"""Tests para profile_database — perfilado de una base DuckDB + relaciones."""

import os
import sys
import tempfile

import duckdb

sys.path.insert(0, os.path.join(os.path.dirname(__file__), "..", ".."))

from pipelines.profile_database import profile_database


def _build_related_db(path: str) -> None:
    """Crea una DuckDB con 2 tablas relacionadas: customers <- orders.

    customers.id es clave; orders.customer_id contiene solo ids de customers,
    de modo que orders.customer_id -> customers.id es una FK detectable por
    containment.
    """
    conn = duckdb.connect(path)
    try:
        conn.execute(
            "CREATE TABLE customers (id INTEGER, name VARCHAR, city VARCHAR)"
        )
        conn.execute(
            "INSERT INTO customers VALUES "
            "(1,'Ana','Madrid'),(2,'Luis','Sevilla'),"
            "(3,'Marta','Bilbao'),(4,'Jon','Vigo')"
        )
        conn.execute(
            "CREATE TABLE orders (order_id INTEGER, customer_id INTEGER, total DOUBLE)"
        )
        conn.execute(
            "INSERT INTO orders VALUES "
            "(10,1,99.5),(11,1,12.0),(12,2,45.0),"
            "(13,3,7.25),(14,4,200.0),(15,2,33.3)"
        )
    finally:
        conn.close()


def test_profile_database_two_related_tables():
    with tempfile.TemporaryDirectory() as d:
        db_path = os.path.join(d, "shop.duckdb")
        _build_related_db(db_path)

        res = profile_database(db_path, write_report=False)

        # status ok y dos tablas perfiladas
        assert res["status"] == "ok", res
        prof = res["db_profile"]
        assert prof["n_tables"] == 2

        # los TableProfiles completos llegan para ambas tablas
        assert len(prof["table_profiles"]) == 2
        profiled_tables = {tp["table"] for tp in prof["table_profiles"]}
        assert profiled_tables == {"customers", "orders"}

        # se detecta la relacion orders.customer_id -> customers.id
        fks = prof["fk_candidates"]
        assert any(
            fk.get("from_table") == "orders"
            and fk.get("from_col") == "customer_id"
            and fk.get("to_table") == "customers"
            and fk.get("to_col") == "id"
            for fk in fks
        ), fks

        # el join graph trae un diagrama mermaid
        graph = prof["join_graph"]
        assert "mermaid" in graph
        assert isinstance(graph["mermaid"], str)
        assert graph["mermaid"].startswith("graph LR")

        # no se reportan paths cuando write_report=False
        assert res["report_md_path"] is None
        assert res["report_json_path"] is None


def test_profile_database_writes_report(tmp_path):
    db_path = os.path.join(str(tmp_path), "shop2.duckdb")
    _build_related_db(db_path)
    report_dir = os.path.join(str(tmp_path), "reports")

    res = profile_database(db_path, report_dir=report_dir, write_report=True)

    assert res["status"] == "ok", res
    assert res["report_md_path"] is not None
    assert res["report_json_path"] is not None
    assert os.path.exists(res["report_md_path"])
    assert os.path.exists(res["report_json_path"])
    md = open(res["report_md_path"], encoding="utf-8").read()
    assert "# EDA base —" in md
    assert "## Relaciones inter-tabla" in md
    assert "```mermaid" in md