pdf/tests/edits/test_replace_text.py

#!/usr/bin/env python3
"""Round-trip tests for the `replace_text` op (line-level text rewriting).

`replace_text` targets stable page-object indices (from the document model's
run.object_indices) rather than coordinates, copies the original styling, and
reflows the rest of the line by the width delta. Verifies that an edit actually
replaces the text and round-trips through save/reload.

Run:  gateway/.venv/Scripts/python.exe tests/edits/test_replace_text.py
"""

from __future__ import annotations

import json
import sys
from pathlib import Path

ROOT = Path(__file__).resolve().parents[2]
sys.path.insert(0, str(ROOT / "gateway"))
import pdfengine  # noqa: E402

CORPUS = ROOT / "corpus" / "fonts" / "utf-8.pdf"


def _load():
    return pdfengine.PdfDocument.load_from_memory(CORPUS.read_bytes(), "")


def _runs(model):
    return [r for p in model.paragraphs for l in p.lines for r in l.runs]


def _alltext(model):
    return " ".join(r.text for r in _runs(model))


def _first_editable(model):
    for r in _runs(model):
        if list(r.object_indices):
            return r
    return None


def _op(run, new_text):
    return {"version": "1.0", "operations": [{
        "id": "r1", "type": "replace_text", "pageIndex": 0,
        "data": {"objectIndices": list(run.object_indices), "text": new_text,
                 "internalFontId": run.internal_font_id, "fontSize": run.font_size},
    }]}


def test_roundtrip_replaces_text():
    doc = _load()
    run = _first_editable(doc.get_page(0).extract_document_model())
    assert run is not None, "no run with object_indices to edit"
    original_fragment = run.text.strip().split()[0]
    doc.apply_edits(json.dumps(_op(run, "REPLACED LINE")))
    out = doc.save_full()
    text2 = _alltext(pdfengine.PdfDocument.load_from_memory(out, "").get_page(0).extract_document_model())
    assert "REPLACED LINE" in text2, "new text not found after round-trip"
    assert original_fragment not in text2, f"original text '{original_fragment}' should be gone"
    print(f"  ok  roundtrip: replaced run (objs {list(run.object_indices)}) -> 'REPLACED LINE'")


def test_reflow_keeps_other_text():
    # Replacing the first run must not destroy text on other lines.
    doc = _load()
    model = doc.get_page(0).extract_document_model()
    run = _first_editable(model)
    other_lines = [r.text for r in _runs(model) if list(r.object_indices) != list(run.object_indices)]
    doc.apply_edits(json.dumps(_op(run, "X")))
    text2 = _alltext(pdfengine.PdfDocument.load_from_memory(doc.save_full(), "").get_page(0).extract_document_model())
    survived = sum(1 for t in other_lines if t.strip() and t.strip() in text2)
    assert survived >= max(1, len(other_lines) // 2), "too much surrounding text lost"
    print(f"  ok  reflow: {survived}/{len(other_lines)} other runs preserved")


def test_empty_indices_is_noop():
    doc = _load()
    before = _alltext(doc.get_page(0).extract_document_model())
    doc.apply_edits(json.dumps({"version": "1.0", "operations": [{
        "id": "r1", "type": "replace_text", "pageIndex": 0,
        "data": {"objectIndices": [], "text": "X", "internalFontId": "", "fontSize": 12.0}}]}))
    after = _alltext(pdfengine.PdfDocument.load_from_memory(doc.save_full(), "").get_page(0).extract_document_model())
    assert before == after, "empty objectIndices should be a no-op"
    print("  ok  empty objectIndices is a safe no-op")


def main() -> int:
    tests = [test_roundtrip_replaces_text, test_reflow_keeps_other_text, test_empty_indices_is_noop]
    failed = 0
    for t in tests:
        try:
            t()
        except AssertionError as exc:
            print(f"  FAIL {t.__name__}: {exc}")
            failed += 1
    print(f"\n{len(tests) - failed}/{len(tests)} passed.")
    return 1 if failed else 0


if __name__ == "__main__":
    raise SystemExit(main())