"""The `status_server=` switch on eco.acquisition.daq_client.Daq.

Daq.__init__ builds PVs and an Assembly, so these tests drive the three
status-related callbacks on a bare Daq instance with only the attributes
they actually touch set - the point is the branch logic and the fallback,
not the rest of the class.
"""

import time
import types

import pytest

from eco.acquisition.daq_client import Daq


class FakeScan:
    def __init__(self, runno=42):
        self.daq_run_number = types.SimpleNamespace(
            get_current_value=lambda: runno
        )
        self._scratch = {}
        self.scan_parameters = {}
        self._values_done = [1]

    def counter_scratch(self, name):
        return self._scratch.setdefault(name, {})

    def values_done(self):
        return self._values_done

    def set_scan_parameter(self, key, value):
        self.scan_parameters[key] = value


class FakeStatusClient:
    base_url = "http://fake:8091"

    def __init__(self, snapshot=None, health=None, fail_with=None):
        self._snapshot = snapshot or {
            "status": {"bernina.a": 1},
            "status_channels": {"bernina.a": "PV:A"},
            "status_times": {"bernina.a": 0.0},
            "selections": {},
            "saved_to": "/data/p1/run0042/aux/status.json",
            "generation": 3,
            "snapshot_seconds": 0.12,
        }
        self._health = health or {
            "ready": True, "n_initialized": 80, "n_target_names": 82, "n_failed": 2
        }
        self._fail_with = fail_with
        self.calls = []

    def snapshot(self, **kwargs):
        self.calls.append(kwargs)
        if self._fail_with:
            raise self._fail_with
        return dict(self._snapshot)

    def wait_ready(self, timeout=None, progress=False):
        if self._fail_with:
            raise self._fail_with
        return self._health

    def health(self):
        if self._fail_with:
            raise self._fail_with
        return {"ready": True, "state": "ready",
                "last_init_finished": time.time(), **self._health}


def _daq(status_server=None, namespace=None, strict=False):
    daq = Daq.__new__(Daq)
    daq.name = "daq"
    daq.instrument = None
    daq._pgroup = "p12345"
    daq.namespace = namespace
    daq._status_server = status_server
    daq._status_server_client = status_server
    daq.status_server_timeout = 5.0
    daq.status_server_snapshot_timeout = 60.0
    daq.status_server_strict = strict
    daq.status_server_wait_ready = 30
    daq.status_server_max_age = 12 * 3600
    daq.status_server_stale_action = "ask"
    daq.status_server_stale_timeout = 0.01
    # the pre-existing tests below cover the synchronous path; the async one
    # has its own set further down
    daq.status_server_async = False
    daq.aux_calls = []
    daq.append_aux = lambda *a, **kw: daq.aux_calls.append((a, kw))
    return daq


class FakeAliasTree:
    def __init__(self):
        self.calls = []

    def get_all(self, **kwargs):
        self.calls.append(kwargs)
        return [{"alias": "bernina.a", "channel": "PV:A", "channeltype": "CA"}]


class FakeNamespace:
    def __init__(self):
        self.init_calls = []
        self.status_calls = []
        self.alias = FakeAliasTree()

    def init_all(self, **kwargs):
        self.init_calls.append(kwargs)

    def get_status(self, **kwargs):
        self.status_calls.append(kwargs)
        return {"status": {"bernina.a": 99}, "status_channels": {}}


def test_status_client_is_built_from_a_url_string():
    daq = Daq.__new__(Daq)
    daq._status_server = "http://host:8091/"
    daq._status_server_client = None
    daq.status_server_timeout = 3.0
    daq.status_server_snapshot_timeout = 60.0
    client = daq.status_client
    assert client.base_url == "http://host:8091"
    assert daq.status_client is client  # cached


def test_no_status_server_means_no_client():
    daq = Daq.__new__(Daq)
    daq._status_server = None
    daq._status_server_client = None
    assert daq.status_client is None


def test_init_namespace_uses_server_and_skips_local_init(capsys):
    ns = FakeNamespace()
    daq = _daq(status_server=FakeStatusClient(), namespace=ns)
    daq.init_namespace()
    assert ns.init_calls == []
    assert "Using status server" in capsys.readouterr().out


def test_init_namespace_falls_back_to_local_init_when_server_is_down(capsys):
    ns = FakeNamespace()
    daq = _daq(
        status_server=FakeStatusClient(fail_with=TimeoutError("no server")),
        namespace=ns,
    )
    daq.init_namespace()
    assert len(ns.init_calls) == 1
    assert ns.init_calls[0]["background"] is False
    out = capsys.readouterr().out
    assert "WARNING" in out and "falling back" in out


def test_init_namespace_strict_mode_raises_instead_of_falling_back():
    ns = FakeNamespace()
    daq = _daq(
        status_server=FakeStatusClient(fail_with=TimeoutError("no server")),
        namespace=ns,
        strict=True,
    )
    with pytest.raises(RuntimeError, match="status server"):
        daq.init_namespace()
    assert ns.init_calls == []


def test_start_status_uses_server_and_uploads_the_server_written_file():
    client = FakeStatusClient()
    daq = _daq(status_server=client, namespace=FakeNamespace())
    scan = FakeScan(runno=42)
    daq.append_start_status_to_scan(scan=scan)

    assert client.calls[0]["save"] is True
    assert client.calls[0]["key"] == "status_run_start"
    assert client.calls[0]["run_number"] == 42
    assert client.calls[0]["pgroup"] == "p12345"

    stat = scan.counter_scratch("daq")["namespace_status"]["status_run_start"]
    assert stat["status"] == {"bernina.a": 1}
    # server-only bookkeeping is stripped before it reaches the run table
    assert "saved_to" not in stat and "generation" not in stat

