processor + parser: fix 3 backfill failure modes

## Context

After the first v2 backfill (commit f62c533), 72 of 73 real payslips
landed correctly, but three residual failure modes remained:

1. **doc_id=215** — a 1442-byte empty-text PDF that Claude
   hallucinated a `pay_date=1900-01-01 / gross=0 / net=0` row for.
   Data poison waiting to happen.
2. **doc_id=39** — a P60 End of Year Certificate. Got tagged
   `payslip` in Paperless, has no Paperless title, so the title-based
   filter couldn't catch it; the regex parser then happily pulled
   bogus numbers out of the P60 layout.
3. **doc_id=49** — a real June 2021 variant-A payslip with an
   `EE Discount BIK` line in BOTH Payments and Deductions at 12.00.
   The parser was configured to drop `EE Discount BIK` from
   `other_deductions` (treating it as a known mapped field), which
   caused validate_totals to fail by exactly 12.00.

## This change

### processor.py — defence in depth

- **`NON_PAYSLIP_CONTENT_RE`** — new regex run against the first
  500 chars of pdftotext output. Catches `P60 End of Year
  Certificate` and `Take-home income per month` (Viktor's comp
  estimation spreadsheet). First-500-char scoping keeps it from
  false-positiving a legit payslip that mentions "P60" in a
  footer.
- **Post-extraction sanity checks** — reject a ProcessResult if
  `pay_date.year < 2010` (Viktor joined Meta in 2019) or if
  `gross_pay == net_pay == 0`. These raise rather than insert,
  so the backfill's existing `except Exception` block logs and
  continues without poisoning the DB. Supersedes the 1900-01-01
  case that would otherwise slip through.

### meta_uk.py — variant A BIK fix

Removed `EE Discount BIK` from `VARIANT_A_DEDUCTIONS_KNOWN`. That
set filters items OUT of `other_deductions` (because they have
dedicated schema fields). `EE Discount BIK` has no dedicated
field — it should stay in `other_deductions` like Private Dental
and Private Medical so the validation math balances.

### Fixtures + tests

- New fixture `meta_uk_2021_06_variant_a_bik.txt` — real
  pdftotext from doc_id=49 — encodes the BIK-in-both-columns
  case so a regression would fail this fixture's validation test.
- `test_parses_variant_a_with_ee_discount_bik` — explicitly
  asserts `EE Discount BIK` lands in `other_deductions`.
- `test_rejects_implausible_pay_date`, `test_rejects_zero_gross_zero_net`
  — cover the two sanity-check branches.
- `test_skips_p60_by_content_when_title_is_null` — covers the
  content-based non-payslip filter.

## Test Plan

### Automated

```
$ poetry run pytest
============================== 57 passed in 2.42s ==============================
$ poetry run ruff check .
All checks passed!
$ poetry run mypy .
Success: no issues found in 24 source files
```

### Manual verification (after deploy + re-run backfill)

Expected DB shape:
- Total rows ≈ 71 (88 paperless tags − 15 non-payslip titles −
  2 null-title non-payslips caught by content filter)
- `validated = true` on ≥99% of rows
- No `pay_date < 2010` rows
- No rows with employer IS NULL

## Reproduce locally

1. `cd payslip-ingest && poetry run pytest`
2. Expected: 57 passed, including the 3 new processor tests and
   the 5 parametrised fixture-total-validation tests.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

payslip_ingest/parsers/meta_uk.py

@@ -328,7 +328,6 @@ VARIANT_A_DEDUCTIONS_KNOWN = {
     "Student Loans",
     "Student Loan",
     "RSU Net Gain",
-    "EE Discount BIK",
 }
 
 VARIANT_A_RSU_LABELS = {

payslip_ingest/processor.py

@@ -33,6 +33,17 @@ NON_PAYSLIP_TITLE_RE = re.compile(
     re.IGNORECASE,
 )
 
+# Some Paperless docs have no title at all — the title filter can't catch
+# them. These are detected by content signature in the pdftotext output.
+# Only apply to the first ~500 chars so we don't accidentally false-positive
+# a real payslip that happens to mention "P60" in a footnote somewhere.
+NON_PAYSLIP_CONTENT_RE = re.compile(
+    r"P60 End of Year Certificate"
+    r"|Employer's summary.+tax year ending"
+    r"|Take-home income per month",
+    re.IGNORECASE,
+)
+
 PDFTOTEXT_PATH = shutil.which("pdftotext")
 
