Began extracting data from epub files

README.md

@@ -1,7 +1,7 @@
 # pyShelf
 A simple terminal based ebook server
 
-Frusterated with Calibre being my only option for hosting my eBook collection, I have decided to spin up my own.
+Frustrated with Calibre being my only option for hosting my eBook collection, I have decided to spin up my own.
 
 Calibre is a great organizational tool for your books, however not having a terminal based option for running and maintaining
 a server is cumbersome when running on older equipment. Thus I am creating pyShelf and I hope to be able to provide all

config.py

@@ -3,9 +3,11 @@ class Config:
    def __init__(self):
       self.book_path = "books/"
       self.book_shelf = "data/shelf.json"
+      self.catalogue_db = "data/catalogue.db"
       self.file_array = [
          "data/catalogue.json",
-         "data/shelf.json",
+         self.book_shelf,
+         self.catalogue_db,
          "conf/settings.json"
          ]
       self.auto_scan = True

lib/library.py

@@ -0,0 +1,96 @@
+#!/usr/bin/python
+import json
+import os
+import re
+import zipfile
+from bs4 import BeautifulSoup
+from config import Config
+config = Config()
+
+
+class Catalogue:
+    """Decodes and stores book information"""
+    """Step One: filter_books"""
+    def __init__(self):
+        self.file_list = []
+        with open(config.book_shelf, 'r') as f:
+            try:
+                self.catalogue = json.load(f)
+                self.current_files = self.scan_folder()
+            except Exception:
+                self.filter_books()
+
+    def scan_folder(self, folder=config.book_path):
+        for f in os.listdir(folder):
+            _path = os.path.abspath(folder+'/'+f)
+            #_path = os.path.abspath('.')+'/'+folder+f+'/'
+            _is_dir = os.path.isdir(_path.strip()+'/')
+            if _is_dir:
+                self.file_list.append(self.scan_folder(_path))
+            self.file_list.append(_path)
+
+    def scan_book(self, book):
+        """REMOVE ME?"""
+        _epub = zipfile.ZipFile(book)
+        with _epub as _epub_open:
+            try: _epub_open.open('content.opf'); return True
+            except Exception as e: print(e); return False
+
+    def filter_books(self):
+        """
+        Scan book folder recursively for epub files
+        filter_books(0) -> Catalogue.books
+        filter_books(1) -> self.books[]
+        :param ret: 0 -> create class property -> dump json
+        :param ret: 1 -> create & return class property
+        """
+        self.scan_folder()
+        regx = re.compile(r"\.epub")
+        self.books = list(filter(regx.search, filter(None, self.file_list)))
+        _book_list_expanded = {}
+        with open(config.book_shelf, 'w') as f:
+            for book in self.books:
+                _book_list_expanded[book] = self.process_book(book)
+            json.dump(_book_list_expanded, f)
+        return _book_list_expanded
+
+    def process_book(self, book):
+        """Return dictionary of epub file contents"""
+        f_name = 'content.opf'
+        book = zipfile.ZipFile(book, 'r')
+        details = {}
+        with book as book_zip:
+            details['files'] = []
+            details['path'] = book.filename
+            expanded = book_zip.infolist()
+            regx = re.compile(r'\.opf|cover')
+            for i in expanded:
+                match = re.search(regx, i.filename)
+                if match:
+                    # Returns zip file location of requested files
+                    details['files'].append(match.string)
+        return details
+
+    def extract_metadata(self, book):
+        """
+        Return extracted metadata and cover picture
+        book['path'] == Full path to ebook file
+        book['files'] == list of files from self.process_book(book)
+        """
+        book_zip = zipfile.ZipFile(book['path'], 'r')
+        opf_regx, cover_regx = re.compile(r'\.opf'), re.compile(r'\.jpg|\.png|\.bmp|\.gif')
+        with book_zip as f:
+            content = list(filter(opf_regx.search, book['files']))
+            content = book_zip.open(content[0])
+            soup = BeautifulSoup(content, "xml")
+            title = soup.find("dc:title")
+            author = soup.find("dc:creator")
+            cover = soup.find("meta", attrs={"name" : "cover"})
+        return title
+    def compare_shelf_current(self):
+        try:
+            self.books
+        except Exception:
+            self.filter_books()
+        unique = set(self.books) - set(self.catalogue)
+        return unique