    (args, kwargs) = daq.aux_calls[0]
    assert args[0] == "/data/p1/run0042/aux/status.json"
    assert kwargs["run_number"] == 42


def test_start_status_falls_back_to_local_namespace_on_server_error(monkeypatch):
    ns = FakeNamespace()
    daq = _daq(
        status_server=FakeStatusClient(fail_with=ConnectionError("refused")),
        namespace=ns,
    )
    scan = FakeScan()
    # stop before the local file write, which targets a real beamline path
    monkeypatch.setattr(daq, "get_last_run_number", lambda: 42, raising=False)
    with pytest.raises(Exception):
        # the local branch will fail writing to /sf/bernina/data/... in a
        # test environment; what matters is that it got that far, i.e. it
        # really did fall back instead of silently doing nothing.
        daq.append_start_status_to_scan(scan=scan)
    assert ns.status_calls, "did not fall back to namespace.get_status()"


def test_end_status_uses_server_and_sets_the_scan_parameter():
    client = FakeStatusClient()
    daq = _daq(status_server=client, namespace=FakeNamespace())
    scan = FakeScan(runno=7)
    daq.append_status_to_scan_and_store(scan)
    assert client.calls[0]["key"] == "status_run_end"
    assert scan.scan_parameters["status"] == "aux/status.json"
    assert (
        "status_run_end" in scan.counter_scratch("daq")["namespace_status"]
    )
    assert daq.aux_calls


def test_end_status_skipped_when_no_steps_were_done():
    client = FakeStatusClient()
    daq = _daq(status_server=client, namespace=FakeNamespace())
    scan = FakeScan()
    scan._values_done = []
    daq.append_status_to_scan_and_store(scan)
    assert client.calls == []


def test_async_write_job_is_joined_before_append_aux():
    class AsyncClient(FakeStatusClient):
        def __init__(self):
            super().__init__()
            self._snapshot["write_job_id"] = "job1"
            self.joined = []

        def wait_write_job(self, job_id, timeout=None):
            self.joined.append(job_id)
            return {"state": "done"}

    client = AsyncClient()
    daq = _daq(status_server=client, namespace=FakeNamespace())
    result = daq._status_from_server("status_run_start", 5, "p1", write_async=True)
    assert client.joined == ["job1"]
    assert result[1] == "/data/p1/run0042/aux/status.json"


def test_snapshot_and_health_get_separate_timeouts():
    """A single 10 s timeout made every real snapshot fail: /health answers
    in milliseconds, but a get_status() fan-out over ~14k bernina channels
    takes 10-20 s."""
    daq = Daq.__new__(Daq)
    daq._status_server = "http://host:8091"
    daq._status_server_client = None
    daq.status_server_timeout = 10.0
    daq.status_server_snapshot_timeout = 180.0
    client = daq.status_client
    assert client.timeout == 10.0
    assert client.snapshot_timeout == 180.0


# --------------------------------------------------------------------------
# freshness gate and the async capture path


def _health(ready=True, age_s=0.0, **extra):
    return {"ready": ready, "state": "ready" if ready else "initializing",
            "last_init_finished": time.time() - age_s,
            "n_initialized": 80, "n_target_names": 82, "n_failed": 2, **extra}


class HealthClient(FakeStatusClient):
    base_url = "http://fake:8091"

    def __init__(self, health=None, capture_fail=None, aliases_fail=None,
                 push_fail=None, start_recording_fail=None,
                 capture_recording_fail=None):
        super().__init__()
        self._health_body = health if health is not None else _health()
        self.captures = []
        self.alias_captures = []
        self.pushes = []
        self.recording_starts = []
        self.recording_captures = []
        self.reinits = 0
        self._capture_fail = capture_fail
        self._aliases_fail = aliases_fail
        self._push_fail = push_fail
        self._start_recording_fail = start_recording_fail
        self._capture_recording_fail = capture_recording_fail

    def start_recording(self, **kwargs):
        if self._start_recording_fail:
            raise self._start_recording_fail
        self.recording_starts.append(kwargs)
        return {"n_channels_attached": 100, "n_channels_requested": 120}

    def capture_recording(self, recording_id, pgroup, run_number, **kwargs):
        if self._capture_recording_fail:
            raise self._capture_recording_fail
        call = {"recording_id": recording_id, "pgroup": pgroup,
                "run_number": run_number, **kwargs}
        self.recording_captures.append(call)
        return {"job_id": "j3", "path": "/data/p1/run0042/aux/monitors.esc.h5"}

