FabledCurator/tests/test_patreon_client.py

"""PatreonClient tests — parsing only, no network.

The fixture is loaded from disk and fed directly to the pure parsing methods
(_transform / extract_media / _validate_response / parse_cursor_from_url); no
live requests call is exercised here.
"""

import json
from pathlib import Path

import pytest
import requests

from backend.app.services import patreon_client as pc_mod
from backend.app.services.native_ingest_common import retry_after_seconds
from backend.app.services.patreon_client import (
    MediaItem,
    PatreonAPIError,
    PatreonAuthError,
    PatreonClient,
    PatreonDriftError,
    parse_cursor_from_url,
)

_FIXTURE = Path(__file__).parent / "fixtures" / "patreon_posts_page1.json"


@pytest.fixture
def response():
    return json.loads(_FIXTURE.read_text())


@pytest.fixture
def client():
    # No cookies file → empty session, but we never make a request in tests.
    return PatreonClient(cookies_path=None)


def _post_by_id(response, post_id):
    for post in response["data"]:
        if post["id"] == post_id:
            return post
    raise AssertionError(f"no post {post_id} in fixture")


def test_transform_indexes_included(client, response):
    index = client._transform(response)
    assert index[("media", "9001")]["file_name"] == "gallery-one.jpg"
    assert index[("media", "9003")]["file_name"] == "bonus-pack.zip"
    assert index[("campaign", "5555")]["name"] == "Example Creator"
    # Unknown keys are absent, not errors.
    assert ("media", "0000") not in index


def test_extract_gallery_dedups_image_large(client, response):
    index = client._transform(response)
    post = _post_by_id(response, "1001")
    items = client.extract_media(post, index)

    # Two gallery images; the image_large cover shares filehash aaaa... with
    # gallery image 9001 → collapses to one, leaving exactly 2 items.
    assert len(items) == 2
    assert all(isinstance(m, MediaItem) for m in items)

    kinds = [m.kind for m in items]
    assert kinds == ["images", "images"]  # image_large deduped away

    filenames = {m.filename for m in items}
    assert filenames == {"gallery-one.jpg", "gallery-two.jpg"}

    hashes = {m.filehash for m in items}
    assert hashes == {"a" * 32, "b" * 32}
    assert all(m.post_id == "1001" for m in items)


def test_extract_attachment_postfile_same_file_collapses(client, response):
    index = client._transform(response)
    post = _post_by_id(response, "1002")
    items = client.extract_media(post, index)

    # The attachment (attachments_media 9003) and the post_file are the SAME
    # zip — both carry filehash c…. attachments resolve before postfile, so the
    # postfile collapses into the attachment by filehash → exactly one survives.
    # (postfile-as-a-kind is exercised separately by the video post test.)
    assert len(items) == 1
    item = items[0]
    assert item.kind == "attachments"
    assert item.filename == "bonus-pack.zip"
    assert item.filehash == "c" * 32


def test_extract_inline_content_img(client, response):
    index = client._transform(response)
    post = _post_by_id(response, "1003")
    items = client.extract_media(post, index)

    assert len(items) == 1
    item = items[0]
    assert item.kind == "content"
    assert item.filehash == "d" * 32
    # & in the src must be unescaped to & .
    assert "&" not in item.url
    assert "token=x&v=2" in item.url


def test_extract_video_postfile(client, response):
    index = client._transform(response)
    post = _post_by_id(response, "1004")
    items = client.extract_media(post, index)

    assert len(items) == 1
    item = items[0]
    assert item.kind == "postfile"
    assert item.url.startswith("https://stream.mux.com/")
    assert item.filehash == "e" * 32


def test_parse_cursor_from_url_extracts_cursor(response):
    next_url = response["links"]["next"]
    assert parse_cursor_from_url(next_url) == "NEXTCURSOR123"


def test_parse_cursor_from_url_none_when_absent():
    assert parse_cursor_from_url("https://www.patreon.com/api/posts?sort=-x") is None
    assert parse_cursor_from_url(None) is None


def test_iter_posts_yields_triples_and_paginates(client, response, monkeypatch):
    # Page 1 = the fixture (4 posts, links.next → NEXTCURSOR123); page 2 = one
    # post, no links.next → stop. Monkeypatch _fetch so no network is touched.
    page2 = {
        "data": [{"id": "1005", "type": "post", "attributes": {"title": "older"}}],
        "included": [],
        "links": {},
    }

