pdf

1"""End-to-end tests for the pdf skill helper scripts. No network required."""2from __future__ import annotations3 4import json5import os6import subprocess7import sys8from pathlib import Path9 10import pytest11 12SCRIPTS = Path(__file__).resolve().parent.parent / "scripts"13 14 15def run(script: str, *args: str, expect: int = 0) -> subprocess.CompletedProcess:16    env = dict(os.environ, LC_ALL="C", LANG="C", PYTHONIOENCODING="utf-8")17    proc = subprocess.run(18        [sys.executable, str(SCRIPTS / script), *args],19        capture_output=True, text=True, encoding="utf-8", env=env,20    )21    assert proc.returncode == expect, f"{script} {args}: rc={proc.returncode}\n{proc.stderr}"22    return proc23 24 25@pytest.fixture(scope="module")26def workdir(tmp_path_factory) -> Path:27    return tmp_path_factory.mktemp("pdfwork")28 29 30@pytest.fixture(scope="module")31def sample_image(workdir: Path) -> Path:32    from PIL import Image33    img_path = workdir / "sample.png"34    img = Image.new("RGB", (120, 80), (30, 120, 200))35    img.save(img_path)36    return img_path37 38 39@pytest.fixture(scope="module")40def report_pdf(workdir: Path, sample_image: Path) -> Path:41    spec = {42        "title": "Quarterly Example Report",43        "author": "example-author",44        "elements": [45            {"type": "heading", "text": "Quarterly Example Report", "level": 1},46            {"type": "paragraph", "text": "This is the introduction paragraph with a marker UNIQUEMARK42."},47            {"type": "table", "rows": [["Region", "Units"], ["North", "1250"], ["South", "980"]], "header": True},48            {"type": "image", "path": str(sample_image), "width": 200},49            {"type": "pagebreak"},50            {"type": "heading", "text": "Appendix", "level": 2},51            {"type": "paragraph", "text": "Second page content."},52        ],53    }54    spec_path = workdir / "spec.json"55    spec_path.write_text(json.dumps(spec), encoding="utf-8")56    out = workdir / "report.pdf"57    run("pdf_create.py", str(spec_path), "-o", str(out))58    assert out.exists() and out.stat().st_size > 50059    return out60 61 62def test_create_and_meta(report_pdf: Path):63    meta = json.loads(run("pdf_read.py", str(report_pdf), "--meta").stdout)64    assert meta["page_count"] == 265    assert meta["encrypted"] is False66    assert meta["likely_scanned_pages"] == []67    assert "Quarterly Example Report" in meta["metadata"].get("Title", "")68 69 70def test_extract_text(report_pdf: Path):71    data = json.loads(run("pdf_read.py", str(report_pdf), "--text").stdout)72    assert data["page_count"] == 273    assert "UNIQUEMARK42" in data["pages"][0]74    assert "Appendix" in data["pages"][1]75    assert "Page 1" in data["pages"][0]  # page number footer76 77 78def test_extract_tables(report_pdf: Path, workdir: Path):79    csv_dir = workdir / "csvs"80    data = json.loads(run("pdf_read.py", str(report_pdf), "--tables",81                          "--csv-dir", str(csv_dir)).stdout)82    assert data["table_count"] >= 183    rows = data["tables"][0]["rows"]84    assert rows[0] == ["Region", "Units"]85    assert ["North", "1250"] in rows86    csv_files = list(csv_dir.glob("*.csv"))87    assert csv_files and "Region" in csv_files[0].read_text(encoding="utf-8")88 89 90@pytest.fixture(scope="module")91def form_pdf(workdir: Path) -> Path:92    from reportlab.lib.pagesizes import A493    from reportlab.pdfgen import canvas94    out = workdir / "form.pdf"95    c = canvas.Canvas(str(out), pagesize=A4)96    form = c.acroForm97    c.drawString(72, 760, "Example Form")98    form.textfield(name="surname", x=72, y=700, width=300, height=20, value="")99    form.checkbox(name="agree", x=72, y=660, buttonStyle="check")100    form.radio(name="color", value="red", x=72, y=620, selected=False)101    form.radio(name="color", value="blue", x=110, y=620, selected=True)102    form.choice(name="size", x=72, y=580, width=120, height=20,103                options=["small", "large"], value="small")104    c.save()105    return out106 107 108def test_form_fill_unicode_roundtrip(form_pdf: Path, workdir: Path):109    surname = "Фамилия — ‘test’"110    values = {"surname": surname, "agree": True, "color": "/red", "size": "large"}111    fields_json = workdir / "values.json"112    fields_json.write_text(json.dumps(values, ensure_ascii=False), encoding="utf-8")113    filled = workdir / "filled.pdf"114    run("pdf_fill_form.py", str(form_pdf), "--fields-json", str(fields_json),115        "-o", str(filled))116    data = json.loads(run("pdf_read.py", str(filled), "--fields").stdout)117    fields = data["fields"]118    assert fields["surname"]["value"] == surname119    assert fields["agree"]["value"] in ("/Yes", "/On", "True", "/1")120    assert fields["color"]["value"] == "/red"121    assert fields["size"]["value"] == "large"122 123 124def test_merge_split_rotate(report_pdf: Path, workdir: Path):125    merged = workdir / "merged.pdf"126    out = json.loads(run("pdf_merge.py", str(report_pdf), str(report_pdf),127                         "-o", str(merged), "--bookmarks").stdout)128    assert out["page_count"] == 4129 130    part = workdir / "part.pdf"131    out = json.loads(run("pdf_split.py", str(merged), "--pages", "2-3",132                         "--rotate", "90", "-o", str(part)).stdout)133    assert out["page_count"] == 2134    meta = json.loads(run("pdf_read.py", str(part), "--meta").stdout)135    assert meta["page_count"] == 2136    assert all(p["rotation"] % 360 == 90 for p in meta["pages"])137 138 139def test_watermark(report_pdf: Path, workdir: Path):140    # Build the stamp at mid-page so its text does not overlap existing141    # headings (overlapping glyphs confuse text extraction).142    from reportlab.lib.pagesizes import A4143    from reportlab.pdfgen import canvas144    stamp = workdir / "stamp.pdf"145    c = canvas.Canvas(str(stamp), pagesize=A4)146    c.setFont("Helvetica", 40)147    c.drawString(200, 400, "DRAFT")148    c.save()149    stamped = workdir / "stamped.pdf"150    run("pdf_watermark.py", str(report_pdf), "--stamp", str(stamp), "-o", str(stamped))151    data = json.loads(run("pdf_read.py", str(stamped), "--text").stdout)152    assert all("DRAFT" in page for page in data["pages"])153 154 155def test_encrypt_decrypt_roundtrip(report_pdf: Path, workdir: Path):156    enc = workdir / "enc.pdf"157    run("pdf_secure.py", str(report_pdf), "--encrypt", "-o", str(enc),158        "--user-password", "your-password")159    meta = json.loads(run("pdf_read.py", str(enc), "--meta").stdout)160    assert meta["encrypted"] is True161 162    dec = workdir / "dec.pdf"163    run("pdf_secure.py", str(enc), "--decrypt", "-o", str(dec),164        "--password", "your-password")165    data = json.loads(run("pdf_read.py", str(dec), "--text").stdout)166    assert "UNIQUEMARK42" in data["pages"][0]167 168 169def test_compress(report_pdf: Path, workdir: Path):170    out = workdir / "compressed.pdf"171    run("pdf_split.py", str(report_pdf), "--pages", "1-2", "--compress", "-o", str(out))172    data = json.loads(run("pdf_read.py", str(out), "--text").stdout)173    assert "UNIQUEMARK42" in data["pages"][0]174 175 176# ---------------------------------------------------------------- form creation177 178FORM_SPEC = {179    "title": "Example Intake Form",180    "page_size": "A4",181    "fields": [182        {"name": "surname", "type": "text", "page": 1, "label": "Surname",183         "label_box": [72, 700, 150, 714], "entry_box": [160, 696, 400, 716]},184        {"name": "agree", "type": "checkbox", "page": 1, "label": "I agree",185         "label_box": [72, 660, 150, 674], "entry_box": [160, 658, 176, 674]},186        {"name": "color", "type": "radio", "page": 1, "label": "Color",187         "label_box": [72, 620, 150, 634], "entry_box": [160, 616, 400, 636],188         "options": ["red", "blue"], "value": "blue"},189        {"name": "size", "type": "dropdown", "page": 1, "label": "Size",190         "label_box": [72, 580, 150, 594], "entry_box": [160, 576, 300, 596],191         "options": ["small", "large"], "value": "small"},192    ],193}194 195 196@pytest.fixture(scope="module")197def built_form(workdir: Path) -> Path:198    spec_path = workdir / "formspec.json"199    spec_path.write_text(json.dumps(FORM_SPEC), encoding="utf-8")200    out = workdir / "built_form.pdf"201    result = json.loads(run("pdf_make_form.py", str(spec_path), "-o", str(out)).stdout)202    assert len(result["fields"]) == 4203    return out204 205 206def test_make_form_lists_all_fields(built_form: Path):207    data = json.loads(run("pdf_read.py", str(built_form), "--fields").stdout)208    fields = data["fields"]209    assert set(fields) == {"surname", "agree", "color", "size"}210    assert fields["surname"]["type"] == "text"211    assert fields["agree"]["options"] == ["/Off", "/Yes"]212    assert fields["color"]["value"] == "/blue"  # pre-selected radio213    assert set(fields["size"]["options"]) == {"small", "large"}214    # label text is drawn on the page, not just stored in the widget215    text = json.loads(run("pdf_read.py", str(built_form), "--text").stdout)216    assert "Surname" in text["pages"][0] and "Size" in text["pages"][0]217 218 219def test_make_form_fill_roundtrip(built_form: Path, workdir: Path):220    values = {"surname": "Smith", "agree": True, "color": "/red", "size": "large"}221    vals = workdir / "builtvals.json"222    vals.write_text(json.dumps(values), encoding="utf-8")223    filled = workdir / "built_filled.pdf"224    run("pdf_fill_form.py", str(built_form), "--fields-json", str(vals), "-o", str(filled))225    fields = json.loads(run("pdf_read.py", str(filled), "--fields").stdout)["fields"]226    assert fields["surname"]["value"] == "Smith"227    assert fields["agree"]["value"] == "/Yes"228    assert fields["color"]["value"] == "/red"229    assert fields["size"]["value"] == "large"230 231 232# ------------------------------------------------------------- layout validation233 234def test_form_layout_valid_spec(workdir: Path):235    spec_path = workdir / "layout_ok.json"236    spec_path.write_text(json.dumps(FORM_SPEC), encoding="utf-8")237    report = json.loads(run("pdf_form_layout.py", str(spec_path)).stdout)238    assert report["ok"] is True239    assert report["errors"] == 0240    assert all(f["ok"] for f in report["fields"])241 242 243def test_form_layout_detects_problems(workdir: Path):244    bad = {245        "page_size": "A4",246        "fields": [247            # out of bounds (x1 beyond A4 width)248            {"name": "wide", "type": "text", "page": 1, "label": "Wide",249             "label_box": [10, 700, 60, 714], "entry_box": [70, 696, 900, 716]},250            # two overlapping entry boxes251            {"name": "one", "type": "text", "page": 1, "label": "One",252             "label_box": [10, 600, 60, 614], "entry_box": [70, 596, 300, 616]},253            {"name": "two", "type": "text", "page": 1, "label": "Two",254             "label_box": [10, 560, 60, 574], "entry_box": [200, 600, 400, 620]},255            # label far away from its entry256            {"name": "lost", "type": "text", "page": 1, "label": "Lost",257             "label_box": [10, 100, 60, 114], "entry_box": [400, 500, 500, 520]},258            # too small259            {"name": "tiny", "type": "text", "page": 1,260             "entry_box": [10, 50, 14, 54]},261        ],262    }263    spec_path = workdir / "layout_bad.json"264    spec_path.write_text(json.dumps(bad), encoding="utf-8")265    proc = run("pdf_form_layout.py", str(spec_path), expect=1)266    report = json.loads(proc.stdout)267    assert report["ok"] is False268    by_name = {f["name"]: f for f in report["fields"]}269    assert any("bounds" in p for p in by_name["wide"]["problems"])270    assert any("overlaps field" in p for p in by_name["two"]["problems"])271    assert any("from its entry" in p for p in by_name["lost"]["problems"])272    assert any("minimum size" in p for p in by_name["tiny"]["problems"])273    assert by_name["one"]["ok"]  # first of the overlapping pair reports clean274 275 276# ------------------------------------------------- rasterization (overlay, pages)277 278def _raster_available() -> bool:279    import shutil280    try:281        import pypdfium2  # noqa: F401282        return True283    except ImportError:284        return shutil.which("pdftoppm") is not None285 286 287def test_form_layout_overlay(built_form: Path, workdir: Path):288    spec_path = workdir / "formspec.json"289    out_png = workdir / "overlay.png"290    report = json.loads(run("pdf_form_layout.py", str(spec_path), "--pdf", str(built_form),291                            "--render-overlay", str(out_png)).stdout)292    overlay = report["overlay"]293    if _raster_available():294        assert overlay["rendered"] is True295        assert out_png.exists() and out_png.stat().st_size > 1000296        from PIL import Image297        with Image.open(out_png) as img:298            assert img.width > 100 and img.height > 100299    else:300        assert overlay["rendered"] is False301        assert overlay["missing"]  # install hints present302 303 304def test_form_layout_overlay_blank_page(workdir: Path):305    # No --pdf: overlay is drawn on a blank page, PIL-only, always renders.306    spec_path = workdir / "formspec.json"307    out_png = workdir / "overlay_blank.png"308    report = json.loads(run("pdf_form_layout.py", str(spec_path),309                            "--render-overlay", str(out_png)).stdout)310    assert report["overlay"]["rendered"] is True311    assert out_png.exists()312 313 314def test_page_image_export(report_pdf: Path, workdir: