Updated documentation

docs/assets/css/style.scss

@@ -80,7 +80,7 @@ a.nostyle, a.nostyle:hover, a.nostyle:active, a.nostyle:focus {
   }
 
   img {
-    display: block;
+
     margin-left: auto;
     margin-right: auto;
   }

docs/examples.html

@@ -5,95 +5,73 @@
 permalink: examples/
 
 examples:
- - title: Generating data
-   simu_image: generated_pop.png
-   real_image: unavailable.png
-   text: All the data included in the downloadable archive from the documentation can be reproduced from scratch. This example shows how to produce a dataset from gaussian noise around several paths.
-   link: generating_data
-   simu_only: true
- - title: Plotting cell sets
-   simu_image: notebook_cell_sets.png
-   real_image: unavailable.png
-   text: It is very easy to generate and parse cell sets in python and color each cell set for plotting. This example shows how to parse cell sets files and quickly plot 2 genes from a dataset, coloring every cell set.
-   link: plotting_cell_sets
- - title: Computing transport maps
-   simu_image: tmaps_46_47_1.png
-   real_image: tmaps_46_47_1.png
-   text: This example shows how to compute transport maps from the command line, choosing the OT parameters that best suit your needs before using those maps to compute ancestors and trajectories.
-   link: computing_transport_maps
- - title: Plotting ancestors
-   simu_image: notebook_ancestors.png
-   real_image: unavailable.png
-   text: The time-course data about the cells is then used to infer the evolution of cells over time in gene-expression space. This example shows how to plot infered ancestors and descendants of a given cell set, to show its trajectory.
-   link: plotting_ancestors
- - title: Plotting shared ancestry
-   simu_image: shared_ancestry.png
-   real_image: unavailable.png
-   text: Knowing the trajectory of each cell set, one can then identify the shared ancestors between two cell sets. This example shows how to plot shared ancestry between several cell sets.
-   link: shared_ancestry
- - title: Plotting trajectory trends
-   simu_image: trajectory_trends_1.png
-   real_image: unavailable.png
-   text: While two-dimensional embeddings are fine for vizualizing a whole population, the evolution of each gene can also be more closely monitored over the evolution of a cell set. This example shows how to plot the expression of each gene over time in a cell set.
-   link: trajectory_trends
- - title: Plotting ancestor census
-   simu_image: ancestor_census.png
-   real_image: unavailable.png
-   text: Several files generated by wot can be viewed and plotted as datasets. This example shows how to parse and plot a census file to generate an ancestor census plot. This method can be easily extended to all kinds of dataset files.
-   link: ancestor_census
- - title: Plotting validation summary
-   simu_image: validation_summary.png
-   real_image: unavailable.png
-   text: After using the optimal transport validation tool, a validation summary file is generated. This example show how to parse it and plot the average and standard deviation of the distances between each type of populations (real, random, interpolated) over time.
-   link: plotting_validation_summary
-
-datasets_simulated:
- - true
- - false
+- title: Generating data
+  image: generated_pop.png
+  text: This example shows how to produce a dataset from gaussian noise around several paths.
+  link: generating_data
+
+- title: Plotting cell sets
+  image: notebook_cell_sets.png
+  text: It is very easy to generate and parse cell sets in python and color each cell set for plotting. This example shows how to parse cell sets files and quickly plot 2 genes from a dataset, coloring every cell set.
+  link: plotting_cell_sets
+
+- title: Computing transport maps
+  image: tmaps_46_47_1.png
+  text: This example shows how to compute transport maps from the command line, choosing the OT parameters that best suit your needs before using those maps to compute ancestors and trajectories.
+  link: computing_transport_maps
+
+- title: Plotting ancestors
+  image: notebook_ancestors.png
+  text: The time-course data about the cells is then used to infer the evolution of cells over time in gene-expression space. This example shows how to plot infered ancestors and descendants of a given cell set, to show its trajectory.
+  link: plotting_ancestors
+
+- title: Plotting shared ancestry
+  image: shared_ancestry.png
+  text: Knowing the trajectory of each cell set, one can then identify the shared ancestors between two cell sets. This example shows how to plot shared ancestry between several cell sets.
+  link: shared_ancestry
+
+- title: Plotting trajectory trends
+  image: trajectory_trends.png
+  text: While two-dimensional embeddings are fine for vizualizing a whole population, the evolution of each gene can also be more closely monitored over the evolution of a cell set. This example shows how to plot the expression of each gene over time in a cell set.
+  link: trajectory_trends
+
+- title: Plotting ancestor census
+  image: ancestor_census.png
+  text: Several files generated by wot can be viewed and plotted as datasets. This example shows how to parse and plot a census file to generate an ancestor census plot. This method can be easily extended to all kinds of dataset files.
+  link: ancestor_census
+
+- title: Plotting validation summary
+  image: validation_summary.png
+  text: After using the optimal transport validation tool, a validation summary file is generated. This example show how to parse it and plot the average and standard deviation of the distances between each type of populations (real, random, interpolated) over time.
+  link: plotting_validation_summary
 ---
 
