from __future__ import annotations from ocrtech.engines import PaddleOcrEngine def test_paddleocr_engine_parses_v3_page_results() -> None: engine = object.__new__(PaddleOcrEngine) raw_result = [ { "input_path": "page.png", "rec_texts": ["नेपाल सरकार", "Mixed English 123"], "rec_scores": [0.98, 0.87], "rec_polys": [ [[10, 10], [210, 10], [210, 40], [10, 40]], [[10, 50], [250, 50], [250, 80], [10, 80]], ], } ] lines = engine._parse_result(raw_result, page_index=0) assert [line.text for line in lines] == ["नेपाल सरकार", "Mixed English 123"] assert [line.confidence for line in lines] == [0.98, 0.87] assert [line.line_id for line in lines] == ["p0-l0", "p0-l1"] def test_paddleocr_engine_parses_v3_json_wrapper_results() -> None: engine = object.__new__(PaddleOcrEngine) class FakeOcrResult: def __init__(self) -> None: self.json = { "res": { "rec_texts": ["तालिका", "Table"], "rec_scores": [0.75, 0.66], "rec_polys": [ [[0, 0], [80, 0], [80, 20], [0, 20]], [[0, 30], [80, 30], [80, 50], [0, 50]], ], } } lines = engine._parse_result([FakeOcrResult()], page_index=2) assert [line.text for line in lines] == ["तालिका", "Table"] assert [line.page_index for line in lines] == [2, 2] assert [line.line_id for line in lines] == ["p2-l0", "p2-l1"]