Path):315    out_dir = workdir / "pageimgs"316    result = json.loads(run("pdf_page_image.py", str(report_pdf), "--pages", "1-2",317                            "--dpi", "72", "--out-dir", str(out_dir)).stdout)318    if _raster_available():319        assert result["rendered"] is True320        assert len(result["files"]) == 2321        from PIL import Image322        with Image.open(result["files"][0]) as img:323            # A4 at 72 dpi is ~595x842 px324            assert 500 < img.width < 700325    else:326        assert result["rendered"] is False327        assert result["missing"]328 329 330def test_page_image_bad_range(report_pdf: Path, workdir: Path):331    if not _raster_available():332        pytest.skip("no rasterizer available")333    run("pdf_page_image.py", str(report_pdf), "--pages", "9",334        "--out-dir", str(workdir / "nope"), expect=4)335 336 337# --------------------------------------------------------------------- stamping338 339def test_stamp_text(report_pdf: Path, workdir: Path):340    out = workdir / "stamp_text.pdf"341    run("pdf_stamp.py", str(report_pdf), "-o", str(out),342        "--text", "STAMPMARK77", "--x", "150", "--y", "500",343        "--font-size", "30", "--color", "#cc0000", "--pages", "1")344    data = json.loads(run("pdf_read.py", str(out), "--text").stdout)345    assert "STAMPMARK77" in data["pages"][0]346    assert "STAMPMARK77" not in data["pages"][1]  # only page 1 stamped347 348 349def test_stamp_text_rotated_opacity(report_pdf: Path, workdir: Path):350    out = workdir / "stamp_rot.pdf"351    run("pdf_stamp.py", str(report_pdf), "-o", str(out),352        "--text", "DRAFT", "--x", "150", "--y", "400", "--font-size", "60",353        "--rotation", "45", "--opacity", "0.3")354    # Rotated glyphs confuse pdfplumber's line grouping; verify via pypdf.355    from pypdf import PdfReader356    text = PdfReader(str(out)).pages[0].extract_text()357    assert "DRAFT" in text358 359 360def test_stamp_image(report_pdf: Path, sample_image: Path, workdir: Path):361    out = workdir / "stamp_img.pdf"362    run("pdf_stamp.py", str(report_pdf), "-o", str(out),363        "--image", str(sample_image), "--x", "400", "--y", "60",364        "--width", "100", "--pages", "2")365    from pypdf import PdfReader366    before = PdfReader(str(report_pdf))367    after = PdfReader(str(out))368 369    def image_xobjects(page):370        res = page.get("/Resources", {})371        xo = res.get("/XObject")372        if xo is None:373            return 0374        return sum(1 for k in xo if xo[k].get("/Subtype") == "/Image")375 376    assert image_xobjects(after.pages[1]) > image_xobjects(before.pages[1])377    assert image_xobjects(after.pages[0]) == image_xobjects(before.pages[0])378 379 380# --------------------------------------------------------- metadata + attachments381 382def test_meta_set_and_clear(report_pdf: Path, workdir: Path):383    out = workdir / "meta_set.pdf"384    run("pdf_meta.py", str(report_pdf), "--set-meta", "-o", str(out),385        "--title", "Retitled Example", "--author", "example-author",386        "--subject", "Testing", "--keywords", "alpha, beta")387    meta = json.loads(run("pdf_read.py", str(out), "--meta").stdout)["metadata"]388    assert meta["Title"] == "Retitled Example"389    assert meta["Author"] == "example-author"390    assert meta["Subject"] == "Testing"391    assert meta["Keywords"] == "alpha, beta"392 393    cleared = workdir / "meta_clear.pdf"394    run("pdf_meta.py", str(out), "--clear-meta", "-o", str(cleared))395    meta = json.loads(run("pdf_read.py", str(cleared), "--meta").stdout)["metadata"]396    assert "Title" not in meta or meta.get("Title") in ("", None)397 398 399def test_attachments_roundtrip(report_pdf: Path, workdir: Path):400    payload = workdir / "payload.txt"401    payload.write_text("attachment payload UNIQUEATTACH99\n", encoding="utf-8")402    with_att = workdir / "with_att.pdf"403    run("pdf_meta.py", str(report_pdf), "--attach", str(payload), "-o", str(with_att))404 405    listing = json.loads(run("pdf_meta.py", str(with_att), "--list-attachments").stdout)406    assert listing["attachment_count"] == 1407    assert listing["attachments"] == ["payload.txt"]408 409    ext_dir = workdir / "extracted"410    result = json.loads(run("pdf_meta.py", str(with_att),411                            "--extract-attachments", str(ext_dir)).stdout)412    assert len(result["extracted"]) == 1413    extracted = Path(result["extracted"][0])414    assert "UNIQUEATTACH99" in extracted.read_text(encoding="utf-8")415