    def fake_fetch(campaign_id, cursor):
        return page2 if cursor == "NEXTCURSOR123" else response

    monkeypatch.setattr(client, "_fetch", fake_fetch)

    seen = list(client.iter_posts("5555"))
    # 4 posts on page 1 + 1 on page 2.
    assert [p["id"] for p, _idx, _cur in seen] == ["1001", "1002", "1003", "1004", "1005"]
    # page_cursor is the cursor that FETCHED the post's page.
    cursors = {p["id"]: cur for p, _idx, cur in seen}
    assert cursors["1001"] is None and cursors["1005"] == "NEXTCURSOR123"
    # included_index is the page's flattened resources, ready for extract_media.
    page1_index = next(idx for p, idx, _cur in seen if p["id"] == "1001")
    assert page1_index[("media", "9001")]["file_name"] == "gallery-one.jpg"


def test_missing_data_raises_drift(client):
    with pytest.raises(PatreonDriftError):
        client._validate_response({"included": []})


def test_non_list_data_raises_drift(client):
    with pytest.raises(PatreonDriftError):
        client._validate_response({"data": {"id": "1"}})


def test_media_missing_file_name_falls_back_to_url_basename(client):
    # A gallery image with a valid URL but no file_name is NOT drift — Patreon
    # serves these (operator-flagged 2026-06-07, BlenderKnight post 73665615);
    # fall back to the URL basename like gallery-dl does.
    post = {
        "id": "2000",
        "type": "post",
        "attributes": {"title": "broken"},
        "relationships": {"images": {"data": [{"id": "7777", "type": "media"}]}},
    }
    index = {
        ("media", "7777"): {
            "download_url": "https://c10.patreonusercontent.com/4/orig/"
            + "f" * 32
            + "/x.jpg"
        }
    }
    items = client.extract_media(post, index)
    assert len(items) == 1
    assert items[0].filename == "x.jpg"
    assert items[0].kind == "images"


def test_media_referenced_but_absent_raises_drift(client):
    post = {
        "id": "2001",
        "type": "post",
        "attributes": {"title": "dangling"},
        "relationships": {"images": {"data": [{"id": "8888", "type": "media"}]}},
    }
    with pytest.raises(PatreonDriftError):
        client.extract_media(post, {})


# -- HTTP status classification (plan #697 step 4) -------------------------


class _FakeResp:
    def __init__(self, status_code, *, json_data=None, raise_json=False, headers=None):
        self.status_code = status_code
        self._json_data = json_data
        self._raise_json = raise_json
        self.headers = headers or {}

    def json(self):
        if self._raise_json:
            raise ValueError("Expecting value: line 1 column 1")
        return self._json_data


def _client_returning(monkeypatch, resp):
    client = PatreonClient(cookies_path=None)
    monkeypatch.setattr(client._session, "get", lambda *a, **k: resp)
    return client


@pytest.mark.parametrize("status", [401, 403])
def test_fetch_auth_status_raises_auth_error(monkeypatch, status):
    client = _client_returning(monkeypatch, _FakeResp(status))
    with pytest.raises(PatreonAuthError) as ei:
        client._fetch("5555", None)
    assert ei.value.status_code == status


def test_fetch_non_json_body_is_auth_not_drift(monkeypatch):
    # An HTML login/challenge page (non-JSON) means the session expired — auth,
    # actionable as "rotate cookies", NOT API drift.
    client = _client_returning(monkeypatch, _FakeResp(200, raise_json=True))
    with pytest.raises(PatreonAuthError):
        client._fetch("5555", None)


@pytest.mark.parametrize("status", [404, 429, 500])
def test_fetch_other_status_raises_api_error_with_code(monkeypatch, status):
    client = _client_returning(monkeypatch, _FakeResp(status))
    with pytest.raises(PatreonAPIError) as ei:
        client._fetch("5555", None)
    assert ei.value.status_code == status
    # Not auth — the ingester maps these by status, not to auth_error.
    assert not isinstance(ei.value, PatreonAuthError)


def test_fetch_ok_returns_payload(monkeypatch):
    client = _client_returning(monkeypatch, _FakeResp(200, json_data={"data": []}))
    assert client._fetch("5555", None) == {"data": []}


