uk_bin_collection 0.140.0__py3-none-any.whl → 0.141.0__py3-none-any.whl

uk_bin_collection/tests/input.json

@@ -1493,6 +1493,15 @@
         "wiki_name": "Oxford City Council",
         "wiki_note": "You will need to use [FindMyAddress](https://www.findmyaddress.co.uk/search) to find the UPRN."
     },
+    "PeterboroughCityCouncil": {
+        "house_number": "7 Arundel Road, Peterborough, PE4 6JJ",
+        "postcode": "PE4 6JJ",
+        "skip_get_url": true,
+        "url": "https://report.peterborough.gov.uk/waste",
+        "web_driver": "http://selenium:4444",
+        "wiki_name": "Peterborough City Council",
+        "wiki_note": "Pass the full address as it appears o nthe Peterborough website and postcode in their respective parameters. This parser requires a Selenium webdriver."
+    },
     "PerthAndKinrossCouncil": {
         "uprn": "124032322",
         "url": "https://www.pkc.gov.uk",

uk_bin_collection/uk_bin_collection/councils/PeterboroughCityCouncil.py

@@ -0,0 +1,167 @@
+import time
+
+from bs4 import BeautifulSoup
+from selenium.webdriver.common.by import By
+from selenium.webdriver.support import expected_conditions as EC
+from selenium.webdriver.support.ui import Select, WebDriverWait
+
+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 parse_data(self, page: str, **kwargs) -> dict:
+        driver = None
+        try:
+            user_poan = kwargs.get("paon")
+            user_postcode = kwargs.get("postcode")
+            if not user_postcode:
+                raise ValueError("No postcode provided.")
+            check_postcode(user_postcode)
+
+            headless = kwargs.get("headless")
+            web_driver = kwargs.get("web_driver")
+            driver = create_webdriver(web_driver, headless, None, __name__)
+            page = "https://report.peterborough.gov.uk/waste"
+
+            driver.get(page)
+
+            wait = WebDriverWait(driver, 30)
+
+            try:
+                # Cookies confirmed working in selenium
+                accept_cookies_button = wait.until(
+                    EC.element_to_be_clickable(
+                        (
+                            By.XPATH,
+                            "//button/span[contains(text(), 'I Accept Cookies')]",
+                        )
+                    )
+                )
+                accept_cookies_button.click()
+            except:
+                print(
+                    "Accept cookies banner not found or clickable within the specified time."
+                )
+                pass
+
+            postcode_input = wait.until(
+                EC.presence_of_element_located((By.XPATH, '//input[@id="postcode"]'))
+            )
+
+            postcode_input.send_keys(user_postcode)
+
+            postcode_go_button = wait.until(
+                EC.element_to_be_clickable((By.XPATH, '//input[@id="go"]'))
+            )
+
+            postcode_go_button.click()
+
+            # Wait for the select address drop down to be present
+            select_address_input = wait.until(
+                EC.presence_of_element_located((By.XPATH, '//input[@id="address"]'))
+            )
+
+            select_address_input.click()
+            time.sleep(2)
+
+            select_address_input_item = wait.until(
+                EC.presence_of_element_located(
+                    (By.XPATH, f"//li[contains(text(), '{user_poan}')]")
+                )
+            )
+
+            select_address_input_item.click()
+
+            address_continue_button = wait.until(
