Merge pull request #7 from domonik/dev

Dev

RAPDOR/tests/test_datastructures.py

@@ -70,11 +70,11 @@ def multi_intensities(intensities):
 
 def drop_replicates(design, rnase_rep, ctrl_rep):
     if rnase_rep > 0:
-        rnase = design[design["Treatment"] == "RNase"].groupby("Replicate").apply(lambda x: list(x.index)).sample(n=rnase_rep).sum()
+        rnase = design[design["Treatment"] == "RNase"].groupby("Replicate").apply(lambda x: list(x.index), include_groups=False).sample(n=rnase_rep).sum()
     else:
         rnase = []
     if ctrl_rep > 0:
-        ctrl = design[design["Treatment"] == "Control"].groupby("Replicate").apply(lambda x: list(x.index)).sample(n=ctrl_rep).sum()
+        ctrl = design[design["Treatment"] == "Control"].groupby("Replicate").apply(lambda x: list(x.index), include_groups=False).sample(n=ctrl_rep).sum()
     else:
         ctrl = []
     rnase = design.loc[rnase]
@@ -147,14 +147,13 @@ def test_different_columns(intensities, design):
 
 
 @pytest.mark.parametrize(
-    "normalize,kernel_size,distance,permanova,nr_samples, distance_cutoff",
+    "normalize,kernel_size,distance,permanova,nr_samples",
     [
-        (True, 3, "Jensen-Shannon-Distance", True, 10, 0.1),
-        (True, 3, "Jensen-Shannon-Distance", True, 10, 0),
-        (False, None, None, False, None, None)
+        (True, 3, "Jensen-Shannon-Distance", True, 10),
+        (False, None, None, False, None)
     ]
 )
-def test_serialization(normalize, kernel_size, distance, permanova, nr_samples, distance_cutoff, rapdordata):
+def test_serialization(normalize, kernel_size, distance, permanova, nr_samples, rapdordata):
     if normalize:
         rapdordata.normalize_array_with_kernel(kernel_size)
     if distance:
@@ -164,7 +163,7 @@ def test_serialization(normalize, kernel_size, distance, permanova, nr_samples,
         rapdordata.rank_table(['Mean Distance', "ANOSIM R"], [False, False])
 
 
-        rapdordata.calc_permanova_p_value(10, threads=1, distance_cutoff=distance_cutoff)
+        rapdordata.calc_permanova_p_value(10, threads=1)
     s = rapdordata.to_jsons()
     loaded_data = RAPDORData.from_json(s)
     assert loaded_data == rapdordata

RAPDOR/visualize/assets/additional.js

@@ -854,10 +854,10 @@ window.dash_clientside = Object.assign({}, window.dash_clientside, {
             ["Figure Creation", null, "/figure_factory", false, simulateClickFormCheckInput(0)],
             ["Figure Creation", ["ff-tut-preview"], "/figure_factory", false, simulateClickID("ff-default", 1)],
             ["Figure Creation", ["ff-tut-preview"], "/figure_factory", false, multiClickInOne([simulateClickFormCheckInput(2), simulateClickID("ff-default", null) ])],
-            ["Dimension Reduction", null, "/analysis", true],
-            ["Dimension Reduction", ["distribution-panel", "dim-red-tab", "dim-red-tut"], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
-            ["Dimension Reduction", ["dim-red-tut", "distribution-panel", "dim-red-tab", ], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
-            ["Dimension Reduction", ["dim-red-tut", "distribution-panel", "dim-red-tab", ], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
+            ["Bubble Plot", null, "/analysis", true],
+            ["Bubble Plot", ["distribution-panel", "dim-red-tab", "dim-red-tut"], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
+            ["Bubble Plot", ["dim-red-tut", "distribution-panel", "dim-red-tab", ], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
+            ["Bubble Plot", ["dim-red-tut", "distribution-panel", "dim-red-tab", ], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
             ["Finish",null, null, false, simulateClickID("dim-red-tab", "dim-red-tut")],
 
