From 5d1d29bb00349fe3fa21fd218c14c5cd2d54b7f0 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Karolina=20Boczo=C5=84?= <karboc3@st.amu.edu.pl>
Date: Thu, 23 Apr 2020 18:05:51 +0000
Subject: [PATCH] xml parser added

---
 backend/xmlParser.py | 32 ++++++++++++++++++++++++++++++++
 1 file changed, 32 insertions(+)
 create mode 100644 backend/xmlParser.py

diff --git a/backend/xmlParser.py b/backend/xmlParser.py
new file mode 100644
index 0000000..276e5fb
--- /dev/null
+++ b/backend/xmlParser.py
@@ -0,0 +1,32 @@
+#!/usr/bin/env python3
+
+import argparse
+from bs4 import BeautifulSoup
+import json
+
+
+# arguments
+parser = argparse.ArgumentParser(description='Process some xml files.')
+parser.add_argument('filename', help='xml forum file')
+args = parser.parse_args()
+
+# make a soup
+with open(args.filename) as forum:
+  soup = BeautifulSoup(forum, "xml")
+
+# put json together
+out = {}
+out['id'] = soup.forum.get('id')
+out['name'] = soup.forum.find('name').text
+out['discussions'] = []
+for d in soup.forum.find_all('discussion'):
+  out['discussions'].append({
+    'id': d.get('id'),
+    'title': d.find('name').text,
+    'posts': [
+      {'id': p.get('id'), 'author': p.userid.text,
+      'message': p.message.get_text()} for p in d.find_all('post')]
+  })
+
+with open('parsed.json', 'w') as outfile:
+  json.dump(out, outfile, ensure_ascii=False, indent=2)