop all the containers if running because they are all mapped to their respective ports in the localhost. Then before running the services locally please change the container name to localhost in the URLs used in code to call the other services.
For example the YTComment.py the hostname in post call to preprocessing_script must be changed. e.g. preprocess_url = 'http://localhost:5002/api/preprocess' . Perform these changes in urls used for calls in other services as well.
To check the existing configuration used for running test cases go through main.yml file in .Github/workflows
Please mention the requirements to execute your testcases in requirements-test.txt file