 
@@ -901,10 +901,10 @@ window.dash_clientside = Object.assign({}, window.dash_clientside, {
             ["Figure Creation", null, "/figure_factory", false, simulateClickFormCheckInput(0)],
             ["Figure Creation", ["ff-tut-preview"], "/figure_factory", false, simulateClickID("ff-default", 1)],
             ["Figure Creation", ["ff-tut-preview"], "/figure_factory", false, multiClickInOne([simulateClickFormCheckInput(2), simulateClickID("ff-default", null) ])],
-            ["Dimension Reduction", null, "/analysis", true],
-            ["Dimension Reduction", ["distribution-panel", "dim-red-tab", "dim-red-tut"], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
-            ["Dimension Reduction", ["dim-red-tut", "distribution-panel", "dim-red-tab", ], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
-            ["Dimension Reduction", ["dim-red-tut", "distribution-panel", "dim-red-tab", ], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
+            ["Bubble Plot", null, "/analysis", true],
+            ["Bubble Plot", ["distribution-panel", "dim-red-tab", "dim-red-tut"], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
+            ["Bubble Plot", ["dim-red-tut", "distribution-panel", "dim-red-tab", ], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
+            ["Bubble Plot", ["dim-red-tut", "distribution-panel", "dim-red-tab", ], "/analysis", true, simulateClickID("dim-red-tab", "dim-red-tut")],
             ["Finish",null, null, false, simulateClickID("dim-red-tab", "dim-red-tut")],
 
         ]

RAPDOR/visualize/assets/tutorial.json

@@ -33,10 +33,10 @@
   "We now select rows to plot from the table. You can also select some more. In the next step we will visit the Figure Factory.",
   "Here is the Figure Factory. As you can see the plot contains the distributions from the selected rows. But it does not look nice yet. We can improve that.",
   "Usually the default settings are fine for plots. However, you can tweak some settings using the Input fields.",
-  "This is a dimension reduction plot displaying all rows from the table but highlighting the selected ones. We can also find it on the analysis page. Let`s have a look.",
-  "Now that we are back at the analysis page we can visit the dimension reduction tab.",
-  "Here you see the interactive dimension reduction. If you hover over the datapoints, the corresponding distribution is shown in the distribution panel. Just try yourself.",
-  "The <b>relative fraction shift</b> is an indicator of how many fractions the protein shifted. The <b>relative distribution change</b> shows whether a protein was more uniformly distributed after treatment (positive values) or accumulated in a less fractions (negative values).",
+  "This is a bubble plot plot displaying all rows from the table but highlighting the selected ones. We can also find it on the analysis page. Let`s have a look.",
+  "Now that we are back at the analysis page we can visit the bubble plot tab.",
+  "Here you see the interactive bubble plot. If you hover over the datapoints, the corresponding distribution is shown in the distribution panel. Just try yourself.",
+  "The <b>relative fraction shift</b> is an indicator of how many fractions the protein shifted. The <b>relative distribution change</b> shows whether a protein was more uniformly distributed after treatment (negative values) or accumulated in a less fractions (positive values).",
   "You can also filter this plot via a cutoff at the right. E.g. you can filter for data points with a specific Rank range. Just try yourself. First set the Cutoff Type to Rank and then use the slider to restrict the range.",
   "This is it. You finished the tutorial. Now you can go to the upload page and upload your own data."
 ]

