refactor/fix: splitting out project view functionality for future exp…

…ansion

src/aind_qc_portal/docdb/database.py

@@ -129,8 +129,8 @@ def get_all():
 
 
 @pn.cache(ttl=TIMEOUT_1H)
-def get_project(project: str):
-    filter = {"data_description.project_name": project}
+def get_project(project_name: str):
+    filter = {"data_description.project_name": project_name}
     limit = 0
     paginate_batch_size = 500
     response = client.retrieve_docdb_records(
@@ -145,6 +145,9 @@ def get_project(project: str):
             "session.session_start_time": 1,
             "data_description.data_level": 1,
             "data_description.project_name": 1,
+            "rig.rig_id": 1,
+            "session.experimenter_full_name": 1,
+            "quality_control": 1,
         },
         limit=limit,
         paginate_batch_size=paginate_batch_size,
@@ -153,6 +156,36 @@ def get_project(project: str):
     return response
 
 
+@pn.cache(ttl=TIMEOUT_1H)
+def get_project_custom(project_name: str, fields: list):
+    """Get all records that match a project name, with custom fields
+
+    Parameters
+    ----------
+    project_name : str
+    fields : list
+        List of fields to retain from DocDB record
+
+    Returns
+    -------
+    list
+        List of dictionaries containing the fields requested
+    """
+    filter = {"data_description.project_name": project_name}
+    limit = 0
+    paginate_batch_size = 500
+    response = client.retrieve_docdb_records(
+        filter_query=filter,
+        projection={
+            "_id": 1,
+        } | {field: 1 for field in fields},
+        limit=limit,
+        paginate_batch_size=paginate_batch_size,
+    )
+
+    return response
+
+
 @pn.cache
 def get_subjects():
     filter = {

src/aind_qc_portal/panel/metric.py

@@ -168,7 +168,7 @@ def metric_panel(self):
             options=["Pass", "Fail", "Pending"],
             name="Metric status",
         )
-        
+
         if pn.state.user == "guest":
             self.state_selector.disabled = True
         else:

src/aind_qc_portal/panel/quality_control.py

@@ -213,15 +213,20 @@ def status_panel(self):
 
     def panel(self):
         """Build a Panel object representing this QC action"""
-        if not self._has_data or not self._data:
-            return pn.widgets.StaticText(value="No QC object available")
 
         # build the header
         md = f"""
 <span style="font-size:14pt">Quality control for {self.asset_name}</span>
 """
         header = pn.pane.Markdown(md)
 
+        if not self._has_data or not self._data:
+            return pn.Row(
+                pn.HSpacer(),
+                pn.Column(header,
+                          pn.widgets.StaticText(value="No QC object available", styles={"font-size": "22pt"}), styles=OUTER_STYLE),
+                pn.HSpacer())
+
         # build the display box: this shows the current state in DocDB of this asset
         # if any evaluations are failing, we'll show a warning
         failing_eval_str = ""

src/aind_qc_portal/projects/dataset.py

@@ -0,0 +1,114 @@
+import pandas as pd
+import param
+
+from aind_qc_portal.docdb.database import get_project, get_project_custom
+from aind_qc_portal.utils import format_link, qc_color
+from aind_data_schema.core.quality_control import QualityControl, Status
+
+class ProjectDataset(param.Parameterized):
+    """Generic dataset class, loads default data for all projects"""
+    subject_filter = param.String(default="")
+
+    def __init__(self, project_name: str):
+        """Create a ProjectDataset object"""
+
+        self.project_name = project_name
+        self._df = None
+        self.exposed_columns = [
+            "subject_id", "Date", "name", "Operator", "S3 link", "Status", "Subject view", "QC view", "session_type", "raw"
+        ]
+        self._get_assets()
+
+    def _get_assets(self):
+        """Get all assets with this project name"""
+        print(self.project_name)
+        records = get_project(self.project_name)
+
+        data = []
+        for record in records:
+            subject_id = record.get('subject', {}).get('subject_id')
+
+            # rig, operator, QC notes get bubbled up? qc status,
+            # custom genotype mapping. Do this for learning-mfish
+
+            # reconstruct the QC object, if possible
+            if record.get('quality_control'):
+                qc = QualityControl(**record.get('quality_control'))
+            else:
+                qc = None
+
+            record_data = {
+                '_id': record.get('_id'),
+                'raw': record.get('data_description', {}).get('data_level') == 'raw',
+                'project_name': record.get('data_description', {}).get('project_name'),
+                'location': record.get('location'),
+                'name': record.get('name'),
+                'session_start_time': record.get('session', {}).get('session_start_time'),
+                'session_type': record.get('session', {}).get('session_type'),
+                'subject_id': subject_id,
+                'operator': list(record.get('session', {}).get('experimenter_full_name')),
+                'Status': qc.status().value if qc else "No QC",
+            }
+            data.append(record_data)
+
+        if len(data) == 0:
+            self._df = None
+            return
+
+        self._df = pd.DataFrame(data)
+        self._df["timestamp"] = pd.to_datetime(self._df["session_start_time"], format='mixed', utc=True)
+        self._df["Date"] = self._df["timestamp"].dt.strftime("%Y-%m-%d %H:%M:%S")
+        self._df["S3 link"] = self._df["location"].apply(lambda x: format_link(x, text="S3 link"))
+        self._df["Subject view"] = self._df["_id"].apply(lambda x: format_link(f"/qc_asset_app?id={x}"))
+        self._df["qc_link"] = self._df["_id"].apply(lambda x: f"/qc_app?id={x}")
+        self._df["QC view"] = self._df.apply(lambda row: format_link(row["qc_link"]), axis=1)
+        self._df["Operator"] = self._df["operator"].apply(lambda x: ", ".join(x))
+        self._df.sort_values(by="timestamp", ascending=True, inplace=True)
+        self._df.sort_values(by="subject_id", ascending=True, inplace=True)
+
+    def filtered_data(self):
+        if self.subject_filter:
+            filtered_df = self._df[self._df["subject_id"].str.contains(self.subject_filter, case=False, na=False)]
+        else:
+            filtered_df = self._df
+
+        return filtered_df[self.exposed_columns]
+
+    @property
+    def data(self):
+
+        return self.filtered_data()[self.exposed_columns].style.map(qc_color, subset=["Status"])
+
+
+    @property
+    def timestamp_data(self):
+        if self.subject_filter:
+            filtered_df = self._df[self._df["subject_id"].str.contains(self.subject_filter, case=False, na=False)]
+        else:
+            filtered_df = self._df
+
+        return filtered_df[["timestamp"]]
+
+
+class LearningmFishDataset(ProjectDataset):
+
+    def __init__(self, project_name: str):
+        if project_name != "Learning mFISH-V1omFISH":
+            raise ValueError("This class is only for Learning mFISH-V1omFISH")
+
+        super().__init__(project_name=project_name)
+
+        self._get_mfish_assets()
+
+    def _get_mfish_assets(self):
+        """Load additional information needed for the Learning mFISH-V1omFISH project
+
+        Extra data should be appended to the self._df dataframe and then needs to be added to the
+        list of exposed columns.
+        """
+        data = get_project_custom(self.project_name, [""])
+
+
+mapping = {
+    "Learning mFISH-V1omFISH": LearningmFishDataset,
+}