 
@@ -71,8 +82,25 @@ async def process_document(
         return ProcessResult(doc_id=doc_id, status="skipped_non_payslip")
     pdf_bytes = await paperless.download_document(doc_id)
 
+    # Content-level non-payslip check (catches P60s with no Paperless title,
+    # personal income spreadsheets, etc.) before we burn extractor budget.
+    text_peek = _pdftotext(pdf_bytes) or ""
+    if NON_PAYSLIP_CONTENT_RE.search(text_peek[:500]):
+        log.info("skipping doc_id=%s — content matches non-payslip signature", doc_id)
+        return ProcessResult(doc_id=doc_id, status="skipped_non_payslip")
+
     extracted, which = await _extract(pdf_bytes, metadata, extractor)
 
+    # Sanity check: Viktor joined Meta UK in 2019. Any pay_date earlier
+    # than 2010 or a zero gross almost certainly means the extractor
+    # hallucinated on a non-payslip PDF that slipped past the title filter.
+    # Reject rather than poison the DB with a 1900-01-01 ghost row.
+    if extracted.pay_date.year < 2010:
+        raise ValueError(
+            f"doc_id={doc_id} extractor={which} produced implausible pay_date={extracted.pay_date}")
+    if extracted.gross_pay == 0 and extracted.net_pay == 0:
+        raise ValueError(f"doc_id={doc_id} extractor={which} produced zero gross and net")
+
     validated = validate_totals(extracted)
     if not validated:
         log.warning(

tests/fixtures/meta_uk_2021_06_variant_a_bik.txt

@@ -0,0 +1,43 @@
+254680A        Mr Viktor Barzin                                                                                    Facebook UK Ltd
+
+NI No : SZ762223D          NI Letter : A        Tax Code : 0T      Pay By : BACS             Date : 30 Jun 2021            Period : 3
+Description                                              Rate   Units                                This Period           This Year
+
+Salary                                                                                                5,096.65           15,289.95
+AE Pension                                                                                             (152.90)            (458.70)
+Laundry Expense                                                                                          40.00              120.00
+RSU Gain Taxable                                                                                                          8,518.63
+RSU Gain Nicable                                                                                                          8,518.63
+Transportation Allowance                                                                                                     73.10
+RSU Net Cash UK                                                                                                              31.38
+Private Dental Insurance                                                                                 15.61               46.83
+Private Medical Insurance                                                                                84.50              253.50
+EE Discount BIK                                                                                          12.00               12.00
+
+
+
+
+                                                                                     Total            5,095.86
+
+Tax                                                                                                  1,410.07             7,657.00
+National Insurance                                                                                     423.17             1,440.88
+RSU Net Gain                                                                                                             13,033.50
+Private Dental Insurance                                                                                 15.61               46.83
+Private Medical Insurance                                                                                84.50              253.50
+EE Discount BIK                                                                                          12.00               12.00
+Student Loans                                                                                          244.00             1,504.00
+
+
+
+
+                                                                                     Total            2,189.35
+Tax District : Pay As You Earn
+
+Tax Reference : 846/BA09294                                                        Net Pay             2,906.51
+
+Taxable Pay : This Period £5095.86 : To Date £23855.31
+Employers NIC This Period : 587.71
+Employers NIC To Date : 2,945.48
+Employers Pension This Period : 458.70
+Employers Pension To Date :1,376.10
+

tests/test_meta_uk_parser.py

@@ -86,6 +86,31 @@ def test_parses_variant_c_2022_11() -> None:
     assert result.ytd_gross == Decimal("131034.64")
 
