add psycode support (#8)

adding psycode versions of workflow widgets and experiment objects

src/np_workflows/experiments/openscope_psycode/__init__.py

@@ -0,0 +1,2 @@
+from .main_psycode_pilot import new_experiment, Hab, Ephys, validate_selected_workflow
+from .psycode_workflow_widget import PsyCode_workflow_widget

src/np_workflows/experiments/openscope_psycode/main_psycode_pilot.py

@@ -0,0 +1,217 @@
+import configparser
+import contextlib
+import copy
+import dataclasses
+import datetime
+import enum
+import functools
+import pathlib
+import platform
+import shutil
+import threading
+import time
+import zlib
+from typing import ClassVar, Literal, NamedTuple, NoReturn, Optional, TypedDict
+
+import IPython
+import IPython.display
+import ipywidgets as ipw
+import np_config
+import np_logging
+import np_services
+import np_session
+import np_workflows
+import PIL.Image
+import pydantic
+from pyparsing import Any
+from np_services import (
+    Service,
+    Finalizable,
+    ScriptCamstim, SessionCamstim, 
+    SessionCamstim,
+    OpenEphys,
+    Sync,
+    VideoMVR,
+    NewScaleCoordinateRecorder,
+    MouseDirector,
+)
+
+logger = np_logging.getLogger(__name__)
+
+
+class PsyCodeSession(enum.Enum):
+    """Enum for the different sessions available, each with different param sets."""
+
+    PRETEST = "pretest"
+    HAB = "hab"
+    EPHYS = "ephys"
+
+
+class PsyCodeMixin:
+    """Provides project-specific methods and attributes, mainly related to camstim scripts."""
+
+    workflow: PsyCodeSession
+    """Enum for particular workflow/session, e.g. PRETEST, HAB_60, HAB_90,
+    EPHYS."""
+
+    session: np_session.PipelineSession
+    mouse: np_session.Mouse
+    user: np_session.User
+    platform_json: np_session.PlatformJson
+
+    @property
+    def recorders(self) -> tuple[Service, ...]:
+        """Services to be started before stimuli run, and stopped after. Session-dependent."""
+        match self.workflow:
+            case PsyCodeSession.PRETEST | PsyCodeSession.EPHYS:
+                return (Sync, VideoMVR, OpenEphys)
+            case PsyCodeSession.HAB:
+                return (Sync, VideoMVR)
+
+    @property
+    def stims(self) -> tuple[Service, ...]:
+        return (SessionCamstim, )
+
+    def initialize_and_test_services(self) -> None:
+        """Configure, initialize (ie. reset), then test all services."""
+
+        MouseDirector.user = self.user.id
+        MouseDirector.mouse = self.mouse.id
+
+        OpenEphys.folder = self.session.folder
+
+        NewScaleCoordinateRecorder.log_root = self.session.npexp_path
+        NewScaleCoordinateRecorder.log_name = self.platform_json.path.name
+
+        SessionCamstim.labtracks_mouse_id = self.mouse.id
+        SessionCamstim.lims_user_id = self.user.id
+
+        self.configure_services()
+
+        super().initialize_and_test_services()
+
+    def update_state(self) -> None:
+        "Store useful but non-essential info."
+        self.mouse.state['last_session'] = self.session.id
+        self.mouse.state['last_PsyCode_session'] = str(self.workflow)
+        if self.mouse == 366122:
+            return
+        match self.workflow:
+            case PsyCodeSession.PRETEST:
+                return
+            case PsyCodeSession.HAB:
+                self.session.project.state['latest_hab'] = self.session.id
+            case PsyCodeSession.EPHYS:
+                self.session.project.state['latest_ephys'] = self.session.id
+                self.session.project.state['sessions'] = self.session.project.state.get('sessions', []) + [self.session.id]
+
+    def run_stim(self) -> None:
+
+        self.update_state()
+
+        if not SessionCamstim.is_ready_to_start():
+            raise RuntimeError("SessionCamstim is not ready to start.")
+
+        np_logging.web(f'PsyCode_{self.workflow.name.lower()}').info(f"Started session {self.mouse.mtrain.stage['name']}")
+        SessionCamstim.start()
+
+        with contextlib.suppress(Exception):
+            while not SessionCamstim.is_ready_to_start():
+                time.sleep(2.5)
+
+        if isinstance(SessionCamstim, Finalizable):
+            SessionCamstim.finalize()
+
+        with contextlib.suppress(Exception):
+            np_logging.web(f'PsyCode_{self.workflow.name.lower()}').info(f"Finished session {self.mouse.mtrain.stage['name']}")
+
+
+    def copy_data_files(self) -> None: 
+        super().copy_data_files()
+
+        # When all processing completes, camstim Agent class passes data and uuid to
+        # /camstim/lims BehaviorSession class, and write_behavior_data() writes a
+        # final .pkl with default name YYYYMMDDSSSS_mouseID_foragingID.pkl
+        # - if we have a foraging ID, we can search for that
+        if None == (stim_pkl := next(self.session.npexp_path.glob(f'{self.session.date:%y%m%d}*_{self.session.mouse}_*.pkl'), None)):
+            logger.warning('Did not find stim file on npexp matching the format `YYYYMMDDSSSS_mouseID_foragingID.pkl`')
+            return
+        assert stim_pkl
+        if not self.session.platform_json.foraging_id:
+            self.session.platform_json.foraging_id = stim_pkl.stem.split('_')[-1]
+        new_stem = f'{self.session.folder}.stim'
+        logger.debug(f'Renaming stim file copied to npexp: {stim_pkl} -> {new_stem}')
+        stim_pkl = stim_pkl.rename(stim_pkl.with_stem(new_stem))
+
+        # remove other stim pkl, which is nearly identical, if it was also copied
+        for pkl in self.session.npexp_path.glob('*.pkl'):
+            if (
+                self.session.folder not in pkl.stem
+                and 
+                abs(pkl.stat().st_size - stim_pkl.stat().st_size) < 1e6
+            ):
+                logger.debug(f'Deleting extra stim pkl copied to npexp: {pkl.stem}')
+                pkl.unlink()
+
+
+def validate_selected_workflow(session: PsyCodeSession, mouse: np_session.Mouse) -> None:
+    for workflow in ('hab', 'ephys'):
+        if (
+            workflow in session.value.lower()
+            and workflow not in mouse.mtrain.stage['name'].lower()
+        ) or (
+            session.value.lower() == 'ephys' and 'hab' in mouse.mtrain.stage['name'].lower()
+        ):
+            raise ValueError(f"Workflow selected ({session.value}) does not match MTrain stage ({mouse.mtrain.stage['name']}): please check cells above.")
+
+
+class Hab(PsyCodeMixin, np_workflows.PipelineHab):
+    def __init__(self, *args, **kwargs):
+        self.services = (
+            MouseDirector,
+            Sync,
+            VideoMVR,
+            self.imager,
+            NewScaleCoordinateRecorder,
+            SessionCamstim,
+        )
+        super().__init__(*args, **kwargs)
+
+
+class Ephys(PsyCodeMixin, np_workflows.PipelineEphys):
+    def __init__(self, *args, **kwargs):
+        self.services = (
+            MouseDirector,
+            Sync,
+            VideoMVR,
+            self.imager,
+            NewScaleCoordinateRecorder,
+            SessionCamstim,
+            OpenEphys,
+        )
+        super().__init__(*args, **kwargs)
+
+
+# --------------------------------------------------------------------------------------
+
+
+def new_experiment(
+    mouse: int | str | np_session.Mouse,
+    user: str | np_session.User,
+    workflow: PsyCodeSession,
+) -> Ephys | Hab:
+    """Create a new experiment for the given mouse and user."""
+    match workflow:
+        case PsyCodeSession.PRETEST | PsyCodeSession.EPHYS:
+            experiment = Ephys(mouse, user)
+        case PsyCodeSession.HAB:
+            experiment = Hab(mouse, user)
+        case _:
+            raise ValueError(f"Invalid workflow type: {workflow}")
+    experiment.workflow = workflow
+
+    with contextlib.suppress(Exception):
+        np_logging.web(f'PsyCode_{experiment.workflow.name.lower()}').info(f"{experiment} created")
+
+    return experiment
+