    def push_status(self, pgroup, run_number, values, key="status_run_start"):
        if self._push_fail:
            raise self._push_fail
        call = {"pgroup": pgroup, "run_number": run_number, "values": values,
                "key": key}
        self.pushes.append(call)
        return {"status": "ok", "n_values": len(values)}

    def health(self):
        return self._health_body

    def capture(self, **kwargs):
        if self._capture_fail:
            raise self._capture_fail
        self.captures.append(kwargs)
        return {"job_id": "j1", "path": "/data/p1/run0042/aux/status.json"}

    def capture_aliases(self, **kwargs):
        if self._aliases_fail:
            raise self._aliases_fail
        self.alias_captures.append(kwargs)
        return {"job_id": "j2", "path": "/data/p1/run0042/aux/aliases.json"}

    def wait_write_job(self, job_id, timeout=None):
        return {"state": "done", "path": "/data/p1/run0042/aux/status.json"}

    def reinit(self, **kwargs):
        self.reinits += 1
        self._health_body = _health(age_s=0.0)
        return self._health_body


def _daq_fresh(client, **kw):
    daq = _daq(status_server=client, namespace=FakeNamespace())
    daq.status_server_max_age = kw.pop("max_age", 12 * 3600)
    daq.status_server_stale_action = kw.pop("stale_action", "ask")
    daq.status_server_stale_timeout = kw.pop("stale_timeout", 0.01)
    daq.status_server_async = kw.pop("async_", True)
    return daq


def test_fresh_server_is_used():
    daq = _daq_fresh(HealthClient(_health(age_s=3600)))
    assert daq.use_status_server() is True


def test_unready_server_is_not_used(capsys):
    daq = _daq_fresh(HealthClient(_health(ready=False)))
    assert daq.use_status_server(verbose=True) is False
    assert "initializing" in capsys.readouterr().out


def test_unreachable_server_is_not_used(capsys):
    daq = _daq_fresh(HealthClient())
    daq._status_server_client._health_body = None

    def boom():
        raise ConnectionError("refused")

    daq._status_server_client.health = boom
    assert daq.use_status_server(verbose=True) is False
    assert "not reachable" in capsys.readouterr().out


def test_stale_server_falls_back_when_told_to(capsys):
    client = HealthClient(_health(age_s=20 * 3600))
    daq = _daq_fresh(client, stale_action="local")
    assert daq.use_status_server(verbose=True) is False
    assert client.reinits == 0
    assert "13.0 h" not in capsys.readouterr().out  # 20 h, not the limit


def test_stale_server_can_be_trusted_anyway():
    client = HealthClient(_health(age_s=20 * 3600))
    daq = _daq_fresh(client, stale_action="use")
    assert daq.use_status_server() is True
    assert client.reinits == 0


def test_stale_server_is_restarted_and_then_used(capsys):
    client = HealthClient(_health(age_s=20 * 3600))
    daq = _daq_fresh(client, stale_action="restart")
    assert daq.use_status_server() is True
    assert client.reinits == 1
    assert "refreshed" in capsys.readouterr().out


def test_stale_prompt_without_a_tty_takes_the_restart(capsys):
    """A scan started from a script has nobody to answer; the timeout takes
    the option that leaves the next run fast."""
    client = HealthClient(_health(age_s=20 * 3600))
    daq = _daq_fresh(client, stale_action="ask")
    assert daq.use_status_server() is True
    assert client.reinits == 1
    out = capsys.readouterr().out
    assert "20.0 h ago" in out and "restarting" in out


def test_stale_prompt_answered_with_l_uses_the_local_namespace(monkeypatch):
    import inputimeout as _inputimeout

    client = HealthClient(_health(age_s=20 * 3600))
    daq = _daq_fresh(client, stale_action="ask")
    monkeypatch.setattr(_inputimeout, "inputimeout", lambda **kw: "l")
    assert daq.use_status_server() is False
    assert client.reinits == 0


def test_max_age_none_disables_the_check():
    client = HealthClient(_health(age_s=1000 * 3600))
    daq = _daq_fresh(client, max_age=None)
    assert daq.use_status_server() is True
    assert client.reinits == 0


def test_scan_start_delegates_the_whole_capture_to_the_server(capsys):
    client = HealthClient()
    daq = _daq_fresh(client)
    scan = FakeScan(runno=42)
    daq.append_start_status_to_scan(scan=scan)

    assert client.captures == [
        {"pgroup": "p12345", "run_number": 42, "key": "status_run_start",
         "upload": True, "keep_status": True}
    ]
    # the client neither waits for the values nor uploads the file itself
    assert client.calls == []
    assert daq.aux_calls == []
    block = scan.counter_scratch("daq")["namespace_status"]["status_run_start"]
    assert block["written_by_status_server"].endswith("status.json")
    assert "delegated" in capsys.readouterr().out


