uk_bin_collection 0.85.0__tar.gz → 0.85.2__tar.gz

{uk_bin_collection-0.85.0 → uk_bin_collection-0.85.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: uk_bin_collection
-Version: 0.85.0
+Version: 0.85.2
 Summary: Python Lib to collect UK Bin Data
 Author: Robert Bradley
 Author-email: robbrad182@gmail.com

{uk_bin_collection-0.85.0 → uk_bin_collection-0.85.2}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "uk_bin_collection"
-version = "0.85.0"
+version = "0.85.2"
 description = "Python Lib to collect UK Bin Data"
 readme = "README.md"
 authors = ["Robert Bradley <robbrad182@gmail.com>"]

{uk_bin_collection-0.85.0 → uk_bin_collection-0.85.2}/uk_bin_collection/tests/input.json

@@ -68,6 +68,13 @@
         "wiki_name": "Bedfordshire Council",
         "wiki_note": "In order to use this parser, you must provide a valid postcode and a uprn retrieved from the councils website for your specific address"
     },
+    "BelfastCityCouncil": {
+        "skip_get_url": true,
+        "url": "https://online.belfastcity.gov.uk/find-bin-collection-day/Default.aspx",
+        "wiki_name": "BelfastCityCouncil",
+        "postcode": "BT10 0GY",
+        "uprn": "185086469"
+    },
     "BexleyCouncil": {
         "house_number": "1 Dorchester Avenue, Bexley",
         "postcode": "DA5 3AH",
@@ -415,8 +422,8 @@
         "wiki_note": "Use [this site](https://www.gbcbincalendars.co.uk/) to find the collections for your address. Use the -n parameter to add them in a comma-separated list inside quotes, such as: 'Friday G4, Friday J'."
     },
     "GlasgowCityCouncil": {
-        "url": "https://www.glasgow.gov.uk/forms/refuseandrecyclingcalendar/PrintCalendar.aspx?UPRN=906700034497",
-        "wiki_command_url_override": "https://www.glasgow.gov.uk/forms/refuseandrecyclingcalendar/PrintCalendar.aspx?UPRN=XXXXXXXX",
+        "url": "https://onlineservices.glasgow.gov.uk/forms/RefuseAndRecyclingWebApplication/CollectionsCalendar.aspx?UPRN=906700034497",
+        "wiki_command_url_override": "https://onlineservices.glasgow.gov.uk/forms/RefuseAndRecyclingWebApplication/CollectionsCalendar.aspx?UPRN=XXXXXXXX",
         "wiki_name": "Glasgow City Council",
         "wiki_note": "Replace XXXXXXXX with UPRN."
     },

uk_bin_collection-0.85.2/uk_bin_collection/uk_bin_collection/councils/BelfastCityCouncil.py

@@ -0,0 +1,100 @@
+import logging
+from datetime import datetime
+
+import requests
+import urllib
+
+from bs4 import BeautifulSoup
+from uk_bin_collection.uk_bin_collection.common import *
+from uk_bin_collection.uk_bin_collection.get_bin_data import AbstractGetBinDataClass
+
+
+
+# import the wonderful Beautiful Soup and the URL grabber
+class CouncilClass(AbstractGetBinDataClass):
+    """
+    Concrete classes have to implement all abstract operations of the
+    base class. They can also override some operations with a default
+    implementation.
+    """
+
+    def get_session_variable(self, soup, id) -> str:
+        """Extract ASP.NET variable from the HTML."""
+        element = soup.find("input", {"id": id})
+        if element:
+            return element.get("value")
+        else:
+            raise ValueError(f"Unable to find element with id: {id}")
+
+    def parse_data(self, page: str, **kwargs) -> dict:
+        bin_data = {"bins": []}
+        headers = {
+            "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:129.0) Gecko/20100101 Firefox/119.0"
+        }
+
+        session = requests.Session()
+        session.headers.update(headers)
+        
+        user_uprn = kwargs.get("uprn")
+        user_postcode = kwargs.get("postcode")
+        URL = "https://online.belfastcity.gov.uk/find-bin-collection-day/Default.aspx"
+
+        # Build initial ASP.NET variables for Postcode Find address
+        response = session.get(URL)
+        response.raise_for_status()
+        soup = BeautifulSoup(response.text, "html.parser")
+        form_data = {
+            "__EVENTTARGET": "",
+            "__EVENTARGUMENT": "",
+            "__VIEWSTATE": self.get_session_variable(soup, "__VIEWSTATE"),
+            "__VIEWSTATEGENERATOR": self.get_session_variable(soup, "__VIEWSTATEGENERATOR"),
+            "__SCROLLPOSITIONX": "0",
+            "__SCROLLPOSITIONY": "0",
+            "__EVENTVALIDATION": self.get_session_variable(soup, "__EVENTVALIDATION"),
+            "ctl00$MainContent$searchBy_radio": "P",
+            "ctl00$MainContent$Street_textbox": "",
+            "ctl00$MainContent$Postcode_textbox": user_postcode,
+            "ctl00$MainContent$AddressLookup_button": "Find address"
+        }
+
+        # Build intermediate ASP.NET variables for uprn Select address
+        response = session.post(URL, data=form_data)
+        response.raise_for_status()
+        soup = BeautifulSoup(response.text, "html.parser")
+        form_data = {
+            "__EVENTTARGET": "",
+            "__EVENTARGUMENT": "",
+            "__VIEWSTATE": self.get_session_variable(soup, "__VIEWSTATE"),
+            "__VIEWSTATEGENERATOR": self.get_session_variable(soup, "__VIEWSTATEGENERATOR"),
+            "__SCROLLPOSITIONX": "0",
+            "__SCROLLPOSITIONY": "0",
+            "__EVENTVALIDATION": self.get_session_variable(soup, "__EVENTVALIDATION"),
+            "ctl00$MainContent$searchBy_radio": "P",
+            "ctl00$MainContent$Street_textbox": "",
+            "ctl00$MainContent$Postcode_textbox": user_postcode,
+            "ctl00$MainContent$lstAddresses": user_uprn,
+            "ctl00$MainContent$SelectAddress_button": "Select address"
+        }
+
+        # Actual http call to get Bins Data
+        response = session.post(URL, data=form_data)
+        response.raise_for_status()
+        soup = BeautifulSoup(response.text, "html.parser")
+        
+        # Find Bins table and data
+        table = soup.find("div", {"id": "binsGrid"})
+        if table:
+            rows = table.find_all("tr")
+            for row in rows:
+                columns = row.find_all("td")
+                if len(columns) >= 4:
+                    collection_type = columns[0].get_text(strip=True)
+                    collection_date_raw = columns[3].get_text(strip=True)
+                    # if the month number is a single digit there are 2 spaces, stripping all spaces to make it consistent
+                    collection_date = datetime.strptime(collection_date_raw.replace(" ", ""),'%a%b%d%Y')
+                    bin_entry = {
+                        "type": collection_type,
+                        "collectionDate": collection_date.strftime(date_format),
+                    }
+                    bin_data["bins"].append(bin_entry)
+        return bin_data