Reworked backend with assistant giving json response of all mentioned locations, in place of using other LLM packets like Spacey, and thus lowering size of the backend.
Backend processes have also been refactored and optimized to lower the overall amount of Api request being made, in hopes of both saving the Api some slack, and speeding up the overall prosess of getting data from backend.
The frontend has also been tweeked to accept the new format of data being retrieved by the backend
Changes in backend
Changes how GPT-response and locations are passed through and processed
Changes how search-tasks are constructed and prepared
Resources from Bing are filtered based on confidence-level on accuracy
Function for geometry has been compressed
Geocode-function integrated in main-process
Entities are returned as objects with display_name and coordinates, instead of a loose array of separate tuples
Added more doc-strings to backend, using AutoDocString extention
Added types for better clarity in functions
Some renaming of values for clarification
Changes in frontend
Moved some html elements in MapComponent
Removed BASE_URL from ApiUtils. To add localhost as base url in local setup: add NEXT_PUBLIC_BACKEND_URL=http://localhost:8000 to .env.local file within the frontend folder
Reworked backend with assistant giving json response of all mentioned locations, in place of using other LLM packets like Spacey, and thus lowering size of the backend.
Backend processes have also been refactored and optimized to lower the overall amount of Api request being made, in hopes of both saving the Api some slack, and speeding up the overall prosess of getting data from backend.
The frontend has also been tweeked to accept the new format of data being retrieved by the backend
Changes in backend
Changes in frontend
BASE_URL
from ApiUtils. To add localhost as base url in local setup: addNEXT_PUBLIC_BACKEND_URL=http://localhost:8000
to.env.local
file within the frontend folder