# --------------------------------------------------------------------------
# aliases: same server-first, local-fallback shape as status


def test_aliases_delegate_to_the_server_and_never_touch_the_local_namespace():
    """The bug this exists to fix: before this branch existed,
    copy_aliases_to_scan always called self.namespace.alias.get_all()
    unconditionally, even in status-server mode - against a namespace that
    status-server mode deliberately never initializes."""
    client = HealthClient()
    ns = FakeNamespace()
    daq = _daq_fresh(client)
    daq.namespace = ns
    scan = FakeScan(runno=42)
    daq.copy_aliases_to_scan(scan)

    assert client.alias_captures == [
        {"pgroup": "p12345", "run_number": 42, "upload": True,
         "channeltypes": None}
    ]
    assert ns.alias.calls == [], "local namespace.alias was touched"
    assert scan.scan_parameters["aliases"] == "aux/aliases.json"
    job = scan.counter_scratch("daq")["status_jobs"]["aliases"]
    assert job["path"].endswith("aliases.json")


def test_aliases_only_sent_on_the_first_step():
    client = HealthClient()
    daq = _daq_fresh(client)
    daq.namespace = FakeNamespace()
    scan = FakeScan(runno=42)
    scan._values_done = [1, 2, 3]  # not the first step
    daq.copy_aliases_to_scan(scan)
    assert client.alias_captures == []


def test_aliases_send_aliases_now_overrides_the_step_guard():
    client = HealthClient()
    daq = _daq_fresh(client)
    daq.namespace = FakeNamespace()
    scan = FakeScan(runno=42)
    scan._values_done = [1, 2, 3]
    daq.copy_aliases_to_scan(scan, send_aliases_now=True)
    assert len(client.alias_captures) == 1


def test_aliases_capture_failure_falls_back_to_the_local_namespace(monkeypatch):
    client = HealthClient(aliases_fail=ConnectionError("refused"))
    ns = FakeNamespace()
    daq = _daq_fresh(client)
    daq.namespace = ns
    scan = FakeScan(runno=42)
    monkeypatch.setattr(daq, "get_last_run_number", lambda **kw: 42, raising=False)
    with pytest.raises(Exception):
        # the local branch fails writing to /sf/bernina/data/... in a test
        # environment; what matters is that it got that far.
        daq.copy_aliases_to_scan(scan)
    assert ns.alias.calls, "did not fall back to namespace.alias.get_all()"


# --------------------------------------------------------------------------
# compare_channels: server-first (compares the SERVER's OWN namespace, so
# this session's namespace is never forced), local fallback - see
# eco.aliases.channel_lists (tested directly in tests/test_channel_lists.py)
# and eco.status_server's /channels/compare endpoint (tested in
# tests/test_status_server.py).


class ChannelAliasTree:
    """Just enough of eco.aliases.Alias for compare_channels' local path -
    unlike FakeAliasTree above, actually honors channeltypes so these tests
    can put JF/BS-tagged entries in and see them (not) come back out."""

    def __init__(self, entries):
        self._entries = entries
        self.calls = []

    def get_all(self, joiner=".", channeltypes=None):
        self.calls.append(channeltypes)
        return [
            e for e in self._entries
            if not channeltypes or e["channeltype"] in channeltypes
        ]


class FakeChannelList:
    def __init__(self, value):
        self._value = value

    def get_current_value(self):
        return self._value


def _channels_namespace(entries):
    ns = FakeNamespace()
    ns.alias = ChannelAliasTree(entries)
    return ns


def test_compare_channels_delegates_to_the_server():
    client = HealthClient()
    client.compare_channels = lambda list_names=None: {
        "channels_JF": {"channeltype": "JF", "missing": ["JF03"],
                         "exceeding": [], "n_required": 2, "n_recorded": 1}
    }
    daq = _daq_fresh(client)
    result = daq.compare_channels(verbose=False)
    assert result["channels_JF"]["missing"] == ["JF03"]


def test_compare_channels_server_failure_falls_back_to_local(capsys):
    client = HealthClient()

    def boom(list_names=None):
        raise ConnectionError("refused")

    client.compare_channels = boom
    daq = _daq_fresh(client)
    daq.namespace = _channels_namespace([
        {"alias": "jf03", "channel": "JF03", "channeltype": "JF"},
        {"alias": "jf04", "channel": "JF04", "channeltype": "JF"},
    ])
    daq.channels = {"channels_JF": FakeChannelList(["JF03"])}
    result = daq.compare_channels(verbose=False)
    assert result["channels_JF"]["missing"] == ["JF04"]
    assert result["channels_JF"]["exceeding"] == []
    assert "WARNING" in capsys.readouterr().out


def test_compare_channels_local_reports_missing_and_exceeding():
    daq = _daq(status_server=None)
    daq.namespace = _channels_namespace([
        {"alias": "jf03", "channel": "JF03", "channeltype": "JF"},
    ])
    daq.channels = {"channels_JF": FakeChannelList(["JF03", "JF_OLD"])}
    result = daq.compare_channels(verbose=False)
    assert result["channels_JF"]["missing"] == []
    assert result["channels_JF"]["exceeding"] == ["JF_OLD"]