# -- verify_auth (credential probe; (True/False/None, message) contract) ---


def test_verify_auth_ok_when_page_authenticates(monkeypatch):
    client = _client_returning(monkeypatch, _FakeResp(200, json_data={"data": []}))
    ok, msg = client.verify_auth("5555")
    assert ok is True
    assert "valid" in msg.lower()


@pytest.mark.parametrize("status", [401, 403])
def test_verify_auth_false_when_rejected(monkeypatch, status):
    client = _client_returning(monkeypatch, _FakeResp(status))
    ok, _msg = client.verify_auth("5555")
    assert ok is False


def test_verify_auth_inconclusive_on_drift(monkeypatch):
    # 200 + JSON but missing top-level 'data' → drift → inconclusive, NOT a
    # credential verdict (our parser is stale, the cookie may be fine).
    client = _client_returning(monkeypatch, _FakeResp(200, json_data={"included": []}))
    ok, msg = client.verify_auth("5555")
    assert ok is None
    assert "api shape changed" in msg.lower()


def test_verify_auth_inconclusive_on_network(monkeypatch):
    import requests
    client = PatreonClient(cookies_path=None)

    def _raise(*a, **k):
        raise requests.ConnectionError("boom")
    monkeypatch.setattr(client._session, "get", _raise)
    ok, _msg = client.verify_auth("5555")
    assert ok is None


# -- 429 backoff + pacing (plan #703) --------------------------------------


def test_retry_after_seconds_honors_numeric_header():
    assert retry_after_seconds(_FakeResp(429, headers={"Retry-After": "7"}), 1) == 7.0


def test_retry_after_seconds_exponential_without_header():
    resp = _FakeResp(429)
    assert retry_after_seconds(resp, 1) == 2.0
    assert retry_after_seconds(resp, 2) == 4.0
    assert retry_after_seconds(resp, 3) == 8.0


def test_retry_after_seconds_caps():
    assert retry_after_seconds(_FakeResp(429), 10) == 30.0
    assert retry_after_seconds(_FakeResp(429, headers={"Retry-After": "999"}), 1) == 30.0


def _client_with_sequence(monkeypatch, responses):
    """A client whose session yields `responses` in order; time.sleep stubbed so
    backoff never really sleeps. Returns (client, slept_list)."""
    client = PatreonClient(cookies_path=None)
    it = iter(responses)
    monkeypatch.setattr(client._session, "get", lambda *a, **k: next(it))
    slept: list[float] = []
    monkeypatch.setattr(pc_mod.time, "sleep", lambda s: slept.append(s))
    return client, slept


def test_fetch_retries_then_succeeds_on_429(monkeypatch):
    client, slept = _client_with_sequence(monkeypatch, [
        _FakeResp(429, headers={"Retry-After": "5"}),
        _FakeResp(200, json_data={"data": []}),
    ])
    assert client._fetch("5555", None) == {"data": []}
    assert 5.0 in slept  # backed off once before the retry


def test_fetch_persistent_429_raises_after_retries(monkeypatch):
    client, _slept = _client_with_sequence(monkeypatch, [_FakeResp(429)] * 10)
    with pytest.raises(PatreonAPIError) as ei:
        client._fetch("5555", None)
    assert ei.value.status_code == 429
    assert not isinstance(ei.value, PatreonAuthError)


def test_fetch_persistent_429_surfaces_retry_after(monkeypatch):
    """plan #708 B1: a terminal 429 carries the server's raw Retry-After seconds
    (uncapped here — the cooldown clamps) so the platform cooldown matches it."""
    client, _slept = _client_with_sequence(
        monkeypatch, [_FakeResp(429, headers={"Retry-After": "120"})] * 10,
    )
    with pytest.raises(PatreonAPIError) as ei:
        client._fetch("5555", None)
    assert ei.value.status_code == 429
    assert ei.value.retry_after == 120.0


def test_request_sleep_paces_before_fetch(monkeypatch):
    client = PatreonClient(cookies_path=None, request_sleep=1.5)
    monkeypatch.setattr(
        client._session, "get",
        lambda *a, **k: _FakeResp(200, json_data={"data": []}),
    )
    slept: list[float] = []
    monkeypatch.setattr(pc_mod.time, "sleep", lambda s: slept.append(s))
    client._fetch("5555", None)
    assert slept == [1.5]