-<ul class="nav nav-pills my-5 mx-5 float-right" role="tablist">
-  <li class="nav-item"><a href="#simulated" class="mx-2 py-3 nav-link active" data-toggle="pill">Simulated data</a></li>
-  <li class="nav-item"><a href="#real" class="mx-2 py-3 nav-link" data-toggle="pill">Real data</a></li>
-</ul>
 
-<div class="tab-content">
-{% for simulated in page.datasets_simulated %}
-<div class="container my-5 tab-pane {% if forloop.first %}show active{% endif %}"
-     {% if simulated %} id="simulated" {% else %} id="real" {% endif %}>
-  <h3 class="text-center mt-5">Running example on how to use <b>wot</b> on {% if simulated %} simulated data {% else %} real data <br/>(currently unavailable){% endif %}</h3>
+<div class="container my-5 tab-pane {% if forloop.first %}show active{% endif %}">
+    <h3 class="text-center mt-5">Examples on how to use <b>wot</b> on simulated data</h3>
 
-  <h5 class="text-center mb-5">These examples are also available as a <a class="nounderline" href="{{site.baseurl}}/notebook">python notebook</a></h5>
+    <hr class="featurette-divider">
 
-  <hr class="featurette-divider">
+    {% for example in page.examples %}
 
-  {% for example in page.examples %}
-    {% if example['simu_only'] != true or simulated %}
-      <a href="{{ example['link'] }}" class="nostyle">
+    <a href="{{ example['link'] }}" class="nostyle">
         <div class="row featurette my-2">
-          <div class="col-md-5 {% cycle('order-1', 'order-3') %}">
-            <img class="featurette-image img-responsive mx-auto d-block"
-                 {% if simulated %}
-                   src="{{site.baseurl}}/images/{{ example['simu_image'] }}"
-                 {% else %}
-                   src="{{site.baseurl}}/images/{{ example['real_image'] }}"
-                 {% endif %}
-                 width=300 height=300 alt="Generic placeholder image">
-          </div>
-
-          <div class="col-md-7 order-2 my-5">
-              <h2 class="featurette-heading my-3">
-                {{ example['title'] }}
-              </h2>
-            <p class="lead my-3">{{ example['text'] }}</p>
-          </div>
-        </div>
-      </a>
+            <div class="col-md-5 {% cycle('order-1', 'order-3') %}">
+                <img class="featurette-image img-responsive mx-auto d-block"
+                     src="{{site.baseurl}}/images/{{ example['image'] }}"
+                     width=300 height=300>
+            </div>
 
-      <hr class="featurette-divider">
-    {% endif  %}
-  {% endfor %}
+            <div class="col-md-7 order-2 my-5">
+                <h2 class="featurette-heading my-3">
+                    {{ example['title'] }}
+                </h2>
+                <p class="lead my-3">{{ example['text'] }}</p>
+            </div>
+        </div>
+    </a>
 
+    <hr class="featurette-divider">
+    {% endfor %}
 </div>
-{% endfor %}
-</div>
+

docs/examples/ancestor_census.md

@@ -14,7 +14,7 @@ You can compute an ancestor census for your cell sets with:
 ```sh
 wot census --tmap tmaps \
  --cell_set cell_sets.gmt \
- --out census --time 15
+ --out census --time 7
 ```
 