+                EC.element_to_be_clickable((By.XPATH, '//input[@value="Continue"]'))
+            )
+
+            address_continue_button.click()
+
+            your_collections_heading = wait.until(
+                EC.presence_of_element_located(
+                    (By.XPATH, "//h2[contains(text(), 'Your collections')]")
+                )
+            )
+
+            results_page = wait.until(
+                EC.presence_of_element_located(
+                    (By.XPATH, "//div[@class='waste__collections']")
+                )
+            )
+
+            soup = BeautifulSoup(results_page.get_attribute("innerHTML"), "html.parser")
+
+            data = {"bins": []}
+            output_date_format = "%d/%m/%Y"
+            input_date_format = "%A, %d %B %Y"  # Expect: Thursday, 17 April 2025
+
+            # Each bin section is within a waste-service-wrapper div
+            collection_panels = soup.find_all("div", class_="waste-service-wrapper")
+
+            for panel in collection_panels:
+                try:
+                    # Bin type
+                    bin_type_tag = panel.find("h3", class_="waste-service-name")
+                    if not bin_type_tag:
+                        continue
+                    bin_type = bin_type_tag.get_text(strip=True)
+
+                    # Get 'Next collection' date
+                    rows = panel.find_all("div", class_="govuk-summary-list__row")
+                    next_collection = None
+                    for row in rows:
+                        key = row.find("dt", class_="govuk-summary-list__key")
+                        value = row.find("dd", class_="govuk-summary-list__value")
+                        if key and value and "Next collection" in key.get_text():
+                            raw_date = " ".join(value.get_text().split())
+
+                            # ✅ Remove st/nd/rd/th suffix from the day (e.g. 17th → 17)
+                            cleaned_date = re.sub(
+                                r"(\d{1,2})(st|nd|rd|th)", r"\1", raw_date
+                            )
+                            next_collection = cleaned_date
+                            break
+
+                    if not next_collection:
+                        continue
+
+                    print(f"Found next collection for {bin_type}: '{next_collection}'")
+
+                    parsed_date = datetime.strptime(next_collection, input_date_format)
+                    formatted_date = parsed_date.strftime(output_date_format)
+
+                    data["bins"].append(
+                        {
+                            "type": bin_type,
+                            "collectionDate": formatted_date,
+                        }
+                    )
+
+                except Exception as e:
+                    print(
+                        f"Error processing panel for bin '{bin_type if 'bin_type' in locals() else 'unknown'}': {e}"
+                    )
+
+            # Sort the data
+            data["bins"].sort(
+                key=lambda x: datetime.strptime(x["collectionDate"], output_date_format)
+            )
+        except Exception as e:
+            # Here you can log the exception if needed
+            print(f"An error occurred: {e}")
+            # Optionally, re-raise the exception if you want it to propagate
+            raise
+        finally:
+            # This block ensures that the driver is closed regardless of an exception
+            if driver:
+                driver.quit()
+        return data

