feat(core): add verified Russian C-SSRS Recent source content

tests/test_cssrs_content.py

@@ -0,0 +1,58 @@
+"""Protect the reviewed PDF transcription and its connection to Recent scoring."""
+
+import hashlib
+import json
+from itertools import product
+from pathlib import Path
+
+from app.core.cssrs_content import load_cssrs_russian
+from app.core.questionnaire import score_cssrs_recent
+
+CLINICAL = Path(__file__).resolve().parents[1] / "docs" / "clinical"
+
+
+def test_source_hashes_and_complete_transcription():
+    form = load_cssrs_russian()
+    manifest = json.loads((CLINICAL / "review-status.json").read_text())["cssrs_russian_pdf"]
+    assert hashlib.sha256((CLINICAL / manifest["file"]).read_bytes()).hexdigest() == manifest["sha256"]
+    source = (CLINICAL / manifest["source_text"]).read_bytes()
+    assert hashlib.sha256(source).hexdigest() == manifest["source_text_sha256"]
+    assert form.version == manifest["content_version"]
+    # Reconstruct the complete extracted page, including numbering, guidance,
+    # original typos, answer order and the color legend. No wording is dropped.
+    blocks = [form.title, form.subtitle, form.period_label_original, form.initial_instruction]
+    blocks.extend(option.label for option in form.options)
+    for index, item in enumerate(form.items):
+        blocks.append(f"{index + 1}. {item.text}" if index < 6 else item.text)
+        if item.interviewer_guidance:
+            blocks.append(item.interviewer_guidance)
+        if item.id == "q2":
+            blocks.append(form.branch_instruction)
+    blocks.extend(["Low Risk", "Moderate Risk", "High Risk"])
+    assert " ".join(blocks) == source.decode().strip()
+    assert [item.recall_period for item in form.items] == ["past_month"] * 5 + ["lifetime", "past_3_months"]
+
+
+def test_all_source_branches_and_color_categories_match_scorer():
+    form = load_cssrs_russian()
+    colors = {item.id: item.yes_risk for item in form.items}
+    assert list(colors.values()) == ["low", "low", "moderate", "high", "high", "moderate", "high"]
+    levels = ["none", "low", "moderate", "high"]
+    for values in product((0, 1), repeat=7):
+        all_answers = dict(zip(colors, values, strict=True))
+        asked = ["q1", "q2"]
+        if all_answers["q2"]:
+            asked.extend(["q3", "q4", "q5"])
+        asked.append("q6_lifetime")
+        if all_answers["q6_lifetime"]:
+            asked.append("q6_recent")
+        answers = {}
+        for item_id in asked:
+            partial = score_cssrs_recent(answers)
+            assert not partial.complete
+            assert partial.missing[0] == item_id
+            answers[item_id] = all_answers[item_id]
+        result = score_cssrs_recent(answers)
+        expected = max((levels.index(colors[key]) for key, value in answers.items() if value), default=0)
+        assert result.complete and result.risk == levels[expected]
+        assert result.required == tuple(asked)