# -- fetch_post_detail_content (best-effort full-body enrichment) ----------


def test_fetch_post_detail_content_returns_legacy_content(monkeypatch):
    """Old posts still carry the flat `content` HTML — returned as-is. The detail
    fetch uses the DEFAULT fieldset (no sparse `fields[post]`, which nulls the
    body) — #842."""
    payload = {"data": {"id": "42", "attributes": {"content": "<p>full body</p>"}}}
    captured: dict = {}
    client = PatreonClient(cookies_path=None)

    def _get(url, params=None, timeout=None):
        captured["url"] = url
        captured["params"] = params
        return _FakeResp(200, json_data=payload)

    monkeypatch.setattr(client._session, "get", _get)
    assert client.fetch_post_detail_content("42") == "<p>full body</p>"
    assert captured["url"].endswith("/api/posts/42")
    # No sparse fieldset — that's what nulled the body on the live API.
    assert not captured["params"] or "fields[post]" not in captured["params"]


def test_fetch_post_detail_content_converts_content_json_string(monkeypatch):
    """Current posts: `content` is null and the body lives in the
    `content_json_string` ProseMirror doc → converted to HTML (#842)."""
    doc = (
        '{"type":"doc","content":[{"type":"paragraph","content":'
        '[{"type":"text","text":"Hello"}]}]}'
    )
    payload = {"data": {"attributes": {"content": None, "content_json_string": doc}}}
    client = _client_returning(monkeypatch, _FakeResp(200, json_data=payload))
    assert client.fetch_post_detail_content("42") == "<p>Hello</p>"


def test_fetch_post_detail_content_empty_body_is_none(monkeypatch):
    payload = {"data": {"attributes": {"content": "   ", "content_json_string": ""}}}
    client = _client_returning(monkeypatch, _FakeResp(200, json_data=payload))
    assert client.fetch_post_detail_content("42") is None


def test_fetch_post_detail_content_blank_id_skips_request(monkeypatch):
    calls: list = []
    client = PatreonClient(cookies_path=None)
    monkeypatch.setattr(
        client._session, "get",
        lambda *a, **k: calls.append(1) or _FakeResp(200, json_data={}),
    )
    assert client.fetch_post_detail_content("") is None
    assert calls == []


def test_fetch_post_detail_content_swallows_http_and_transport_errors(monkeypatch):
    # Non-200 → None (best-effort; a missing body must never fail the walk).
    client = _client_returning(monkeypatch, _FakeResp(500))
    assert client.fetch_post_detail_content("42") is None
    # Non-JSON body (HTML challenge) → None, not a raise.
    client = _client_returning(monkeypatch, _FakeResp(200, raise_json=True))
    assert client.fetch_post_detail_content("42") is None
    # Transport error → None.
    client = PatreonClient(cookies_path=None)

    def _boom(*a, **k):
        raise requests.ConnectionError("boom")

    monkeypatch.setattr(client._session, "get", _boom)
    assert client.fetch_post_detail_content("42") is None


# -- post_record_key (media-less post seen-ledger gate) --------------------


def test_post_record_key_returns_synthetic_key_and_id():
    assert PatreonClient.post_record_key({"id": "987"}) == ("post:987", "987")
    assert PatreonClient.post_record_key({"id": 987}) == ("post:987", "987")


def test_post_record_key_none_without_id():
    assert PatreonClient.post_record_key({}) is None
    assert PatreonClient.post_record_key({"id": None}) is None


# -- post_is_gated (#874 tier-gated access filter) --------------------------


def test_post_is_gated_only_on_explicit_false():
    # The blurred-preview case: account can't view the post → gate it.
    assert PatreonClient.post_is_gated(
        {"attributes": {"current_user_can_view": False}}
    ) is True


def test_post_is_gated_viewable_and_missing_are_not_gated():
    # Explicit True, a missing flag, an explicit None, and a missing attributes
    # block ALL count as viewable — gate only on an explicit False so an absent
    # field never over-filters accessible posts.
    assert PatreonClient.post_is_gated(
        {"attributes": {"current_user_can_view": True}}
    ) is False
    assert PatreonClient.post_is_gated({"attributes": {}}) is False
    assert PatreonClient.post_is_gated(
        {"attributes": {"current_user_can_view": None}}
    ) is False
    assert PatreonClient.post_is_gated({}) is False