def test_compare_channels_raises_without_namespace_or_server():
    daq = _daq(status_server=None, namespace=None)
    with pytest.raises(RuntimeError, match="no local namespace"):
        daq.compare_channels(verbose=False)


def test_compare_channels_prints_a_summary_by_default(capsys):
    daq = _daq(status_server=None)
    daq.namespace = _channels_namespace([
        {"alias": "jf03", "channel": "JF03", "channeltype": "JF"},
    ])
    daq.channels = {"channels_JF": FakeChannelList(["JF03"])}
    daq.compare_channels()
    out = capsys.readouterr().out
    assert "channels_JF" in out and "OK" in out


# --------------------------------------------------------------------------
# start_scan_monitoring / end_scan_monitoring - wired into
# callbacks_start_scan/callbacks_end_scan (see Daq.__init__). There is no
# local fallback - a local recording would need this session's own
# namespace to hold a live monitor per channel for the scan's whole
# duration, exactly the per-session cost the status server exists to avoid
# - so these only ever do anything for a server-backed, daq-counter scan
# (any scan that runs this Daq's callbacks at all already implies "daq is a
# counter"); a scan with no status server configured gets no
# namespace_monitor.h5, the same way it already gets no server-backed
# aliases.json/status.json.


def test_start_scan_monitoring_starts_a_recording_named_for_the_run():
    client = HealthClient()
    daq = _daq_fresh(client)
    scan = FakeScan(runno=42)
    result = daq.start_scan_monitoring(scan)

    assert client.recording_starts == [
        {"recording_id": "p12345_run0042", "names": None, "mode": "throttle",
         "min_interval": 0.1, "pgroup": "p12345", "run_number": 42}
    ]
    assert result["n_channels_attached"] == 100
    assert (
        scan.counter_scratch("daq")["monitoring_recording_id"]
        == "p12345_run0042"
    )


def test_start_scan_monitoring_warns_about_failed_required_channels(capsys):
    client = HealthClient()
    client.start_recording = lambda **kw: {
        "n_channels_attached": 100, "n_channels_requested": 120,
        "failed_required": ["daq.something"],
    }
    daq = _daq_fresh(client)
    scan = FakeScan(runno=42)
    daq.start_scan_monitoring(scan)
    out = capsys.readouterr().out
    assert "REQUIRED" in out and "daq.something" in out


def test_start_scan_monitoring_is_a_noop_without_a_status_server():
    daq = _daq_fresh(None)
    daq._status_server = None
    daq._status_server_client = None
    scan = FakeScan(runno=42)
    assert daq.start_scan_monitoring(scan) is None


def test_start_scan_monitoring_is_a_noop_when_the_server_is_not_used(capsys):
    client = HealthClient(_health(ready=False))
    daq = _daq_fresh(client)
    scan = FakeScan(runno=42)
    assert daq.start_scan_monitoring(scan) is None
    assert client.recording_starts == []
    assert "monitoring_recording_id" not in scan.counter_scratch("daq")


def test_start_scan_monitoring_failure_falls_back_to_doing_nothing(capsys):
    client = HealthClient(start_recording_fail=ConnectionError("refused"))
    daq = _daq_fresh(client)
    scan = FakeScan(runno=42)
    assert daq.start_scan_monitoring(scan) is None
    assert "WARNING" in capsys.readouterr().out
    assert "monitoring_recording_id" not in scan.counter_scratch("daq")


def test_end_scan_monitoring_is_a_noop_when_nothing_was_started():
    client = HealthClient()
    daq = _daq_fresh(client)
    scan = FakeScan(runno=42)
    assert daq.end_scan_monitoring(scan) is None
    assert client.recording_captures == []


def test_end_scan_monitoring_captures_the_recording_started_for_this_scan():
    client = HealthClient()
    daq = _daq_fresh(client)
    scan = FakeScan(runno=42)
    daq.start_scan_monitoring(scan)
    job = daq.end_scan_monitoring(scan)

    assert client.recording_captures == [
        {"recording_id": "p12345_run0042", "pgroup": "p12345",
         "run_number": 42, "upload": True, "filename": "namespace_monitor.h5"}
    ]
    assert job["job_id"] == "j3"
    assert (
        scan.counter_scratch("daq")["status_jobs"]["recording"]["job_id"] == "j3"
    )
    # registered the same way copy_aliases_to_scan/append_start_status_to_scan
    # register "aliases"/"status", so scan_info_rel.json carries it too
    assert scan.scan_parameters["monitors"] == "aux/namespace_monitor.h5"


def test_end_scan_monitoring_failure_does_not_raise(capsys):
    client = HealthClient(capture_recording_fail=ConnectionError("refused"))
    daq = _daq_fresh(client)
    scan = FakeScan(runno=42)
    daq.start_scan_monitoring(scan)
    assert daq.end_scan_monitoring(scan) is None
    assert "WARNING" in capsys.readouterr().out


