Building a PDF Reader in 500 Lines of Python with PyQt and SQLite

After two weeks of iteration, the PyReadon PDF reader implements core features: adding and deleting books, reading books, saving reading progress, and viewing book metadata. This version optimizes the book data structure, supports simultaneous reading of multiple books, adds left-click page turning, and uses SQLite to persist the library and per-book information such as file path and current page.

Book Class Design

The Book class encapsulates book metadata, page count, and reading status. The __eq__ special method compares two Book instances by their file name ( fname) to determine if they represent the same book.

class Book:
    def __init__(self, fname):
        # File name
        self.fname = fname
        # Whether the book has been read
        self.flag = None
        self._info = None
        self._page = 0
        self.get_meta_data(self.fname)

    def __eq__(self, other):
        if hasattr(other, 'fname'):
            return self.fname == other.fname
        return False

Simultaneous Reading of Multiple Books

An internal read_list stores Book objects currently being read. Each book retains its own page attribute so the program remembers the last page for every book. On startup, read_list is initialized with a placeholder None followed by all books whose flag indicates they were being read when the app last closed.

self.read_list = [None]
self.read_list.extend(book for book in self.booklist if book.flag)

Left-Click Page Turning

The MyArea widget overrides mousePressEvent. The mouse's x-coordinate ( event.pos().x()) is compared against the widget width: clicks in the left third trigger self.left() (previous page), clicks in the right third trigger self.right() (next page).

# Mouse left-click page turning
def mousePressEvent(self, event):
    pos = event.pos().x()
    width = self.size().width()
    if event.button() == Qt.LeftButton:
        if pos > width * 2 / 3:
            self.right()
        elif pos < width / 3:
            self.left()

SQLite Persistence

SQLite is chosen for its serverless, zero-configuration, and fast characteristics. On launch, read_db() reads book records from PDF.db:

If PDF.db does not exist, create it and a book_info table with columns path, page, flag.

Query all rows, construct a namedtuple for each, instantiate a Book object, restore its page and flag, and yield the book.

book_db = 'PDF.db'
book_info = namedtuple('info', 'path page flag')

def read_db():
    os.chdir(os.path.dirname(os.path.realpath(__file__)))
    if not os.path.exists(book_db):
        conn = sqlite3.connect(book_db)
        conn.execute("CREATE TABLE book_info(path, page, flag)")
        conn.close()
    conn = sqlite3.connect(book_db)
    for row in conn.execute('SELECT * FROM book_info'):
        info = book_info(*row)
        book = Book(info.path)
        book.page = info.page
        book.flag = info.flag
        yield book
    conn.close()

Saving the library state uses save2db(), which first clears the table via remove_db() then bulk-inserts current book data:

def save2db(booklist):
    conn = sqlite3.connect(book_db)
    conn.executemany("INSERT INTO book_info Values (?,?,?)",
        ((book.fname, book.page, book.flag) for book in booklist))
    conn.commit()
    conn.close()

def remove_db():
    conn = sqlite3.connect(book_db)
    conn.execute('DELETE FROM book_info')
    conn.commit()
    conn.close()

Viewing Book Metadata

A right-click context menu triggers QMessageBox.about() to display detailed metadata extracted from the PDF: path, format, title, author, creator, and producer.

elif action == item3:
    index = row_num * 8 + col_num
    book = self.booklist[index]
    info = book.info
    fmt = f'Path: {info.path}

' \
        f'Format: {info.format}

' \
        f'Title: {info.title}

' \
        f'Author: {info.author}

' \
        f'Creator: {info.creator}

' \
        f'Producer: {info.producer}'
    QMessageBox.about(self, 'Document Info', fmt)

About Dialog via Qt Designer

An Info dialog designed in Qt Designer is bound to the main window's info toolbar action:

info = Info()
reader.infobar.triggered.connect(info.show)

The complete source code is available at https://github.com/cassieeric/Python-Application.