nasim/ProductQuery
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

updates readme

Browse Source
This commit is contained in:
nasim 2024-12-17 21:18:46 +01:00
parent 661e3b390f
commit 8f5f916bfe
1 changed files with 3 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
README.md
View File

@ -42,7 +42,6 @@ This repository contains a **PDF-based product query assistant** built using **L
```bash ```bash
pip install -r requirements.txt pip install -r requirements.txt
``` ```
--- ---
## **Configuration** ## **Configuration**
@ -70,7 +69,6 @@ Replace the placeholder API key or use environment variables for production:
CREATE_FAISS_INDEX = True # To create a new index CREATE_FAISS_INDEX = True # To create a new index
CREATE_FAISS_INDEX = False # To load an existing index CREATE_FAISS_INDEX = False # To load an existing index
``` ```
On initial start set this to TRUE
--- ---
## **Usage** ## **Usage**
@ -91,7 +89,7 @@ Replace the placeholder API key or use environment variables for production:
## **Folder Structure** ## **Folder Structure**
``` ```
xbo-product-query/ CodeChallenge/
│ │
├── app/ ├── app/
│ ├── main.py # Application entry point │ ├── main.py # Application entry point
@ -115,9 +113,8 @@ When scaling up following optimization can be applied:
- Allow appending new documents to the existing FAISS index without rebuilding it entirely. - Allow appending new documents to the existing FAISS index without rebuilding it entirely.
2. **Memory and Disk Management** 2. **Use a Vector Store DB**
- Use a persistent FAISS index stored on disk, which can be loaded as needed. - Use a Vector DB for scalibility
- Enable FAISS's `diskann` mode to load only the necessary portions of the index into memory, reducing RAM consumption.
3. **Batch Processing of Queries** 3. **Batch Processing of Queries**
- Break large queries into smaller batches to improve response times. - Break large queries into smaller batches to improve response times.