From 7d925a0ada749a8d450639bfab21454590f3b6b9 Mon Sep 17 00:00:00 2001
From: Phil Harle <philharle@gmail.com>
Date: Fri, 24 Apr 2020 14:41:00 +0100
Subject: [PATCH] Add Newcastle City Council

---
 outputs/NewcastleCityCouncil.json | 16 +++++++++++
 scripts/NewcastleCityCouncil.py   | 46 +++++++++++++++++++++++++++++++
 2 files changed, 62 insertions(+)
 create mode 100644 outputs/NewcastleCityCouncil.json
 create mode 100644 scripts/NewcastleCityCouncil.py

diff --git a/outputs/NewcastleCityCouncil.json b/outputs/NewcastleCityCouncil.json
new file mode 100644
index 0000000000..da2a1e3124
--- /dev/null
+++ b/outputs/NewcastleCityCouncil.json
@@ -0,0 +1,16 @@
+{
+    "bins": [
+        {
+            "BinType": "Domestic Waste",
+            "NextCollectionDate": "2020-05-05"
+        },
+        {
+            "BinType": "Recycling",
+            "NextCollectionDate": "2020-04-28"
+        },
+        {
+            "BinType": "Garden Waste",
+            "NextCollectionDate": "2020-04-24"
+        }
+    ]
+}
\ No newline at end of file
diff --git a/scripts/NewcastleCityCouncil.py b/scripts/NewcastleCityCouncil.py
new file mode 100644
index 0000000000..d24f6774f7
--- /dev/null
+++ b/scripts/NewcastleCityCouncil.py
@@ -0,0 +1,46 @@
+#!/usr/bin/env python3
+from urllib.request import Request, urlopen
+import json
+from bs4 import BeautifulSoup
+from datetime import datetime
+
+user_agent = 'Mozilla/5.0 (Windows NT 6.1; Win64; x64)'
+headers = {'User-Agent': user_agent}
+
+#Replace URL
+req = Request('https://community.newcastle.gov.uk/my-neighbourhood/ajax/getBinsNew.php?uprn=XXXXXXXXXXXX')
+req.add_header('User-Agent', user_agent)
+
+fp = urlopen(req).read()
+page = fp.decode("utf8")
+
+soup = BeautifulSoup(page, features="html.parser")
+soup.prettify()
+
+#Form a JSON wrapper
+data = {"bins":[]}
+
+#Loops the strong elements
+for element in soup.find_all("strong"):
+    #Domestic Waste is formatted differenty to other bins
+    if "Green Bin (Domestic Waste) details:" in str(element):
+        collectionInfo = element.next_sibling.find('br').next_element
+    else:
+        collectionInfo = element.next_sibling.next_sibling.next_sibling.next_sibling
+    
+    binType = str(element)[str(element).find("(")+1:str(element).find(")")]
+    collectionDate = str(datetime.strptime(str(collectionInfo).replace('Next collection : ',''), '%d-%b-%Y').date())
+    
+    dict_data = {
+        "BinType": binType,
+        "NextCollectionDate": collectionDate
+        }
+
+    #Add data to the main JSON Wrapper
+    data["bins"].append(dict_data)
+
+
+##Make the JSON
+json_data = json.dumps(data,sort_keys=True, indent=4)
+
+print(json_data)
\ No newline at end of file