def test_start_scan_monitoring_is_wired_into_callbacks_start_scan_after_run_number():
    """Needs scan.daq_run_number (set by
    count_run_number_up_and_attach_to_scan) and the server-in-use decision
    (cached by append_start_status_to_scan) - both must run first."""
    import inspect

    source = inspect.getsource(Daq.__init__)
    assert (
        source.index("count_run_number_up_and_attach_to_scan")
        < source.index("append_start_status_to_scan")
        < source.index("self.start_scan_monitoring")
    )


def test_end_scan_monitoring_is_wired_into_callbacks_end_scan_before_scan_info():
    """Must run before callbacks_end_scan's copy_scan_info_to_raw, which is
    what actually writes the "monitors" scan parameter end_scan_monitoring
    registers into scan_info_rel.json. copy_scan_info_to_raw also appears
    earlier, in callbacks_end_step - rindex() to get the callbacks_end_scan
    occurrence, not that one."""
    import inspect

    source = inspect.getsource(Daq.__init__)
    assert (
        source.index("self.end_scan_monitoring")
        < source.rindex("self.copy_scan_info_to_raw")
    )


def test_scan_end_delegates_and_sets_the_scan_parameter():
    client = HealthClient()
    daq = _daq_fresh(client)
    scan = FakeScan(runno=7)
    daq.append_status_to_scan_and_store(scan)
    assert client.captures[0]["key"] == "status_run_end"
    assert scan.scan_parameters["status"] == "aux/status.json"


def test_the_server_decision_is_taken_once_per_scan():
    """Start from the server and end from the local namespace would be a
    quietly inconsistent run."""
    client = HealthClient()
    daq = _daq_fresh(client)
    scan = FakeScan(runno=7)
    calls = []
    real = daq.use_status_server
    daq.use_status_server = lambda **kw: (calls.append(1), real(**kw))[1]
    daq.append_start_status_to_scan(scan=scan)
    daq.append_status_to_scan_and_store(scan)
    assert len(calls) == 1


def test_capture_failure_falls_back_to_the_local_path(monkeypatch):
    client = HealthClient(capture_fail=ConnectionError("refused"))
    ns = FakeNamespace()
    daq = _daq_fresh(client)
    daq.namespace = ns
    scan = FakeScan()
    monkeypatch.setattr(daq, "get_last_run_number", lambda **kw: 42, raising=False)
    with pytest.raises(Exception):
        daq.append_start_status_to_scan(scan=scan)
    assert ns.status_calls, "did not fall back to namespace.get_status()"


def test_write_status_waits_for_the_job_when_called_standalone(monkeypatch):
    client = HealthClient()
    daq = _daq_fresh(client)
    monkeypatch.setattr(daq, "get_last_run_number", lambda **kw: 42, raising=False)
    path = daq.write_status(pgroup="p1", run_number=42)
    assert str(path).endswith("run0042/aux/status.json")
    assert client.captures[0]["run_number"] == 42


# --------------------------------------------------------------------------
# the run table is filled from the server's values, not from its own CA reads


class RunTableSpy:
    def __init__(self):
        self.calls = []

    def append_run(self, runno, metadata=None, d=None, **kw):
        self.calls.append({"runno": runno, "metadata": metadata, "d": d})


class RunTableScan(FakeScan):
    """FakeScan plus the attributes the run-table callback reads."""

    def __init__(self, runno=42):
        super().__init__(runno)
        get = lambda v: types.SimpleNamespace(get_current_value=lambda: v)
        self.description = get("a scan")
        self.values_todo = get([[0.0], [1.0]])
        self.counters_names = get(["daq"])
        self.scan_command = get("ascan(...)")
        self.pulses_per_step = [10, 10]
        self.adjustables = [types.SimpleNamespace(name="dummy", Id="PV:DUMMY")]
        self.initial_values = get({"dummy": 0.5})


def _runtable_daq(client, run_table):
    daq = _daq_fresh(client)
    daq.run_table = run_table
    return daq


def test_runtable_gets_the_flat_status_mapping_not_the_whole_block():
    rt = RunTableSpy()
    daq = _runtable_daq(HealthClient(), rt)
    scan = RunTableScan()
    scan.counter_scratch("daq")["namespace_status"] = {
        "status_run_start": {"status": {"bernina.a": 1}, "status_channels": {}}
    }
    daq._create_runtable_metadata_append_status_to_runtable(scan)
    assert rt.calls[0]["d"] == {"bernina.a": 1}


def test_runtable_append_is_deferred_until_the_capture_finishes():
    """Filling it inline would either block the scan or make the run table do
    the CA fan-out the server exists to remove."""
    rt = RunTableSpy()
    client = HealthClient()
    daq = _runtable_daq(client, rt)
    scan = RunTableScan(runno=7)

    collected = {"job_id": "j1", "state": "done", "status": {"bernina.b": 2}}
    client.wait_write_job = lambda job_id, timeout=None, include_status=False: collected

    daq.append_start_status_to_scan(scan=scan)
    daq._create_runtable_metadata_append_status_to_runtable(scan)