RAPDOR/visualize/assets/tutorialDisplayMode.json

@@ -24,10 +24,10 @@
   "We now select rows to plot from the table. You can also select some more. In the next step we will visit the Figure Factory.",
   "Here is the Figure Factory. As you can see the plot contains the distributions from the selected rows. But it does not look nice yet. We can improve that.",
   "Usually the default settings are fine for plots. However, if you want to tweak some settings you need to use a local copy which enables the input fields.",
-  "This is a dimension reduction plot displaying all rows from the table but highlighting the selected ones. We can also find it on the analysis page. Let`s have a look.",
-  "Now that we are back at the analysis page we can visit the dimension reduction tab.",
-  "Here you see the interactive dimension reduction. If you hover over the datapoints, the corresponding distribution is shown in the distribution panel. Just try yourself.",
-  "The <b>relative fraction shift</b> is an indicator of how many fractions the protein shifted. The <b>relative distribution change</b> shows whether a protein was more uniformly distributed after treatment (positive values) or accumulated in a less fractions (negative values).",
+  "This is a bubble plot plot displaying all rows from the table but highlighting the selected ones. We can also find it on the analysis page. Let`s have a look.",
+  "Now that we are back at the analysis page we can visit the bubble plot tab.",
+  "Here you see the interactive bubble plot. If you hover over the datapoints, the corresponding distribution is shown in the distribution panel. Just try yourself.",
+  "The <b>relative fraction shift</b> is an indicator of how many fractions the protein shifted. The <b>relative distribution change</b> shows whether a protein was more uniformly distributed after treatment (negative values) or accumulated in a less fractions (positive values).",
   "You can also filter this plot via a cutoff at the right. E.g. you can filter for data points with a specific Rank range. Just try yourself. First set the Cutoff Type to Rank and then use the slider to restrict the range.",
   "This is it. You finished the tutorial."
 

RAPDOR/visualize/callbacks/plotCallbacks.py

@@ -5,7 +5,7 @@
 from dash.exceptions import PreventUpdate
 from plotly import graph_objs as go
 from RAPDOR.plots import plot_replicate_distribution, plot_distribution, plot_barcode_plot, plot_heatmap, \
-    plot_dimension_reduction, empty_figure, DEFAULT_TEMPLATE, DEFAULT_TEMPLATE_DARK, plot_bars
+    plot_dimension_reduction, empty_figure, DEFAULT_TEMPLATE, DEFAULT_TEMPLATE_DARK, plot_bars, plot_distance_and_var
 from dash_extensions.enrich import Serverside, callback
 from RAPDOR.datastructures import RAPDORData
 import logging
@@ -56,8 +56,7 @@ def update_distribution_plot(key, recomp, primary_color, secondary_color, replic
             if rapdordata.categorical_fraction:
                 fig = plot_bars(array, rapdordata.internal_design_matrix, x=rapdordata.fractions, offset=i,
                                 colors=colors, yname=yname)
-                if night_mode:
-                    fig.update_traces(error_y=dict(color="white"), marker=dict(line=dict(width=1, color="white")))
+                fig.update_traces(error_y=dict(width=15))
             else:
                 fig = plot_distribution(array, rapdordata.internal_design_matrix, offset=i, colors=colors, show_outliers=True, yname=yname)
         if not night_mode:
@@ -106,6 +105,12 @@ def update_distribution_plot(key, recomp, primary_color, secondary_color, replic
         legend=dict(font=dict(size=14)),
         legend2=dict(font=dict(size=14))
     )
+    if isinstance(rapdordata.fractions[0], str):
+        fig.update_xaxes(
+            tickvals=list(range(len(rapdordata.fractions))),
+            ticktext=[val.replace(" ", "<br>").replace("<br>&<br>", " &<br>") for val in rapdordata.fractions],
+            tickmode="array"
+        )
 
     fig.update_xaxes(dtick=1, title=None)
     fig.update_xaxes(fixedrange=True)
@@ -241,7 +246,7 @@ def calc_clusters(
         if rapdordata.cluster_features is None:
             rapdordata.calc_distribution_features()
             logger.info("Calculated Cluster Features")
-            logger.info("Running Dimension Reduction - because cluster features changed")
+            logger.info("Running Bubble Plot - because cluster features changed")
         if cluster_method is not None:
             if cluster_method == "HDBSCAN":
                 kwargs = dict(min_cluster_size=hdb_min_cluster_size, cluster_selection_epsilon=hdb_epsilon)
@@ -267,22 +272,33 @@ def calc_clusters(
 @callback(
     Output("cutoff-type", "value"),
     Output("cutoff-type", "options"),
+    Output("plot-cutoff-name", "children"),
+    Output("cutoff-type", "clearable"),
     Input("data-store", "data"),
+    Input("dataset-plot-type", "value"),
     State("cutoff-type", "value"),
 