src/np_workflows/experiments/openscope_psycode/psycode_workflow_widget.py

@@ -0,0 +1,83 @@
+import configparser
+import contextlib
+import copy
+import enum
+import functools
+from typing import ClassVar, Literal, NamedTuple, NoReturn, Optional, TypedDict
+
+import IPython.display
+import ipywidgets as ipw
+import np_config
+import np_logging
+import np_session
+import np_workflows
+from pyparsing import Any
+
+from np_workflows.experiments.openscope_psycode.main_psycode_pilot import PsyCodeSession
+
+global_state = {}
+"""Global variable for persisting widget states."""
+
+# for widget, before creating a experiment --------------------------------------------- #
+
+class SelectedSession:
+    def __init__(self, session: str | PsyCodeSession, mouse: str | int | np_session.Mouse):
+        if isinstance(session, str):
+            session = PsyCodeSession(session)
+        self.session = session
+        self.mouse = str(mouse)
+
+    def __repr__(self) -> str:
+        return f"{self.__class__.__name__}({self.session}, {self.mouse})"
+
+
+def PsyCode_workflow_widget(
+    mouse: str | int | np_session.Mouse,
+) -> SelectedSession:
+    """Select a stimulus session (hab, pretest, ephys) to run.
+
+    An object with mutable attributes is returned, so the selected session can be
+    updated along with the GUI selection. (Preference would be to return an enum
+    directly, and change it's value, but that doesn't seem possible.)
+
+    """
+
+    selection = SelectedSession(PsyCodeSession.PRETEST, mouse)
+
+    session_dropdown = ipw.Select(
+        options=tuple(_.value for _ in PsyCodeSession),
+        description="Session",
+    )
+
+    def update_selection():
+        selection.__init__(str(session_dropdown.value), str(mouse))
+
+    if (previously_selected_value := global_state.get('selected_session')):
+        session_dropdown.value = previously_selected_value
+        update_selection()
+
+    console = ipw.Output()
+    with console:
+        if last_session := np_session.Mouse(selection.mouse).state.get('last_PsyCode_session'):
+            print(f"{mouse} last session: {last_session}")
+        print(f"Selected: {selection.session}")
+
+    def update(change):
+        if change["name"] != "value":
+            return
+        if (options := getattr(change["owner"], "options", None)) and change[
+            "new"
+        ] not in options:
+            return
+        if change["new"] == change["old"]:
+            return
+        update_selection()
+        with console:
+            print(f"Selected: {selection.session}")
+        global_state['selected_session'] = selection.session.value
+
+    session_dropdown.observe(update, names='value')
+
+    IPython.display.display(ipw.VBox([session_dropdown, console]))
+
+    return selection