    deadline = time.time() + 5
    while time.time() < deadline and not rt.calls:
        time.sleep(0.02)
    assert rt.calls, "run table row was never appended"
    assert rt.calls[0]["d"] == {"bernina.b": 2}
    assert rt.calls[0]["runno"] == 7


def test_runtable_row_is_still_written_when_the_capture_fails(capsys):
    """A run with no run-table row is worse than one filled the slow way."""
    rt = RunTableSpy()
    client = HealthClient()
    daq = _runtable_daq(client, rt)
    scan = RunTableScan(runno=8)

    def _boom(job_id, timeout=None, include_status=False):
        raise ConnectionError("server went away")

    client.wait_write_job = _boom
    daq.append_start_status_to_scan(scan=scan)
    daq._create_runtable_metadata_append_status_to_runtable(scan)

    deadline = time.time() + 5
    while time.time() < deadline and not rt.calls:
        time.sleep(0.02)
    assert rt.calls and rt.calls[0]["d"] == {}
    assert "could not collect run 8 status" in capsys.readouterr().out


def test_runtable_append_is_inline_when_the_status_is_already_there():
    rt = RunTableSpy()
    daq = _runtable_daq(HealthClient(), rt)
    daq.status_server_async = False
    scan = RunTableScan(runno=9)
    scan.counter_scratch("daq")["namespace_status"] = {
        "status_run_start": {"status": {"bernina.c": 3}}
    }
    daq._create_runtable_metadata_append_status_to_runtable(scan)
    assert rt.calls[0]["d"] == {"bernina.c": 3}   # no thread involved


# --------------------------------------------------------------------------
# scans.acquiring_scan.* is pushed to the server -- see
# eco.status_server.namespace_store.NamespaceMonitorStore.push_status: the
# server can never poll these itself (no CA channel), so the client hands
# over the same values it already builds `metadata` from.


def test_push_acquiring_scan_status_sends_the_real_scan_values():
    rt = RunTableSpy()
    client = HealthClient()
    client.wait_write_job = lambda job_id, timeout=None, include_status=False: {
        "state": "done", "status": {}
    }
    daq = _runtable_daq(client, rt)
    scan = RunTableScan(runno=7)

    daq.append_start_status_to_scan(scan=scan)
    daq._create_runtable_metadata_append_status_to_runtable(scan)

    assert len(client.pushes) == 1
    push = client.pushes[0]
    assert push["pgroup"] == daq.pgroup
    assert push["run_number"] == 7
    assert push["key"] == "status_run_start"
    values = push["values"]
    assert values["scans.acquiring_scan.description"] == "a scan"
    assert values["scans.acquiring_scan.scan_command"] == "ascan(...)"
    assert values["scans.acquiring_scan.adjustables_names"] == ["dummy"]
    assert values["scans.acquiring_scan.initial_values"] == {"dummy": 0.5}
    assert values["scans.acquiring_scan.number_of_steps"] == 2
    assert "scans.acquiring_scan.start_time" in values


def test_push_is_skipped_when_status_is_already_inline():
    """No server job means nothing will ever read a pushed value back, so
    there is nothing to push."""
    rt = RunTableSpy()
    client = HealthClient()
    daq = _runtable_daq(client, rt)
    daq.status_server_async = False
    scan = RunTableScan(runno=9)
    scan.counter_scratch("daq")["namespace_status"] = {
        "status_run_start": {"status": {"bernina.c": 3}}
    }
    daq._create_runtable_metadata_append_status_to_runtable(scan)
    assert client.pushes == []


def test_push_failure_does_not_break_the_runtable_append(capsys):
    """A failed push must not cost the run its run-table row -- the row
    still gets everything under metadata.* regardless."""
    rt = RunTableSpy()
    client = HealthClient(push_fail=ConnectionError("server went away"))
    client.wait_write_job = lambda job_id, timeout=None, include_status=False: {
        "state": "done", "status": {}
    }
    daq = _runtable_daq(client, rt)
    scan = RunTableScan(runno=13)

    daq.append_start_status_to_scan(scan=scan)
    daq._create_runtable_metadata_append_status_to_runtable(scan)

    deadline = time.time() + 5
    while time.time() < deadline and not rt.calls:
        time.sleep(0.02)
    assert rt.calls, "run table row was never appended"
    assert rt.calls[0]["runno"] == 13
    assert "could not push scans.acquiring_scan status" in capsys.readouterr().out


def test_only_the_start_block_keeps_its_values_on_the_server():
    """Holding a few MB for a block nobody collects is just a leak."""
    client = HealthClient()
    daq = _daq_fresh(client)
    daq.append_start_status_to_scan(scan=RunTableScan(runno=11))
    daq.append_status_to_scan_and_store(RunTableScan(runno=11))
    assert client.captures[0]["keep_status"] is True
    assert client.captures[1]["keep_status"] is False


