Invoice Similarity and Feature Extraction

Overview

This project demonstrates how to extract text from PDF invoices, compute similarity between invoices using cosine similarity, and extract specific features from the invoices using regular expressions.

Concepts

1. PDF Text Extraction

Objective: Extract text from PDF files.

• Library Used: PyPDF2

• Function: extractText(pdf_path)

  This function reads a PDF file and extracts its text content. It handles exceptions that might occur during file reading.

2. TF-IDF Vectorization

Objective: Convert text documents into numerical vectors to measure their similarity.

• Library Used: sklearn.feature_extraction.text.TfidfVectorizer

• Function: calculateSimilarity(text1, text2)

  The TfidfVectorizer transforms the text into TF-IDF vectors, which are then used to compute the cosine similarity between two text documents.

3. Cosine Similarity

Objective: Measure the similarity between two vectors.

• Library Used: sklearn.metrics.pairwise.cosine_similarity

• Function: calculateSimilarity(text1, text2)

  Cosine similarity calculates the cosine of the angle between two vectors. It provides a measure of similarity ranging from -1 (completely dissimilar) to 1 (completely similar).

4. Feature Extraction

Objective: Extract specific features from text using regular expressions.

• Library Used: re

• Function: extractFeatures(text)

  This function uses regular expressions to extract specific features from the invoice text, such as invoice number, date, and amount.

5. Invoice Database Management

Objective: Manage a collection of invoices and find the most similar one.

• Class: InvoiceDatabase
• Methods:
  • addInvoice(invoice): Adds an invoice to the database.
  • findMostSimilar(inputInvoice): Finds the most similar invoice to the given input invoice by comparing their text similarity.