Search Engine (VectorDB)

ادارة الذاكرةThe project is a mini search engine designed to search for extensive and specialized textual information. The concept of the project has been applied to data extracted from programming channels to provide answers to frequently asked questions in the field of programming. The project has been implemented on a Telegram bot, allowing users to ask questions to the bot and receive answers based on reliable and accurate information.

How It Works:

The project consists of two main parts: system operation and information handling.

Information:

YouTube was identified as the best source for valuable programming information. Common programming questions and their answers were extracted and summarized using AI. This data was then compressed and stored in a vector database as patterns using embedding technology.

Search Engine System:

The system searches for information through pattern similarity rather than literal text comparison. Using a vector database, it employs 'cosine distance' to find the closest matching patterns. The question is converted into an embedding and compared with stored data. Search accuracy depends on the quality of stored information and the embedding model. Retrieved information with high similarity is summarized to answer the question. للجوال و تحيليل البيانات الاضافة مميزات اضافية للالعاب