t2a_v2/tests/test_referentiels.py

"""Tests pour le gestionnaire de référentiels et les routes Flask associées."""

from __future__ import annotations

import json
from pathlib import Path
from unittest.mock import patch, MagicMock

import pytest

from src.viewer.referentiels import ReferentielManager
from src.config import ALLOWED_EXTENSIONS, UPLOAD_MAX_SIZE_MB


# ---------------------------------------------------------------------------
# Tests ReferentielManager
# ---------------------------------------------------------------------------

class TestReferentielManager:

    @pytest.fixture
    def manager(self, tmp_path):
        return ReferentielManager(tmp_path / "refs")

    def test_add_file(self, manager):
        ref = manager.add_file("guide.pdf", b"fake pdf content")
        assert ref["filename"] == "guide.pdf"
        assert ref["extension"] == ".pdf"
        assert ref["status"] == "uploaded"
        assert ref["size_bytes"] == len(b"fake pdf content")
        assert ref["chunks_count"] == 0

    def test_list_all(self, manager):
        manager.add_file("a.txt", b"hello")
        manager.add_file("b.csv", b"col1,col2")
        assert len(manager.list_all()) == 2

    def test_get(self, manager):
        ref = manager.add_file("guide.pdf", b"content")
        found = manager.get(ref["id"])
        assert found is not None
        assert found["filename"] == "guide.pdf"

    def test_get_not_found(self, manager):
        assert manager.get("nonexistent") is None

    def test_remove(self, manager):
        ref = manager.add_file("guide.pdf", b"content")
        assert manager.remove(ref["id"]) is True
        assert len(manager.list_all()) == 0
        assert manager.get(ref["id"]) is None

    def test_remove_not_found(self, manager):
        assert manager.remove("nonexistent") is False

    def test_add_file_invalid_extension(self, manager):
        with pytest.raises(ValueError, match="Extension"):
            manager.add_file("malware.exe", b"evil")

    def test_add_file_too_large(self, manager):
        big_data = b"x" * (UPLOAD_MAX_SIZE_MB * 1024 * 1024 + 1)
        with pytest.raises(ValueError, match="volumineux"):
            manager.add_file("big.pdf", big_data)

    def test_persistence(self, tmp_path):
        """L'index persiste entre les instances."""
        dir_path = tmp_path / "refs"
        m1 = ReferentielManager(dir_path)
        m1.add_file("a.txt", b"hello")

        m2 = ReferentielManager(dir_path)
        assert len(m2.list_all()) == 1
        assert m2.list_all()[0]["filename"] == "a.txt"

    def test_file_stored_on_disk(self, manager, tmp_path):
        ref = manager.add_file("test.txt", b"file content here")
        stored_path = manager._dir / ref["stored_name"]
        assert stored_path.exists()
        assert stored_path.read_bytes() == b"file content here"

    def test_remove_deletes_file(self, manager):
        ref = manager.add_file("test.txt", b"content")
        stored_path = manager._dir / ref["stored_name"]
        assert stored_path.exists()
        manager.remove(ref["id"])
        assert not stored_path.exists()


# ---------------------------------------------------------------------------
# Tests chunking générique
# ---------------------------------------------------------------------------

class TestChunking:

    def test_chunk_txt(self, tmp_path):
        from src.medical.rag_index import chunk_user_file

        txt_file = tmp_path / "test.txt"
        txt_file.write_text(
            "Premier paragraphe avec assez de mots pour le seuil.\n\n"
            "Deuxième paragraphe avec encore plus de mots pour dépasser le minimum.\n\n"
            "Court\n\n"
            "Troisième paragraphe qui devrait aussi être un chunk valide.",
            encoding="utf-8",
        )

        chunks = chunk_user_file(txt_file, "test_doc")
        assert len(chunks) >= 2  # au moins 2 paragraphes assez longs
        assert all(c.document == "test_doc" for c in chunks)

    def test_chunk_csv(self, tmp_path):
        from src.medical.rag_index import chunk_user_file

        csv_file = tmp_path / "test.csv"
        csv_file.write_text(
            "code,description,note\n"
            "K85.1,Pancréatite aiguë biliaire,diagnostic fréquent\n"
            "I10,Hypertension essentielle,comorbidité courante\n",
            encoding="utf-8",
        )

        chunks = chunk_user_file(csv_file, "csv_doc")
        assert len(chunks) == 2
        assert "K85.1" in chunks[0].text
        assert "I10" in chunks[1].text

    def test_chunk_unsupported_extension(self, tmp_path):
        from src.medical.rag_index import chunk_user_file

        bad_file = tmp_path / "test.xyz"
        bad_file.write_text("content")

        chunks = chunk_user_file(bad_file, "bad")
        assert chunks == []


# ---------------------------------------------------------------------------
# Tests routes Flask
# ---------------------------------------------------------------------------

class TestReferentielRoutes:

    @pytest.fixture
    def app(self, tmp_path):
        """Crée une app Flask de test avec un manager temporaire."""
        from src.viewer.app import create_app
        app = create_app()
        app.config["TESTING"] = True
        return app

    @pytest.fixture
    def client(self, app):
        return app.test_client()

    def test_admin_page_loads(self, client):
        resp = client.get("/admin/referentiels")
        assert resp.status_code == 200
        assert "Référentiels RAG" in resp.data.decode()

    def test_upload_no_file(self, client):
        resp = client.post("/admin/referentiels/upload")
        assert resp.status_code == 400
        data = resp.get_json()
        assert "error" in data

    def test_upload_valid_file(self, client):
        from io import BytesIO
        data = {
            "file": (BytesIO(b"test content"), "doc.txt"),
        }
        resp = client.post("/admin/referentiels/upload", data=data, content_type="multipart/form-data")
        result = resp.get_json()
        assert resp.status_code == 200
        assert result["ok"] is True
        assert result["referentiel"]["filename"] == "doc.txt"

    def test_delete_nonexistent(self, client):
        resp = client.delete("/admin/referentiels/nonexistent")
        assert resp.status_code == 404