| """The printed Fundamentals and forecast table (the app's "Print / PDF"): the report HTML inside the page's own |
| stylesheet, every <details> opened, print media, Letter with 0.4in margins, printed by Chrome and read back with |
| pdfplumber. A long provenance header once pushed both value columns off the printed page (SPGI b3779fcd20362bdf). |
| |
| Skipped where Playwright, Chrome or pdfplumber is unavailable (the Space does not run the test suite).""" |
| import os |
| import pathlib |
| import sys |
| import tempfile |
|
|
| import pytest |
|
|
| sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__)))) |
| sys.path.insert(0, os.path.dirname(os.path.abspath(__file__))) |
|
|
| pytest.importorskip("playwright") |
| pdfplumber = pytest.importorskip("pdfplumber") |
| CHROME = "/Applications/Google Chrome.app/Contents/MacOS/Google Chrome" |
| if not os.path.exists(CHROME): |
| pytest.skip("Chrome is not installed", allow_module_level=True) |
|
|
| import catalog as C |
| import reconcile as RC |
| import report_model as RM |
| from test_spgi_lineage import good, patched |
|
|
| CSS = (pathlib.Path(__file__).resolve().parents[2] / "static" / "app.css").read_text() |
| MARGIN_PT = 0.4 * 72 |
| ROWS = (("Revenue", "REVENUE_TTM", "FORECAST_REVENUE"), ("Operating margin", "OPERATING_MARGIN_TTM", "FORECAST_OPERATING_MARGIN"), |
| ("Net margin", "NET_MARGIN_TTM", "FORECAST_NET_MARGIN"), ("EPS (adjusted basis)", "EPS_ADJUSTED_TTM", "FORECAST_EPS")) |
|
|
|
|
| def print_pdf(report_html, out, extra_css=""): |
| from playwright.sync_api import sync_playwright |
| page = (f"<!doctype html><html><head><meta charset='utf-8'><style>{CSS}{extra_css}</style></head><body class='has-analysis'>" |
| f"<main><section id='analysis' class='analysis'><div id='report'>{report_html}</div></section></main></body></html>") |
| with sync_playwright() as p: |
| b = p.chromium.launch(executable_path=CHROME) |
| pg = b.new_page() |
| pg.set_content(page, wait_until="load") |
| pg.evaluate("document.querySelectorAll('#report details').forEach(d => d.open = true)") |
| pg.emulate_media(media="print") |
| pg.pdf(path=out, format="Letter", print_background=True, |
| margin={k: "0.4in" for k in ("top", "bottom", "left", "right")}) |
| b.close() |
|
|
|
|
| def check_fundamentals(pdf_path, heads, values, footnotes): |
| """Problems with the printed table (empty = the 12 checks pass). |
| heads: the two column headings; values: {row label: (historical, forecast)}; footnotes: texts that must be printed.""" |
| bad = [] |
| with pdfplumber.open(pdf_path) as pdf: |
| page = next((p for p in pdf.pages if "Fundamentals and forecast" in (p.extract_text() or "")), None) |
| if page is None: |
| return ["no page carries the Fundamentals and forecast table"] |
| text = " ".join((page.extract_text() or "").split()) |
| words = page.extract_words(keep_blank_chars=False, use_text_flow=True) |
| top0 = next(w["top"] for w in words if w["text"] == "Fundamentals") |
| for h in heads: |
| if " ".join(h.upper().split()) not in text.upper(): |
| bad.append(f"column heading '{h}' is not fully printed") |
| table_words = [] |
| for label, (hist, fcst) in values.items(): |
| first = label.split()[0] |
| lw = next((w for w in words if w["text"] == first and w["top"] > top0), None) |
| if lw is None: |
| bad.append(f"row '{label}' is not printed") |
| continue |
| line = [w for w in words if abs(w["top"] - lw["top"]) < 4 and w["x0"] >= lw["x0"]] |
| table_words += line |
| got = [w["text"] for w in line] |