 )
-def update_cutoff_selection(rapdordata: RAPDORData, current_selection):
+def update_cutoff_selection(rapdordata: RAPDORData, plot_type, current_selection):
     if rapdordata is None:
         raise PreventUpdate
     options = [option for option in rapdordata.score_columns if
                option in rapdordata.df and is_numeric_dtype(rapdordata.df[option])]
     selection = dash.no_update
+
+    if plot_type == "Bubble Plot":
+        name = "Cutoff Type"
+        clearable = True
+    else:
+        name = "Y Axis"
+        selection = "ANOSIM R" if len(options) > 0 and current_selection not in options else selection
+        clearable = False
+        options = [option for option in options if ("p-Value" in option) or option in ("ANOSIM R", "PERMANOVA F")]
     if len(options) > 0 and current_selection not in options:
         selection = None
     elif len(options) == 0:
         options = dash.no_update
-
-    return selection, options
+    return selection, options, name, clearable
 
 @callback(
     Output("cutoff-range", "min"),
@@ -291,23 +307,30 @@ def update_cutoff_selection(rapdordata: RAPDORData, current_selection):
     Output("cutoff-range", "value"),
     Output("cutoff-range", "disabled"),
     Input("cutoff-type", "value"),
+    Input("dataset-plot-type", "value"),
     State("data-store", "data")
 
 )
-def update_range_slider(cutoff_type, rapdordata: RAPDORData):
+def update_range_slider(cutoff_type, plot_type, rapdordata: RAPDORData):
+    if plot_type != "Bubble Plot":
+        marks = [0, 1]
+        marks_t = {i: f"" for i in marks}
+        marks = []
+        return 0, 1, marks_t, marks, True
     if cutoff_type is not None:
         min_v = np.nanmin(rapdordata.df[cutoff_type])
         max_v = np.nanmax(rapdordata.df[cutoff_type])
         min_v = np.floor(min_v * 100) / 100
         max_v = np.ceil(max_v * 100) / 100
         if "p-Value" in cutoff_type:
             min_v = min(1e-5, min_v)
+            if min_v == 0:
+                min_v = 1e-20
             max_v = 1
             min_v = np.floor(np.log10(min_v))
             max_v = np.ceil(np.log10(max_v))
             marks = np.linspace(min_v, max_v, 5)
             marks = np.floor(marks)
-
             p = np.log10(0.05)
             iidx = np.searchsorted(marks, p)
             marks = np.insert(marks, iidx, p)
@@ -329,11 +352,27 @@ def update_range_slider(cutoff_type, rapdordata: RAPDORData):
         marks_t = None
         d_max = None
         disabled = True
+    logger.info(f"updating range slider: {min_v, max_v, marks_t, d_max, disabled}")
     return min_v, max_v, marks_t, (min_v, d_max), disabled
 
 
 
 
+@callback(
+    Output("showLFC", "on"),
+    Output("showLFC", "disabled"),
+    Output("3d-plot", "disabled"),
+    Input("3d-plot", "on"),
+    Input("dataset-plot-type", "value"),
+)
+def disable_lfc_and_3d(tdplot, plot_type):
+    if plot_type == "Bubble Plot":
+        if tdplot:
+            return False, True, False
+        else:
+            return dash.no_update, False, False
+    else:
+        return dash.no_update, False, True
 
 
 