src/aind_qc_portal/projects/project_view.py

@@ -0,0 +1,103 @@
+from typing import Optional
+import pandas as pd
+import panel as pn
+import altair as alt
+
+from aind_qc_portal.projects.dataset import mapping, ProjectDataset
+from aind_qc_portal.utils import df_timestamp_range, OUTER_STYLE
+from aind_qc_portal.utils import OUTER_STYLE, AIND_COLORS
+
+
+class ProjectView():
+
+    def __init__(self, project_name: str):
+        cls = mapping.get(project_name, ProjectDataset)
+        self.dataset = cls(project_name=project_name)
+        self.project_name = project_name
+
+    @property
+    def has_data(self):
+        return self.dataset.data is not None
+
+    def get_subjects(self):
+        if not self.has_data:
+            return []
+
+        return self.dataset._df["subject_id"].unique()
+
+    def get_data(self) -> Optional[pd.DataFrame]:
+        if not self.has_data:
+            return None
+
+        return self.dataset.filtered_data()
+
+    def get_data_styled(self):
+        if not self.has_data:
+            return None
+
+        return self.dataset.data
+
+    def history_panel(self):
+        """Create a plot showing the history of this asset, showing how assets were derived from each other"""
+        if not self.has_data:
+            return pn.widgets.StaticText(
+                value=f"No data found for project: {self.project_name}"
+            )
+
+        # Calculate the time range to show on the x axis
+        (min_range, max_range, range_unit, format) = df_timestamp_range(
+            self.dataset.timestamp_data
+        )
+
+        chart = (
+            alt.Chart(self.get_data())
+            .mark_bar()
+            .encode(
+                x=alt.X(
+                    "Date:T",
+                    title="Time",
+                    scale=alt.Scale(domain=[min_range, max_range]),
+                    axis=alt.Axis(format=format, tickCount=range_unit),
+                ),
+                y=alt.Y("subject_id:N", title="Subject ID"),
+                tooltip=[
+                    "subject_id",
+                    "session_type",
+                    "Date",
+                ],
+                color=alt.Color("subject_id:N"),
+                href="qc_link:N",
+            )
+            .properties(width=900)
+        )
+
+        return pn.pane.Vega(chart, sizing_mode="stretch_width", styles=OUTER_STYLE)
+
+    def panel(self) -> pn.Column:
+        """Return panel object"""
+
+        md = f"""
+        <h1 style="color:{AIND_COLORS["dark_blue"]};">
+            {self.project_name}
+        </h1>
+        <b>{len(self.dataset.filtered_data())}</b> data assets are associated with this project.
+        """
+
+        header = pn.pane.Markdown(md, width=1000, styles=OUTER_STYLE)
+
+        chart_pane = self.history_panel()
+
+        df_pane = pn.pane.DataFrame(self.get_data_styled(), width=950, escape=False, index=False)
+
+        def update_subject_filter(event):
+            self.dataset.subject_filter = event.new
+            df_pane.object = self.get_data()
+
+        subject_filter = pn.widgets.Select(name="Subject filter", options=list(self.get_subjects()))
+        subject_filter.param.watch(update_subject_filter, "value")
+
+        df_col = pn.Column(subject_filter, df_pane, styles=OUTER_STYLE)
+
+        col = pn.Column(header, chart_pane, df_col)
+
+        return col