 
+def test_parses_variant_a_with_ee_discount_bik() -> None:
+    """June 2021 — variant A. Real pdftotext from doc_id=49.
+
+    Has an `EE Discount BIK` line present in BOTH Payments AND Deductions
+    blocks with value 12.00. Needs to land in `other_deductions` so the
+    validation formula accounts for it (earlier parser version filtered
+    it out, causing an off-by-12.00 validation failure).
+    """
+    result = parse_meta_uk(_load("meta_uk_2021_06_variant_a_bik.txt"))
+
+    assert result.pay_date == date(2021, 6, 30)
+    assert result.salary == Decimal("5096.65")
+    assert result.pension_sacrifice == Decimal("152.90")
+    assert result.rsu_vest == Decimal("0")  # RSU lines are YTD-only this period
+    assert result.rsu_offset == Decimal("0")
+    assert result.gross_pay == Decimal("5095.86")
+    assert result.income_tax == Decimal("1410.07")
+    assert result.national_insurance == Decimal("423.17")
+    assert result.student_loan == Decimal("244.00")
+    assert result.net_pay == Decimal("2906.51")
+    assert result.other_deductions.get("Private Dental Insurance") == Decimal("15.61")
+    assert result.other_deductions.get("Private Medical Insurance") == Decimal("84.50")
+    assert result.other_deductions.get("EE Discount BIK") == Decimal("12.00")
+
+
 def test_parses_variant_a_2021_08() -> None:
     """Aug 2021 — variant A. Real pdftotext from doc_id=43.
 
@@ -135,6 +160,7 @@ def test_raises_on_empty_text() -> None:
     "meta_uk_2025_03.txt",
     "meta_uk_2022_11_variant_c.txt",
     "meta_uk_2021_08_variant_a.txt",
+    "meta_uk_2021_06_variant_a_bik.txt",
 ])
 def test_all_fixtures_validate_totals(fixture_name: str) -> None:
     """Every fixture must satisfy gross - deductions ≈ net within 2p."""

tests/test_processor.py

@@ -202,3 +202,47 @@ async def test_regex_miss_falls_back_to_claude(paperless: AsyncMock, extractor:
     assert result.status == "inserted"
     assert result.extractor == "claude"
     extractor.extract.assert_awaited_once()
+
+
+async def test_rejects_implausible_pay_date(paperless: AsyncMock, extractor: AsyncMock) -> None:
+    """Reject 1900-01-01-style hallucinations before they poison the DB."""
+    bad = _sample_extraction()
+    bad_dict = bad.model_dump()
+    bad_dict["pay_date"] = date(1900, 1, 1)
+    extractor.extract.return_value = ExtractedPayslip.model_validate(bad_dict)
+
+    factory = _SessionFactory([_FakeSession(existing_ids=[])])
+    with pytest.raises(ValueError, match="implausible pay_date"):
+        await process_document(42, factory, paperless, extractor)
+
+
+async def test_skips_p60_by_content_when_title_is_null(paperless: AsyncMock, extractor: AsyncMock,
+                                                       monkeypatch: pytest.MonkeyPatch) -> None:
+    """P60s get the `payslip` tag sometimes, and some have no title in Paperless.
+
+    The title filter can't catch them, so we also check the pdftotext output
+    for the `P60 End of Year Certificate` signature before hitting the
+    extractor.
+    """
+    paperless.get_document.return_value = {"id": 42, "title": None}
+    monkeypatch.setattr(processor, "_pdftotext",
+                        lambda _: "P60 End of Year Certificate\nTax year to 5 April 2021\n")
+
+    factory = _SessionFactory([_FakeSession(existing_ids=[])])
+    result = await process_document(42, factory, paperless, extractor)
+
+    assert result.status == "skipped_non_payslip"
+    extractor.extract.assert_not_called()
+
+
+async def test_rejects_zero_gross_zero_net(paperless: AsyncMock, extractor: AsyncMock) -> None:
+    """Reject the other common hallucination: all zeros on a non-payslip."""
+    bad = _sample_extraction()
+    bad_dict = bad.model_dump()
+    bad_dict["gross_pay"] = Decimal("0")
+    bad_dict["net_pay"] = Decimal("0")
+    extractor.extract.return_value = ExtractedPayslip.model_validate(bad_dict)
+
+    factory = _SessionFactory([_FakeSession(existing_ids=[])])
+    with pytest.raises(ValueError, match="zero gross and net"):
+        await process_document(42, factory, paperless, extractor)