"""Unit test suite for pipeline modules.""" from __future__ import annotations import tempfile import unittest from pathlib import Path import numpy as np from PIL import Image, PngImagePlugin from pipeline.config import load_config, load_taxonomy, load_queries, get_paths from pipeline.manifest import Manifest, normalize_url from pipeline.queries import build_queries from pipeline.validator import validate_image from pipeline.processor import normalize_image from pipeline.dedup import compute_hashes, find_duplicate_groups class TestPipeline(unittest.TestCase): def setUp(self): self.cfg = load_config() self.temp_dir_obj = tempfile.TemporaryDirectory() self.tmp_dir = Path(self.temp_dir_obj.name) self.paths = get_paths(self.cfg, base=self.tmp_dir) self.paths.ensure() self.db_path = self.paths.manifest_db self.manifest = Manifest(self.db_path) def tearDown(self): self.manifest.close() self.temp_dir_obj.cleanup() def test_url_normalization(self): url = "HTTPS://Example.com:443/path/to/img.jpg?utm_source=test&gclid=123&ref=abc#section" norm = normalize_url(url) self.assertEqual(norm, "https://example.com/path/to/img.jpg?ref=abc") def test_manifest_sources_and_hits(self): url = "https://example.com/test_board.jpg" is_new = self.manifest.add_hit(url, "bing", "PCB socket", class_hint="cpu_socket") self.assertTrue(is_new) pending = self.manifest.pending_sources() self.assertEqual(len(pending), 1) self.assertEqual(pending[0]["url_norm"], "https://example.com/test_board.jpg") self.manifest.mark_source(pending[0]["url_norm"], status="downloaded", image_id="img_001") pending_after = self.manifest.pending_sources() self.assertEqual(len(pending_after), 0) def test_manifest_image_lifecycle(self): inserted = self.manifest.add_image( image_id="abc12345", sha256="1234567890abcdef", orig_path="originals/abc12345.jpg", status="downloaded", ) self.assertTrue(inserted) img = self.manifest.get_image("abc12345") self.assertIsNotNone(img) self.assertEqual(img["status"], "downloaded") self.manifest.update_image("abc12345", status="validated", width=1000, height=800) img_updated = self.manifest.get_image("abc12345") self.assertEqual(img_updated["status"], "validated") self.assertEqual(img_updated["width"], 1000) def test_query_builder(self): qcfg = load_queries(self.cfg) queries = build_queries(qcfg) self.assertGreater(len(queries), 0) classes = {q.class_hint for q in queries} self.assertIn("cpu_socket", classes) self.assertIn("ic_chip", classes) def test_validator(self): # Valid PCB image with noise so PNG/JPEG size > 5000 bytes test_img_path = self.tmp_dir / "valid_pcb.png" arr = np.random.randint(0, 255, (800, 1000, 3), dtype=np.uint8) img = Image.fromarray(arr) img.save(test_img_path, format="PNG") res = validate_image(test_img_path, self.cfg["validate"]) self.assertTrue(res.ok) self.assertEqual(res.width, 1000) self.assertEqual(res.height, 800) # Too small file test (< 5000 bytes) small_img_path = self.tmp_dir / "small.jpg" small_img = Image.new("RGB", (100, 100), color=(0, 0, 0)) small_img.save(small_img_path, format="JPEG") res_small = validate_image(small_img_path, self.cfg["validate"]) self.assertFalse(res_small.ok) self.assertEqual(res_small.reason, "too_small_file") # Low resolution test (> 5000 bytes via noise, but < 480px short side) lowres_path = self.tmp_dir / "low_res.png" lowres_arr = np.random.randint(0, 255, (300, 300, 3), dtype=np.uint8) lowres_img = Image.fromarray(lowres_arr) lowres_img.save(lowres_path, format="PNG") self.assertGreater(lowres_path.stat().st_size, 5000) res_lowres = validate_image(lowres_path, self.cfg["validate"]) self.assertFalse(res_lowres.ok) self.assertEqual(res_lowres.reason, "low_resolution") def test_ai_metadata_validation(self): ai_img_path = self.tmp_dir / "ai_gen.png" arr = np.random.randint(0, 255, (800, 1000, 3), dtype=np.uint8) img = Image.fromarray(arr) png_info = PngImagePlugin.PngInfo() png_info.add_text("Software", "Midjourney v6.0") img.save(ai_img_path, format="PNG", pnginfo=png_info) res = validate_image(ai_img_path, self.cfg["validate"]) self.assertFalse(res.ok) self.assertEqual(res.reason, "ai_generated_metadata") def test_processor(self): test_rgba_path = self.tmp_dir / "rgba_pcb.png" arr = np.random.randint(0, 255, (900, 1200, 4), dtype=np.uint8) img = Image.fromarray(arr, mode="RGBA") img.save(test_rgba_path, format="PNG") dst_path = self.tmp_dir / "normalized_pcb.jpg" res = normalize_image(test_rgba_path, dst_path, self.cfg["process"]) self.assertTrue(dst_path.exists()) self.assertEqual(res.width, 1200) with Image.open(dst_path) as norm_img: self.assertEqual(norm_img.mode, "RGB") def test_dedup_hashes(self): img_path = self.tmp_dir / "dedup_test.jpg" arr = np.random.randint(0, 255, (600, 600, 3), dtype=np.uint8) img = Image.fromarray(arr) img.save(img_path, format="JPEG") hashes = compute_hashes(img_path) self.assertIn("phash", hashes) self.assertIn("dhash", hashes) self.assertIn("phash_flip", hashes) self.assertIn("dhash_flip", hashes) def test_cli_all_subcommand_parser(self): import subprocess res = subprocess.run( [".venv/bin/python3", "scraper.py", "all", "--help"], capture_output=True, text=True, cwd=str(Path(__file__).resolve().parent.parent) ) self.assertEqual(res.returncode, 0) self.assertIn("--max-num", res.stdout) def test_splitter(self): from pipeline.splitter import run as run_splitter self.manifest.add_image(image_id="img1", sha256="sha1", status="deduped", cluster_id="c1", orig_path="orig1.jpg") self.manifest.add_image(image_id="img2", sha256="sha2", status="deduped", cluster_id="c1", orig_path="orig2.jpg") self.manifest.add_image(image_id="img3", sha256="sha3", status="deduped", cluster_id="c2", orig_path="orig3.jpg") splits_dir = self.tmp_dir / "splits" stats = run_splitter(self.manifest, self.cfg, self.tmp_dir, splits_dir, copy_files=False) self.assertEqual(stats["total_split"], 3) # Check cluster_id integrity: img1 and img2 must be in the same split img1 = self.manifest.get_image("img1") img2 = self.manifest.get_image("img2") self.assertEqual(img1["split"], img2["split"]) def test_cli_split_subcommand_parser(self): import subprocess res = subprocess.run( [".venv/bin/python3", "scraper.py", "split", "--help"], capture_output=True, text=True, cwd=str(Path(__file__).resolve().parent.parent) ) self.assertEqual(res.returncode, 0) self.assertIn("--train-ratio", res.stdout) if __name__ == "__main__": unittest.main()