lib/pyShelf.py

@@ -0,0 +1,33 @@
+#!/usr/bin/python
+import os
+import zipfile
+from config import Config
+config = Config()
+
+class InitFiles:
+    """First run file creation operations"""
+    def __init__(self, file_array):
+        print("Begining creation of file structure")
+        for _pointer in file_array:
+            if not os.path.isfile(_pointer):
+                self.CreateFile(_pointer)
+
+    def CreateFile(self, _pointer):
+        """Create the file"""
+        if not os.path.isdir(os.path.split(_pointer)[0]):
+            os.mkdir(os.path.split(_pointer)[0])
+        f = open(_pointer, "w+")
+        f.close()
+
+
+class Epub:
+    """All Epub file handling"""
+    def __init__(self):
+        global config
+        self.book_path = config.book_path
+
+    def import_book(self):
+        pass
+
+    def book_list(self):
+        pass

lib/storage.py

@@ -0,0 +1,48 @@
+#!/usr/bin/python
+import sys
+import sqlite3
+sys.path.insert(1,'../')
+from config import Config
+db_pointer = Config().catalogue_db
+
+
+class Storage:
+    """Contains all methods for system storage"""
+
+    def __init__(self):
+        self.db_file = db_pointer
+        self.database()
+
+    def database(self):
+        """Create database cursor"""
+        try:
+            self.db = sqlite3.connect(self.db_file)
+            self.cursor = self.db.cursor()
+            return True
+        except Exception as e:
+            return False
+
+    def create_tables(self):
+        """Create table structure"""
+        q_check = "SELECT * FROM books"
+        q_create = '''CREATE TABLE books(title text, author text,
+        categories text, cover blob,  pages int, progress int,
+        file_name text)'''
+        try:
+            self.cursor.execute(q_check)
+        except Exception as e:
+            self.cursor.execute(q_create)
+
+    def insert_book(self, book):
+        """
+        Insert book in database
+        :returns: True if succeeds False if not
+        """
+        q = '''INSERT INTO books (title, author, categories, cover,
+        pages, progress, file_name) values (%s, %s, %s, %s, 0, %s)''' % ()
+        try:
+            self.cursor.execute(q)
+            return True
+        except Exception as e:
+            print(e)
+            return False

main.py

@@ -1,5 +1,7 @@
 #!/usr/bin/python
 # import zipfile as Zip
+import sys
+sys.path.insert(1, 'lib/')
 from pyShelf import InitFiles
 from config import Config
 from library import Catalogue
@@ -11,5 +13,7 @@ InitFiles(config.file_array)
 Catalogue = Catalogue()
 # Filter Your books
 # This only needs to be run on first run, & when new books are added
-book_list = Catalogue.filter_books(0)
-
+book_list = Catalogue.filter_books()
+for book in book_list:
+    extracted = Catalogue.extract_metadata(book_list[book])
+    print(extracted)

tests/test_library.py

@@ -1,14 +1,16 @@
 import unittest
+import sys
+sys.path.insert(1, 'lib/')
 from library import Catalogue
 
 
-class Testing(unittest.TestCase):
+class LibraryTest(unittest.TestCase):
 
     def test_libray_catalogue(self):
         self.assertIsNotNone(Catalogue())
 
     def test_library_catalogue_filter_books(self):
-        self.assertIsNotNone(Catalogue().filter_books(1))
+        self.assertIsNotNone(Catalogue().filter_books())
 
 if __name__ == '__main__':
     unittest.main()

tests/test_storage.py

@@ -0,0 +1,16 @@
+import unittest
+import sys
+sys.path.insert(1, 'lib/')
+from storage import Storage
+
+storage = Storage()
+
+class StorageTest(unittest.TestCase):
+
+    def test_Storage_databasee(self):
+        self.assertTrue(storage.database())
+
+    def test_Storage_create_tables(self):
+        self.assertIsNot(storage.create_tables(), Exception)
+if __name__ == '__main__':
+    unittest.main()

tests/test_sysio.py

@@ -1,6 +1,8 @@
 import unittest
 import os
 import shutil
+import sys
+sys.path.insert(1, 'lib/')
 from pyShelf import InitFiles
 from pyShelf import Epub