fix: pair Lichfield bin headings with dates within their own card

[alec-pinson]

What

LichfieldDistrictCouncil collects two page-wide lists and zips them together by index:

bins  = soup.find_all("h3", class_="bin-collection-tasks__heading")
dates = soup.find_all("p",  class_="bin-collection-tasks__date")
for i in range(len(dates)):
    bint = " ".join(bins[i].text.split()[2:4])
    date = dates[i].text

That holds only while every heading has a corresponding date element. Two things on the current page break the assumption:

1. The calendar download link (Download your bin collection calendar) reuses the bin-collection-tasks__heading class but is not a bin, so it pads bins.
2. A bin with no scheduled collection renders a bin-collection-tasks__frequency (Collected every Monday) instead of a bin-collection-tasks__date. Food Waste Caddy is in this state on at least some addresses.

When a dateless card sits before the others, every bin is shifted onto the following bin's date and the final bin drops off the end of the loop. The failure is silent — output stays schema-valid, it's just wrong.

Reproduction

This affects the repo's own test UPRN for Lichfield, 100031694085. Ground truth from the page:

Card	Date shown
Black Bin	30th July
Food Waste Caddy	(none — "Collected every Thursday")
Blue Bin	6th August
Brown Bin	6th August
Blue Bag	6th August

Before this PR:

{"bins": [
  {"type": "Black Bin",  "collectionDate": "30/07/2026"},
  {"type": "Food Waste", "collectionDate": "06/08/2026"},
  {"type": "Blue Bin",   "collectionDate": "06/08/2026"},
  {"type": "Brown Bin",  "collectionDate": "06/08/2026"}
]}

Food Waste is given a date the page never showed, and Blue Bag is missing entirely.

After this PR:

{"bins": [
  {"type": "Black Bin", "collectionDate": "30/07/2026"},
  {"type": "Blue Bin",  "collectionDate": "06/08/2026"},
  {"type": "Brown Bin", "collectionDate": "06/08/2026"},
  {"type": "Blue Bag",  "collectionDate": "06/08/2026"}
]}

On a second address (394018646) the misalignment was more visible, because the dateless Food Waste Caddy card is first:

	Before	After	Page
Food Waste	27/07	(omitted)	no date
Blue Bin	27/07	27/07	27th July
Blue Bag	03/08	27/07	27th July
Black Bin	missing	03/08	3rd August

How

• Iterate the headings and read the date from within the same card, skipping cards that have no date. Removes the positional coupling between the two lists.
• Derive the bin name from the heading with its visually-hidden Your next / collection spans removed, rather than slicing words [2:4]. The slice truncated three-word names (Food Waste Caddy → Food Waste). Prefix/suffix stripping is kept as a fallback should those spans ever go away.

No change to the date parsing or year-rollover logic.

Notes

Reported downstream as bins going unavailable in Home Assistant — once a bin stops appearing in the output the coordinator stops creating its entities. Likely related to #1905, which presented the same way and was closed as council-side; this looks like the underlying parser fragility that made the page change bite.

Verified against the live site for both UPRNs above. black --check clean.

Summary by CodeRabbit

• Bug Fixes
  • Improved bin collection information parsing for Lichfield District Council.
  • Preserved multi-word bin names while removing hidden display text.
  • Improved matching between each bin type and its scheduled collection date.
  • Excluded bins without a scheduled collection date to prevent incomplete results.

[coderabbitai]

📝 Walkthrough

Walkthrough

LichfieldDistrictCouncil.parse_data now derives cleaned bin names and matches each bin heading with a date from the same card, skipping cards without scheduled dates.

Changes

Lichfield bin parsing

Layer / File(s)	Summary
Card-local bin extraction uk_bin_collection/uk_bin_collection/councils/LichfieldDistrictCouncil.py	Adds bin-name normalization that excludes visually hidden text and replaces index-based pairing with card-local date extraction, skipping unscheduled cards.

Estimated code review effort: 2 (Simple) | ~10 minutes

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the main fix: pairing Lichfield bin headings with dates inside each card.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

🧹 Nitpick comments (1)

uk_bin_collection/uk_bin_collection/councils/LichfieldDistrictCouncil.py (1)

40-71: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Add regression coverage for the card-local pairing.

Please add fixtures asserting that a dateless card does not shift subsequent dates, calendar headings are ignored, and multi-word names such as “Food Waste Caddy” are preserved.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@uk_bin_collection/uk_bin_collection/councils/LichfieldDistrictCouncil.py`
around lines 40 - 71, The current card-local parsing logic lacks regression
fixtures for its key pairing and heading-handling cases. Add fixtures and tests
covering a dateless card, ensuring later cards retain their own dates; calendar
download headings are ignored; and multi-word names such as “Food Waste Caddy”
are preserved by bin_name.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@uk_bin_collection/uk_bin_collection/councils/LichfieldDistrictCouncil.py`:
- Around line 40-71: The current card-local parsing logic lacks regression
fixtures for its key pairing and heading-handling cases. Add fixtures and tests
covering a dateless card, ensuring later cards retain their own dates; calendar
download headings are ignored; and multi-word names such as “Food Waste Caddy”
are preserved by bin_name.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: d4dbe72c-67d7-4d1c-abcb-8a739bfc2652

📥 Commits

Reviewing files that changed from the base of the PR and between 36d9205 and 97667fa.

📒 Files selected for processing (1)

• uk_bin_collection/uk_bin_collection/councils/LichfieldDistrictCouncil.py

[alec-pinson]

Lichfield recently sent out Food waste bins in the past week or 2 and broke the black bin

Tested the fix:

Will get claude to review code rabbit nitpick

[alec-pinson]

Re the nitpick asking for regression coverage on the card-local pairing:

The principle is right, and arguably this bug is the argument for it — the mis-pairing was silent precisely because the shifted output stayed schema-valid. The existing checks can't catch that class of failure by construction.

The obstacle is that there's no precedent for it here. As far as I can tell:

• uk_bin_collection/tests/ has no per-council unit tests and no HTML fixtures
• councils are validated only by validate_council_outputs.feature, which scrapes the live site and validates the result against output.schema
• nothing imports a council module directly for offline parsing

So adding fixtures for Lichfield alone would introduce an offline test pattern that none of the other 351 councils follow. That seemed like a maintainer decision rather than something to fold into a single-council bugfix, so I've left this PR as just the fix.

@robbrad — happy to add offline fixture tests here (dateless card not shifting subsequent dates, calendar heading ignored, multi-word Food Waste Caddy preserved by bin_name) if you'd like that pattern established, or to keep this PR minimal and raise the test-infrastructure question separately. Either works — let me know which you'd prefer.

[robbrad]

Thanks for this — the visually-hidden-span extraction approach for the bin name is cleaner than the original word-slicing and is exactly what got used. #2179 fixed the same underlying bug independently and also handled a case yours doesn't: a card with no date but a recognizable weekly frequency (e.g. Food Waste Caddy) now gets a derived next-occurrence date instead of being dropped from the output entirely, which matters for keeping that Home Assistant entity from going unavailable. Rather than pick one over the other, I synthesized both into a single fix in the consolidated "July Part 2" release: #2189 — your heading-text extraction combined with #2179's frequency fallback and year-rollover fix. Closing this one out — credited in the PR.

---

Generated by Claude Code