# --------------------------------------------------------------------------
# rate_multiplicator / get_detector_code_frequency: never a raw CA get, never
# a silent None reaching arithmetic (the pulse_id bug, 4th occurrence)


class FakeFreqMonitor:
    """Stands in for CallbackEpics(func="latest"): .data is the same dict
    object the real one mutates in place, so a cached reference stays live."""

    def __init__(self, initial=None):
        self.data = {"value": initial}
        self.started = False

    def start(self, add_current_value=True):
        self.started = True

    def push(self, value):
        self.data["value"] = value


class FreqDetector:
    def __init__(self, value, pvname="TEST:Evt-1-Freq-I", monitorable=True):
        self._value = value
        self.pvname = pvname
        self._monitor = FakeFreqMonitor(value) if monitorable else None

    def get_current_value(self):
        return self._value

    def set_current_value_callback(self, func="accumulate", **kwargs):
        if self._monitor is None:
            raise AttributeError("not monitorable")
        return self._monitor


def _event_master(freq_detector, code=50):
    em = types.SimpleNamespace()
    em.__dict__[f"code{code:03d}"] = types.SimpleNamespace(frequency=freq_detector)
    return em


def _freq_daq(freq_detector, code=50):
    daq = Daq.__new__(Daq)
    daq._event_master = None
    daq._detectors_event_code = None
    daq._frequency_detector = None
    daq._frequency_monitor = None
    daq._frequency_latest = {"value": None}
    event_master = _event_master(freq_detector, code=code)
    # replicate the __init__ snippet directly, since Daq.__new__ skips it
    try:
        daq._frequency_detector = event_master.__dict__[
            f"code{code:03d}"
        ].frequency
        mon = daq._frequency_detector.set_current_value_callback(func="latest")
        mon.start()
        daq._frequency_monitor = mon
        daq._frequency_latest = mon.data
    except Exception:
        pass
    daq._detectors_event_code = code
    return daq


def test_frequency_is_read_from_the_monitor_cache_not_a_fresh_get():
    det = FreqDetector(50.0)
    daq = _freq_daq(det)
    assert daq._frequency_monitor.started is True
    assert daq.get_detector_code_frequency() == 50.0
    assert daq.rate_multiplicator == 2


def test_frequency_cache_follows_live_monitor_updates():
    det = FreqDetector(50.0)
    daq = _freq_daq(det)
    det._monitor.push(25.0)
    assert daq.get_detector_code_frequency() == 25.0


def test_a_transient_none_from_the_monitor_falls_back_to_a_direct_read():
    """The exact failure mode observed live: pyepics silently returns None
    on a transient CA hiccup instead of raising."""
    det = FreqDetector(50.0)
    daq = _freq_daq(det)
    det._monitor.push(None)
    # get_current_value() still works even though the monitor cache is
    # momentarily empty
    assert daq.get_detector_code_frequency() == 50.0


def test_no_value_anywhere_raises_instead_of_dividing_by_none():
    """Before this fix: int(100 / freq) with freq=None -> TypeError, deep
    inside retrieve(), killing a real scan mid-run."""
    det = FreqDetector(None)
    daq = _freq_daq(det)
    with pytest.raises(TimeoutError, match="TEST:Evt-1-Freq-I"):
        daq.get_detector_code_frequency()
    with pytest.raises(TimeoutError):
        daq.rate_multiplicator


def test_a_non_monitorable_frequency_falls_back_to_direct_reads():
    """MasterEventCodeFix (fixed CTA sequencer codes) has no PV to monitor
    at all - set_current_value_callback isn't there, __init__'s attach must
    not blow up, and reads should still work via get_current_value()."""
    det = FreqDetector(50.0, monitorable=False)
    daq = _freq_daq(det)
    assert daq._frequency_monitor is None
    assert daq.get_detector_code_frequency() == 50.0


def test_missing_event_master_does_not_crash_init():
    daq = Daq.__new__(Daq)
    daq._frequency_detector = None
    daq._frequency_monitor = None
    daq._frequency_latest = {"value": None}
    daq._detectors_event_code = None
    with pytest.raises(TimeoutError):
        daq.get_detector_code_frequency()


# --------------------------------------------------------------------------
# _create_runtable_metadata_append_status_to_runtable must honor
# append_status_info=False (it silently didn't - a regression from the old
# combined elog+run_table callback, which had the guard)


def test_runtable_metadata_callback_skips_when_append_status_info_is_false():
    rt = RunTableSpy()
    daq = _runtable_daq(HealthClient(), rt)
    scan = RunTableScan(runno=99)
    daq._create_runtable_metadata_append_status_to_runtable(
        scan, append_status_info=False
    )
    assert rt.calls == [], (
        "run_table.append_run() ran despite append_status_info=False - this "
        "is what makes a session's first scan pay for Run_Table2's Google "
        "Sheets authentication even when status collection was disabled"
    )


def test_runtable_metadata_callback_still_runs_by_default():
    rt = RunTableSpy()
    daq = _runtable_daq(HealthClient(), rt)
    scan = RunTableScan(runno=100)
    daq._create_runtable_metadata_append_status_to_runtable(scan)
    assert len(rt.calls) == 1