This repository contains a collection of Jupyter notebooks designed to demonstrate and facilitate AI-powered data extraction and processing tasks through Orq.ai. Each notebook serves a specific purpose, from extracting structured data from unstructured sources to generating SQL queries from natural language inputs.
This notebook focuses on extracting structured data from PDF invoices. Key Features:
-
Supports a variety of invoice templates.
-
Provides pre-processing steps for noisy or low-quality PDFs.
-
Outputs data in CSV or JSON format.
This notebook extracts data from images of receipts.
Key Features:
-
Identifies merchant names, transaction dates, and itemized costs.
-
Handles image distortion and varying lighting conditions.
-
Outputs data in user-friendly formats like JSON.
This notebook demonstrates natural language processing (NLP) techniques to classify user intents from text inputs. It is designed for applications like chatbots and customer support automation.
Key Features:
-
Uses pre-trained language models for high accuracy.
-
Supports custom intent categories.
-
Provides a confusion matrix and other metrics to evaluate model performance.
This notebook showcases the capability of transforming natural language queries into executable SQL statements. It is particularly useful for data analysts and non-technical users who need to interact with databases.
Key Features:
-
Supports SQL generation for a wide range of database schemas.
-
Provides a validation mechanism for generated SQL queries.