{uk_bin_collection-0.140.0.dist-info → uk_bin_collection-0.141.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: uk_bin_collection
-Version: 0.140.0
+Version: 0.141.0
 Summary: Python Lib to collect UK Bin Data
 Author: Robert Bradley
 Author-email: robbrad182@gmail.com

{uk_bin_collection-0.140.0.dist-info → uk_bin_collection-0.141.0.dist-info}/RECORD

@@ -3,7 +3,7 @@ uk_bin_collection/tests/check_selenium_url_in_input.json.py,sha256=Iecdja0I3XIiY
 uk_bin_collection/tests/council_feature_input_parity.py,sha256=DO6Mk4ImYgM5ZCZ-cutwz5RoYYWZRLYx2tr6zIs_9Rc,3843
 uk_bin_collection/tests/features/environment.py,sha256=VQZjJdJI_kZn08M0j5cUgvKT4k3iTw8icJge1DGOkoA,127
 uk_bin_collection/tests/features/validate_council_outputs.feature,sha256=SJK-Vc737hrf03tssxxbeg_JIvAH-ddB8f6gU1LTbuQ,251
-uk_bin_collection/tests/input.json,sha256=b6MsNMMQbs1-lW6g1_dYz5WTiwt8Puuu3q_BEopiulk,120060
+uk_bin_collection/tests/input.json,sha256=uvi5_CrjVy26H4gkWdoRXCJ1wsJPgntzJB26hXwC5jI,120556
 uk_bin_collection/tests/output.schema,sha256=ZwKQBwYyTDEM4G2hJwfLUVM-5v1vKRvRK9W9SS1sd18,1086
 uk_bin_collection/tests/step_defs/step_helpers/file_handler.py,sha256=Ygzi4V0S1MIHqbdstUlIqtRIwnynvhu4UtpweJ6-5N8,1474
 uk_bin_collection/tests/step_defs/test_validate_council.py,sha256=VZ0a81sioJULD7syAYHjvK_-nT_Rd36tUyzPetSA0gk,3475
@@ -216,6 +216,7 @@ uk_bin_collection/uk_bin_collection/councils/OadbyAndWigstonBoroughCouncil.py,sh
 uk_bin_collection/uk_bin_collection/councils/OldhamCouncil.py,sha256=9dlesCxNoVXlmQaqZj7QFh00smnJbm1Gnjkr_Uvzurs,1771
 uk_bin_collection/uk_bin_collection/councils/OxfordCityCouncil.py,sha256=d_bY0cXRDH4kSoWGGCTNN61MNErapSOf2WSTYDJr2r8,2318
 uk_bin_collection/uk_bin_collection/councils/PerthAndKinrossCouncil.py,sha256=Kos5GzN2co3Ij3tSHOXB9S71Yt78RROCfVRtnh7M1VU,3657
+uk_bin_collection/uk_bin_collection/councils/PeterboroughCityCouncil.py,sha256=lOrDD4jfJ-_C5UwCGqRcQ1G-U1F5X6rf255ypzYEBcg,6300
 uk_bin_collection/uk_bin_collection/councils/PlymouthCouncil.py,sha256=FJqpJ0GJhpjYeyZ9ioZPkKGl-zrqMD3y5iKa07e_i30,3202
 uk_bin_collection/uk_bin_collection/councils/PortsmouthCityCouncil.py,sha256=xogNgVvwM5FljCziiNLgZ_wzkOnrQkifi1dkPMDRMtg,5588
 uk_bin_collection/uk_bin_collection/councils/PowysCouncil.py,sha256=db3Y5FJz-LFDqmVZqPdzcBxh0Q26OFPrbUxlQ7r4vsQ,5896
@@ -325,8 +326,8 @@ uk_bin_collection/uk_bin_collection/councils/YorkCouncil.py,sha256=I2kBYMlsD4bId
 uk_bin_collection/uk_bin_collection/councils/council_class_template/councilclasstemplate.py,sha256=EQWRhZ2pEejlvm0fPyOTsOHKvUZmPnxEYO_OWRGKTjs,1158
 uk_bin_collection/uk_bin_collection/create_new_council.py,sha256=m-IhmWmeWQlFsTZC4OxuFvtw5ZtB8EAJHxJTH4O59lQ,1536
 uk_bin_collection/uk_bin_collection/get_bin_data.py,sha256=YvmHfZqanwrJ8ToGch34x-L-7yPe31nB_x77_Mgl_vo,4545
-uk_bin_collection-0.140.0.dist-info/LICENSE,sha256=vABBUOzcrgfaTKpzeo-si9YVEun6juDkndqA8RKdKGs,1071
-uk_bin_collection-0.140.0.dist-info/METADATA,sha256=3Dw3sim4P-LwGOmRbcX7jSK15SFpT16G2lTDl_wcuj8,19851
-uk_bin_collection-0.140.0.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
-uk_bin_collection-0.140.0.dist-info/entry_points.txt,sha256=36WCSGMWSc916S3Hi1ZkazzDKHaJ6CD-4fCEFm5MIao,90
-uk_bin_collection-0.140.0.dist-info/RECORD,,
+uk_bin_collection-0.141.0.dist-info/LICENSE,sha256=vABBUOzcrgfaTKpzeo-si9YVEun6juDkndqA8RKdKGs,1071
+uk_bin_collection-0.141.0.dist-info/METADATA,sha256=wkZzAu4PUwGpBcnep3fvKUVjYSPYKVf4_wlShHwzfXM,19851
+uk_bin_collection-0.141.0.dist-info/WHEEL,sha256=Nq82e9rUAnEjt98J6MlVmMCZb-t9cYE2Ir1kpBmnWfs,88
+uk_bin_collection-0.141.0.dist-info/entry_points.txt,sha256=36WCSGMWSc916S3Hi1ZkazzDKHaJ6CD-4fCEFm5MIao,90
+uk_bin_collection-0.141.0.dist-info/RECORD,,