| for want, col in ((hist, "historical"), (fcst, "forecast")): |
| if want not in got: |
| bad.append(f"{label}: {col} value '{want}' is not visible (row prints {got})") |
| if hist in got and fcst in got and got.index(hist) > got.index(fcst): |
| bad.append(f"{label}: values out of column order") |
| for w in table_words: |
| if w["x0"] < MARGIN_PT - 2 or w["x1"] > page.width - MARGIN_PT + 2: |
| bad.append(f"'{w['text']}' is outside the page margins ({w['x0']:.0f}–{w['x1']:.0f} of {page.width:.0f})") |
| for i, a in enumerate(table_words): |
| for b in table_words[i + 1:]: |
| if a["x0"] < b["x1"] - 0.5 and b["x0"] < a["x1"] - 0.5 and a["top"] < b["bottom"] - 0.5 and b["top"] < a["bottom"] - 0.5: |
| bad.append(f"'{a['text']}' overlaps '{b['text']}'") |
| for f in footnotes: |
| if " ".join(f.split()) not in text: |
| bad.append(f"footnote not printed: '{f[:60]}…'") |
| return bad |
|
|
|
|
| def _spgi_report(st): |
| cat = C.build_catalog(st, include_deferred=True) |
| return RM.build_user_report(st, cat, {"summary": [], "thesis": [], "risks": [], "seats": {}}, []) |
|
|
|
|
| def _expected(rep): |
| i = next(i for i, b in enumerate(rep.blocks) if b.css == "rp-fund") |
| tbl = rep.blocks[i] |
| vals = {} |
| for row in tbl.content["rows"]: |
| label = "".join(f.text for f in row[0]) |
| vals[label] = ("".join(f.text for f in row[1]), "".join(f.text for f in row[2])) |
| feet = ["".join(f.text for f in b.content) for b in rep.blocks[i + 1:i + 3]] |
| return tbl.content["head"][1:], vals, feet |
|
|
|
|
| def _spgi_state_changes(st): |
| reports = [{"fiscalDateEnding": e} for e in ("2026-06-30", "2026-03-31", "2025-12-31", "2025-09-30")] |
| q = RC.ttm_quality({"status": "RECONCILED", "overrides": [], "unverified_quarters": ["2026-06-30"]}, reports) |
| return [("historical", "reconciliation", dict(st.historical.get("reconciliation"), ttm_quality=q)), |
| ("forecast", "consensus_anchor", {"eps": {"model_weight": 0.0, "reason": "RESTATEMENT"}, |
| "revenue": {"model_weight": 0.0, "reason": "RESTATEMENT"}})] |
|
|
|
|
| def test_printed_fundamentals_table_shows_every_value_and_footnote(good): |
| st = good.state |
| with patched(st, *_spgi_state_changes(st)): |
| rep = _spgi_report(st) |
| heads, vals, feet = _expected(rep) |
| assert len(vals) == 4 and all(h and f for h, f in vals.values()) |
| assert "through 2026-03-31" in feet[0] and "set from sell-side consensus" in feet[1] |
| with tempfile.TemporaryDirectory() as d: |
| out = os.path.join(d, "report.pdf") |
| print_pdf(RM.render_html(rep), out) |
| assert check_fundamentals(out, heads, vals, feet) == [] |
|
|
|
|
| def test_the_check_catches_the_old_clipped_table(good): |
| """Negative control: the old long provenance header with the old print behaviour (no wrapping, a scrolling |
| container) must fail the check, as the published b3779fcd20362bdf PDF did.""" |
| st = good.state |
| with patched(st, *_spgi_state_changes(st)): |
| rep = _spgi_report(st) |
| heads, vals, feet = _expected(rep) |
| tbl = next(b for b in rep.blocks if b.css == "rp-fund") |
| tbl.content["head"][1] = RM.ttm_header(RM.historical_quality(st), True) |
| old_css = "@media print{.rp-scroll{overflow-x:auto!important}.rp-t{width:auto!important;min-width:100%!important}" \ |
| ".rp-t th{white-space:nowrap!important}}" |
| with tempfile.TemporaryDirectory() as d: |
| out = os.path.join(d, "old.pdf") |
| print_pdf(RM.render_html(rep), out, extra_css=old_css) |
| assert check_fundamentals(out, [tbl.content["head"][1], heads[1]], vals, feet) |
|
|
