Mlynarchik Artyom

README.md

@@ -0,0 +1,24 @@
+# Pure Python RSS Reader [PythonHomework]
+
+Python version: v3.8
+
+Current version: v0.2
+
+Code checking: Code correspond to pep8
+#### Usage:
+```shell
+usage: __main__.py [-h] [--version] [--json] [--verbose] [--limit LIMIT] source
+
+Pure Python command-line RSS reader.
+
+positional arguments:
+  source         RSS URL
+
+optional arguments:
+  -h, --help     show this help message and exit
+  --version      Print version info
+  --json         Print result as JSON in stdout
+  --verbose      Outputs verbose status messages
+  --limit LIMIT  Limit news topics if this parameter provided
+  ```
+  JSON scheme is described in `json_schema.json`

json_schema.json

@@ -0,0 +1,31 @@
+{
+    "$schema": "http://json-schema.org/schema#",
+    "title": "feed",
+    "type": "object",
+    "required": ["title", "date", "text", "link", "hrefs"],
+    "properties": {
+      "title": {
+        "type": "string",
+        "description": "Article title"
+      },
+      "date": {
+        "type": "date",
+        "description": "Article published date"
+      },
+      "text": {
+        "type": "string",
+        "description": "Article text"
+      },
+      "link": {
+        "type": "string",
+        "description": "Article static link"
+      },
+      "hrefs": {
+        "type": "array",
+        "items": {
+          "type": "string",
+          "description": "Article href"
+        }
+      }
+    }
+  }

rss_reader/__main__.py

@@ -0,0 +1,34 @@
+import argparse
+import logging
+
+from rss_reader import Reader
+
+
+def main():
+    args = parse_args()
+
+    if args.verbose:
+        logging.basicConfig(format="%(levelname)s: %(message)s", level=logging.INFO)
+    else:
+        logging.basicConfig(format="%(levelname)s: %(message)s")
+
+    reader = Reader(args.source, args.limit, args.json)
+    reader.parse_url()
+
+    reader.print_articles()
+
+
+def parse_args():
+    parser = argparse.ArgumentParser(description='Pure Python command-line RSS reader.')
+
+    parser.add_argument('source', help='RSS URL')
+    parser.add_argument('--version', help='Print version info', action='version', version='%(prog)s 0.2')
+    parser.add_argument('--json', help='Print result as JSON in stdout', action='store_true')
+    parser.add_argument('--verbose', help='Outputs verbose status messages', action='store_true')
+    parser.add_argument('--limit', help='Limit news topics if this parameter provided', type=int)
+
+    return parser.parse_args()
+
+
+if __name__ == '__main__':
+    main()

rss_reader/article.py

@@ -0,0 +1,26 @@
+"""Article module"""
+
+import re
+import time
+import logging
+
+
+class Article:
+    logger = logging.getLogger('__main__.py')
+
+    def __init__(self, title, date, text, link, hrefs):
+        self.title = title
+        self.date = date
+        self.text = self.strip_html_string(text)
+        self.link = link.split('?')[0]
+        self.hrefs = hrefs
+
+    def convert_time_to_unix(self):
+        """Convert datetime to unix time"""
+        pattern_time = time.strptime(self.date, '%a, %d %b %Y %H:%M:%S %z')
+        return int(time.mktime(pattern_time))
+
+    def strip_html_string(self, string):
+        """Remove html tags from a string"""
+        strip_string = re.compile('<.*?>')
+        return re.sub(strip_string, '', string)

rss_reader/json_format.py

@@ -0,0 +1,41 @@
+"""Module with tools for working with Json"""
+
+import json
+import codecs
+import logging
+
+
+class Json:
+    logger = logging.getLogger('__main__.py')
+
+    def __init__(self):
+        self.data = {}
+
+    def __str__(self):
+        """Print JSON-file to console"""
+        self.logger.info('Print JSON-data to console')
+
+        return json.dumps(self.data, ensure_ascii=False, indent=4)
+
+    # def print(self):
+    #     """Print JSON-file to console"""
+    #     self.logger.info('Print JSON-data to file')
+
+        print(json.dumps(self.data, ensure_ascii=False, indent=4))
+
+    def write_to_file(self):
+        """Write JSON-data to file"""
+        self.logger.info('Write JSON-data to file')
+
+        with codecs.open('data.json', 'w', encoding='utf-8') as outfile:
+            json.dump(self.data, outfile, ensure_ascii=False, indent=4)
+
+    def format(self, data):
+        """Format file to JSON-format"""
+        self.logger.info('Format data to JSON appereance')
+
+        self.data = {}
+        self.data['feed'] = []
+
+        for element in data:
+            self.data['feed'].append(element)

rss_reader/rss_reader.py

@@ -0,0 +1,85 @@
+"""RSS-reader module"""
+
+import feedparser
+import logging
+
+from article import Article
+from json_format import Json
+
+
+class Reader:
+    logger = logging.getLogger('__main__.py')
+
+    def __init__(self, link, limit, json):
+        self.link = link
+        self.limit = limit
+        self.articles = []
+        self.json = json
+        self.hrefs = []
+
+    def parse_url(self):
+        """Get RSS xml-file from url"""
+        self.logger.info('Get RSS XML-file from url')
+
+        self.feed = feedparser.parse(self.link)
+        self.parse_xml(self.feed.entries[:self.limit])
+
+    def parse_xml(self, source):
+        """Parse xml-file to articles"""
+        self.logger.info('Parse XML-file to articles')
+
+        for item in source:
+            content = []
+
+            try:
+                for element in item.media_content:
+                    content.append(element['url'])
+            except AttributeError:
+                try:
+                    for element in item.media_thumbnail:
+                        content.append(element['url'])
+                except AttributeError:
+                    content.append('No content!')
+                # content.append('No content!')
+
+            self.articles.append(Article(item.title, item.published, item.description, item.link, content))
+
+        if self.json is True:
+            json_object = Json()
+            feeds = self.articles_to_array()
+            json_object.format(feeds)
+            print(json_object)
+
+    def articles_to_array(self):
+        self.logger.info('Convert articles to array of dicts')
+
+        array = []
+        for article in self.articles:
+            feed_dict = {}
+            feed_dict.update({'title': article.title})
+            feed_dict.update({'date': article.date})
+            feed_dict.update({'text': article.text})
+            feed_dict.update({'link': article.link})
+            feed_dict.update({'hrefs': article.hrefs})
+            array.append(feed_dict)
+
+        return array
+
+    def print_articles(self):
+        self.logger.info('Print articles to console')
+
+        for article in self.articles:
+            self.print_article(article)
+            print('\n-------------------------\n')
+
+    def print_article(self, article):
+        """Print article to console"""
+
+        print(f'Title: {article.title}')
+        print(f'Date:  {article.date}')
+        print(f'Link: {article.link}')
+        print('\nArticle text:')
+        print(article.text)
+        print('\nHrefs:')
+        for href in article.hrefs:
+            print(href)

setup.py

@@ -0,0 +1,19 @@
+import setuptools
+
+setuptools.setup(
+    name="rss-reader",
+    version="0.2",
+    author="Archeex",
+    author_email="qsanich@gmail.com",
+    description="Pure Python command-line RSS reader",
+    packages=setuptools.find_packages(),
+    classifiers=[
+        "Programming Language :: Python :: 3",
+        "License :: OSI Approved :: MIT License",
+        "Operating System :: OS Independent",
+    ],
+    entry_points={
+        'console_scripts': ['rss_reader = rss_reader.__main__:main']
+    },
+    python_requires='>=3.8'
+)