@@ -342,6 +381,7 @@ def update_range_slider(cutoff_type, rapdordata: RAPDORData):
 @callback(
     Output("cluster-graph", "figure"),
     Input("night-mode", "on"),
+    Input("dataset-plot-type", "value"),
     Input("primary-color", "data"),
     Input("secondary-color", "data"),
     Input("plot-dim-red", "data"),
@@ -350,47 +390,60 @@ def update_range_slider(cutoff_type, rapdordata: RAPDORData):
     Input('3d-plot', 'on'),
     Input('cutoff-range', 'value'),
     Input("additional-header-dd", "value"),
-
+    Input("showLFC", "on"),
     State('cutoff-type', 'value'),
     State('data-store', "data"),
 )
-def plot_cluster_results(night_mode, color, color2, plotting, selected_rows, marker_size, td_plot, cutoff_range, add_header, cutoff_type, rapdordata: RAPDORData):
+def plot_cluster_results(night_mode, plot_type, color, color2, plotting, selected_rows, marker_size, td_plot, cutoff_range, add_header, show_lfc, cutoff_type, rapdordata: RAPDORData):
     logger.info(f"running cluster plot triggered via - {ctx.triggered_id}")
-    dim = 2 if not td_plot else 3
-    if dim == 3 and ctx.triggered_id == "cluster-marker-slider":
-        raise PreventUpdate
-    colors = [color, color2]
     if rapdordata is None:
         raise PreventUpdate
-
+    colors = [color, color2]
+    dim = 2 if not td_plot else 3
     if not plotting:
         fig = empty_figure("Data not Calculated<br> Get Scores first")
     else:
         if selected_rows is not None and len(selected_rows) >= 1:
             highlight = rapdordata.df.loc[selected_rows, "RAPDORid"]
         else:
             highlight = None
-        logger.info(f"Cutoff - {cutoff_range}")
-        if cutoff_type is None:
-            cutoff_range = None
-        else:
-            if "p-Value" in cutoff_type:
-                cutoff_range = 10 ** cutoff_range[0], 10 ** cutoff_range[1]
-
-        fig = plot_dimension_reduction(
-            rapdordata,
-            dimensions=dim,
-            colors=colors,
-            highlight=highlight,
-            show_cluster=True if "Cluster" in rapdordata.df else False,
-            marker_max_size=marker_size,
-            second_bg_color="white" if not night_mode else "#181818",
-            bubble_legend_color="black" if not night_mode else "white",
-            title_col=add_header,
-            cutoff_range=cutoff_range,
-            cutoff_type=cutoff_type
+        if plot_type == "Bubble Plot":
+            if dim == 3 and ctx.triggered_id == "cluster-marker-slider":
+                raise PreventUpdate
+
+            logger.info(f"Cutoff - {cutoff_range}")
+            logger.info(f"highlight - {highlight}")
+            if cutoff_type is None:
+                cutoff_range = None
+            else:
+                if "p-Value" in cutoff_type:
+                    cutoff_range = 10 ** cutoff_range[0], 10 ** cutoff_range[1]
+            if show_lfc:
+                highlight_color = "white" if night_mode else "black"
+            else:
+                highlight_color = None
+            fig = plot_dimension_reduction(
+                rapdordata,
+                dimensions=dim,
+                colors=colors,
+                highlight=highlight,
+                show_cluster=True if "Cluster" in rapdordata.df else False,
+                marker_max_size=marker_size,
+                second_bg_color="white" if not night_mode else "#181818",
+                bubble_legend_color="black" if not night_mode else "white",
+                title_col=add_header,
+                cutoff_range=cutoff_range,
+                cutoff_type=cutoff_type,
+                highlight_color=highlight_color,
+                show_lfc=show_lfc
 
-        )
+            )
+        else:
+            if cutoff_type is None:
+                fig = empty_figure("Select Y Axis Type first")
+            else:
+                fig = plot_distance_and_var(rapdordata, colors, title_col=add_header, highlight=highlight, var_type=cutoff_type, show_lfc=show_lfc)
+                fig.update_traces(marker=dict(size=marker_size))
     if not night_mode:
 
         fig.layout.template = DEFAULT_TEMPLATE