Feature/dash cohort

dash/.dockerignore

@@ -0,0 +1,2 @@
+venv/
+.idea/

dash/.gitignore

@@ -0,0 +1 @@
+venv/

dash/Dockerfile

@@ -0,0 +1,10 @@
+FROM python:3.10
+WORKDIR /code
+
+COPY requirements.txt /
+RUN pip install -r /requirements.txt
+COPY ./ ./
+
+EXPOSE 8050
+
+CMD ["python", "./app.py"]

dash/app.py

@@ -0,0 +1,75 @@
+import logging
+import os
+
+import dash
+import dash_bootstrap_components as dbc
+from dash import Dash, Output, Input, dcc
+from dash import html
+
+from util import get_authz
+
+external_stylesheets = [dbc.themes.SKETCHY]
+
+logger = logging.getLogger('dash')
+
+print('DASH_URL_BASE_PATHNAME', os.environ.get('DASH_URL_BASE_PATHNAME', default="~not set~"))
+app = Dash(__name__, external_stylesheets=external_stylesheets, use_pages=True)
+
+
+@app.server.before_request
+def check_privileges():
+    """Do this before every call"""
+    get_authz()
+
+
+# Clientside callback: refresh token by calling fence's /user endpoint
+app.clientside_callback(
+    """
+    // Call fence's /user endpoint, parse the response and update the profile 
+    async function(n_intervals, data) {
+        const response = await fetch(location.origin + '/user/user');
+        if (!response.ok) {
+            console.log('error retrieving user', response )
+            return 'Profile (unauthorized)';
+        } else {
+            const user = await response.json();
+            console.log('clientside_callback you are logged in as:', user.username);
+            return 'Profile (' + user.username + ')';
+        }        
+    }
+    """,
+    Output('nav_item-profile', 'children'),
+    Input('clientside-interval', 'n_intervals')
+)
+
+
+app.layout = dbc.Container([
+    html.H1("ACED", className="display-3"),
+    html.P(
+        "A simple dash app.",
+        className="lead",
+    ),
+
+    dbc.Nav(
+        [
+            dbc.NavItem(
+                dbc.NavLink(f"{page['name']}", href=page["relative_path"], id=f'nav_item-{page["name"].lower()}')
+            )
+            for page in dash.page_registry.values()
+        ]
+    ),
+
+    html.Hr(className="my-2"),
+    # define a timed client side action here
+    dcc.Interval(
+        id='clientside-interval',
+        n_intervals=0,
+        interval=60 * 1000  # in milliseconds check every minute
+    ),
+    # other page contents goes here
+    dash.page_container
+])
+
+
+if __name__ == '__main__':
+    app.run_server(host="0.0.0.0",  debug=True)  #

dash/figures/histogram.py

@@ -0,0 +1,123 @@
+import logging
+
+import dash_bootstrap_components as dbc
+import pandas as pd
+import plotly.express as px
+from dash import html, dcc
+
+logger = logging.getLogger('dash')
+
+
+def histogram_figures(histograms):
+    """An array of histograms figures, dataframes and names.
+    @return fig[], df[], name[]"""
+
+    dfs = []
+    figs = []
+    names = []
+
+    assert len(histograms._aggregation.keys()) == 1  # noqa
+    entity = list(histograms._aggregation.keys())[0]  # noqa
+    for k, v in histograms._aggregation[entity].items():  # noqa
+        v.name = k
+        df = pd.DataFrame(v.histogram)
+        fig = px.histogram(df, x="key", y="count", title=v.name, log_y=len(v.histogram) > 1,
+                           id={
+                               'index': f"{entity}-{k}",
+                               'type': 'query-parameter'
+                           }
+                           )
+        fig.update_layout(legend=dict(orientation="h", title=None),
+                          yaxis_title=None, xaxis_title=None,
+                          plot_bgcolor='rgba(0,0,0,0)'
+                          )
+
+        figs.append(fig)
+        dfs.append(df)
+        names.append(v.name)
+
+    return figs, dfs, names
+
+
+def histogram_selects(histograms):
+    """An array of histograms checklists dataframes, and names.
+    @return fig[], df[], name[]"""
+
+    dfs = []
+    checklists = []
+    names = []
+
+    assert len(histograms._aggregation.keys()) == 1  # noqa
+    entity = list(histograms._aggregation.keys())[0]  # noqa
+    for k, v in histograms._aggregation[entity].items():  # noqa
+        v.name = k
+        if not any([isinstance(h.key, str) for h in v.histogram]):
+            continue
+        df = pd.DataFrame(v.histogram)
+        checklist = dcc.Checklist(
+            id={
+                'index': f"{entity}-{k}",
+                'type': 'query-parameter'
+            },
+            options=[
+                {
+                    'value': str(h.key),
+                    'label': html.Div(
+                         [
+                             html.Div(h.key),
+                             dbc.Badge(h.count, className="ms-1 term-count", color="info",
+                                       id={
+                                           'index': f"{entity}-{k}-{h.key}",
+                                           'type': 'term-count'
+                                       })
+                         ],
+                         style={'display': 'inline-flex', 'paddingLeft': '2em'}
+                    )
+                }
+                for h in v.histogram
+            ],
+            labelStyle={'display': 'flex'}
+        )
+
+        checklists.append(checklist)
+        dfs.append(df)
+        names.append(v.name)
+
+    return checklists, dfs, names
+
+
+def histogram_sliders(histograms):
+    """An array of histograms sliders dataframes, and names.
+    @return slider[], df[], name[]"""
+
+    dfs = []
+    sliders = []
+    names = []
+
+    assert len(histograms._aggregation.keys()) == 1  # noqa
+    entity = list(histograms._aggregation.keys())[0]  # noqa
+    for k, v in histograms._aggregation[entity].items():  # noqa
+        v.name = k
+        if any([isinstance(h.key, str) for h in v.histogram]):
+            continue
+        if not any([isinstance(h.key, list) for h in v.histogram]):
+            continue
+
+        assert len(v.histogram) == 1
+        df = pd.DataFrame(v.histogram)
+        min_ = int(v.histogram[0].key[0])
+        max_ = int(v.histogram[0].key[1])
+        # value = max_,
+        sliders.append(
+            dcc.RangeSlider(min_, max_, int((max_ - min_) / 10),
+                            id={
+                                'index': f"{entity}-{k}",
+                                'type': 'query-parameter'
+                            }
+                            )
+        )
+
+        dfs.append(df)
+        names.append(v.name)
+
+    return sliders, dfs, names

