README.md

---
configs:
- config_name: Metadata
  data_files: pm_info_table.csv
- config_name: Source Data
  data_files: pm_source_text.csv
task_categories:
- text-classification
- question-answering
- sentence-similarity
- token-classification
language:
- ta
pretty_name: Project Madurai
size_categories:
- n<1K
license: apache-2.0
---

<h1>Project Madurai Books Text Dataset</h1>

This dataset card aims to convert the Tamil books available on the Project Madurai  website to the HF dataset. It has been scrapped from [Project Madurai Website](https://www.projectmadurai.org/pmworks.html).


## Dataset Details

- You can see a table above called "Meta Data", which is just an info table.
<b>You can't able to preview the "Source Data" table</b>, due to it being about 300MB.
[Don't open the Dataset in Excel It will lead to a crash of the OS instead open it using Python in pandas or datasets]

- <b>"Project Madurai"</b> was a great initiative in 2008, to make the Tamil Language available to everyone on the internet.<br>
- They converted almost <b>1006 books</b> to the format of PDF and HTML pages.<br>
- By utilizing the website this dataset is created .<br>
- The dataset contains more genres of books from Tamil <b>Literature to Fiction</b> which can be very useful for developing an NLP model for Tamil.

## Dataset Structure

#### First Table Structure
- This Table Contains the Meta Data of the Dataset Which Tells the <b>Index</b>, <b>Work.No</b>[no of the book],<b>Author</b> of the Book,<b>Title</b> of the book, <b>Unicode</b>[The Html Page name which the source of the book is available].
![First_Table](first_table_img.png)
- This Table can be used to find the source of the data in the "Source Data"[second table].
- In one cell there can be 2 or more book id.

#### Second Table Structure
- You can download the second table from "Files and Versions". This can't be preview due to the size.
- This Table Contains the Source Data of the Dataset Which Tells the <b>Index</b>, <b>Book_Id</b>[HTML page name],<b>Link</b> of the Book,<b>tamil_text</b> of the book.
![Second_table](second_table_img.png)
- In this Table, each row contains a single book or half of the book[due to one book in different html pages].

- The 'tamil_text' column contains more than <b>95% of Tamil text, and others are Roman letters, numerals, and english text</b>. While scarping it the Tamil text and english text both in single tag, it can be cleaned in future.

- The 'tamil_text' column may contain <b>delimeters</b> like <b>['',.?!/n-]</b> to process the data easier.

- The Table contains more than 1006 rows due to some books split into different parts.

- You can find the book author's name in the first table using the Book_ID

### Special Thanks
- Project Madurai Team.