dash/figures/project.py

@@ -0,0 +1,31 @@
+
+import pandas as pd
+import plotly.express as px
+from inflection import titleize, pluralize
+from models.project import project_detail_counts
+
+
+def counts():
+    """Horizontal bar graph.
+    @return fig, df"""
+
+    project_counts = list(project_detail_counts())
+    flattened = []
+    for p in project_counts:
+        for k, v in p.items():
+            f = {'name': p.project[0].name}
+            if k == 'project':
+                continue
+            f['entity'] = titleize(pluralize(k.replace('_count', '')))
+            f['count'] = v
+            flattened.append(f)
+    df = pd.DataFrame(flattened)
+    fig = px.bar(df, x="count", y="entity", color='name', orientation='h',
+                 hover_data=["name", "count"],
+                 height=400,
+                 log_x=True)
+    fig.update_layout(legend=dict(orientation="h", title=None), yaxis_title=None, xaxis_title=None,
+                      # paper_bgcolor='rgba(0,0,0,0)',
+                      plot_bgcolor='rgba(0,0,0,0)'
+                      )
+    return fig, df

dash/models/file.py

@@ -0,0 +1,126 @@
+
+from dotwiz import DotWiz
+from util import get_guppy_service
+
+import logging
+logger = logging.getLogger('dash')
+
+
+def get_files(dot_notation=True, variables={"filter": {"AND": []}, "sort": []}):
+    """Fetch histogram of counts for all projects.
+    @param variables: a graphql filter and sort
+    @type dot_notation: bool render results as a lightweight class"""
+    query = """
+        query ($sort: JSON,$filter: JSON,) {
+            file (accessibility: all, offset: 0, first: 1000, , sort: $sort, filter: $filter,) {
+                file_id  
+                patient_id    
+                file_category    
+                file_name    
+                file_size    
+                object_id    
+            }
+            _aggregation {
+              file (filter: $filter, accessibility: all) {
+                _totalCount
+              }
+            }
+        }    
+    """
+    guppy_service = get_guppy_service()
+    data = guppy_service.graphql_query(query, variables=variables)['data']
+    data = DotWiz(data)
+    return [f for f in data.file]
+
+
+def get_file_histograms(dot_notation=True, variables={"filter": {"AND": []}}):
+    """Fetch histogram of counts for all projects.
+    @param variables: a graphql filter
+    @type dot_notation: bool render results as a lightweight class"""
+
+    histogram_query = """
+    query ($filter: JSON) {
+      _aggregation {
+        file (filter: $filter, filterSelf: false, accessibility: all) {
+
+    project_id {
+      histogram {
+        key
+        count
+      }
+    },
+    file_category {
+      histogram {
+        key
+        count
+      }
+    },
+    data_type {
+      histogram {
+        key
+        count
+      }
+    },
+    data_format {
+      histogram {
+        key
+        count
+      }
+    },
+    patient_id {
+      histogram {
+        key
+        count
+      }
+    },
+    patient_gender {
+      histogram {
+        key
+        count
+      }
+    },
+    patient_disability_adjusted_life_years {
+      histogram {
+        key
+        count
+      }
+    },
+    patient_ombCategory {
+      histogram {
+        key
+        count
+      }
+    },
+    patient_ombCategory_detail {
+      histogram {
+        key
+        count
+      }
+    },
+    patient_us_core_birthsex {
+      histogram {
+        key
+        count
+      }
+    },
+    patient_quality_adjusted_life_years {
+      histogram {
+        key
+        count
+      }
+    },
+    patient_maritalStatus_coding_0_display {
+      histogram {
+        key
+        count
+      }
+    }
+        }
+      }
+    }    
+    """
+    guppy_service = get_guppy_service()
+    data = guppy_service.graphql_query(histogram_query, variables=variables)['data']
+    if dot_notation:
+        return DotWiz(data)
+    return data