 See the [CLI documentation]({{site.baseurl}}/cli_documentation#ancestor-census)
@@ -28,31 +28,7 @@ Note : You need to have a transport maps computed for this to work. Please refer
 This will create several census files, that you can then plot as follows :
 
 ```python
-import wot.graphics
-from matplotlib import pyplot
-
-picture_title = "Ancestor census of cells from Tip 1 at day 10"
-
-# Change this to the name of census file you want to plot
-census_file = 'census_tip1.txt'
-
-# Choose a color and a name to display for each cell set in the census
-legend = [
-        [ "#e00000", "Tip 1" ],
-        [ "#00e000", "Tip 2" ],
-        [ "#0000e0", "Tip 3" ]
-        ]
-
-ds = wot.io.read_dataset(census_file)
-
-wot.graphics.legend_figure(pyplot, legend)
-for i in range(ds.X.shape[1]):
-    pyplot.plot(ds.X[:,i], color=legend[i][0])
-
-pyplot.title(picture_title)
-pyplot.xlabel("time")
-pyplot.ylabel("Proportion of ancestors per cell set")
-pyplot.show()
+{% include_relative code/07_plotting_ancestor_census.py %}
 ```
 
 ### Result ###

docs/notebook/00_generating_data.py → docs/examples/code/00_generating_data.py

@@ -1,17 +1,19 @@
 import numpy as np
-import wot.simulate
 from numpy.random import randint
 from numpy.random import random
 
+import wot.simulate
+
 # ------ Configuration variables -------
 matrix_file = 'matrix.txt'
 days_file = 'days.txt'
 covariate_file = 'covariate.txt'
 gene_sets_file = 'gene_sets.gmt'
 
-number_of_timepoints = 51
-covariates_count = 5
-average_cell_count_per_timepoint = 3000
+number_of_timepoints = 8
+covariates_count = 2
+average_cell_count_per_timepoint = 1000
+np.random.seed(1234)
 
 gene_sets = {
     'Stem cells': ['Stem_gene'],

docs/notebook/02_plotting_two_features.py → docs/examples/code/02_plotting_two_features.py

@@ -27,7 +27,7 @@
 pyplot.figure(figsize=(5, 5))
 pyplot.axis('off')
 wot.graphics.plot_2d_dataset(pyplot, ds,
-                             x=gene_x_plot, y=gene_y_plot)
+                             x=gene_x_plot, y=gene_y_plot, colors=ds.obs['color'].values)
 pyplot.autoscale(enable=True, tight=True)
 pyplot.tight_layout(pad=0)
 pyplot.savefig(destination_file)

docs/notebook/03_plotting_cell_sets.py → docs/examples/code/03_plotting_cell_sets.py

@@ -1,3 +1,4 @@
+import numpy as np
 from matplotlib import pyplot
 
 import wot.graphics
@@ -33,17 +34,17 @@
 wot.io.write_gene_sets(cell_sets, cell_sets_file, "gmt")
 
 # Plot the cell sets
-
-wot.set_cell_metadata(ds, 'color', bg_color)
+colors = np.array([bg_color] * len(ds.obs))
 
 for color, cset_name in cell_sets_to_color:
-    wot.set_cell_metadata(ds, 'color', color,
-                          indices=cell_sets[cset_name])
+    cell_ids = cell_sets[cset_name]
+    indices = ds.obs.index.get_indexer_for(cell_ids)
+    colors[indices] = color
 
 pyplot.figure(figsize=(5, 5))
 pyplot.axis('off')
 wot.graphics.plot_2d_dataset(pyplot, ds,
-                             x=gene_x_plot, y=gene_y_plot)
+                             x=gene_x_plot, y=gene_y_plot, s=1, colors=colors)
 wot.graphics.legend_figure(pyplot, cell_sets_to_color)
 pyplot.autoscale(enable=True, tight=True)
 pyplot.tight_layout(pad=0)

docs/notebook/04_plotting_ancestors.py → docs/examples/code/04_plotting_ancestors.py

@@ -10,7 +10,7 @@
 gene_y_plot = 1
 cell_sets_file = 'cell_sets.gmt'
 target_cell_set = "Red blood cells"
-target_timepoint = 50
+target_timepoint = 7
 destination_file = "ancestors.png"
 # --------------------------------------
 
@@ -33,6 +33,7 @@ def color_cells(population):
 pyplot.figure(figsize=(5, 5))
 pyplot.axis('off')
 wot.set_cell_metadata(ds, 'color', bg_color)
+
 wot.graphics.plot_2d_dataset(pyplot, ds, x=gene_x_plot, y=gene_y_plot)
 
 cell_sets = wot.io.read_sets(cell_sets_file, as_dict=True)
@@ -46,7 +47,7 @@ def color_cells(population):
     color_cells(population)
 
 wot.graphics.plot_2d_dataset(pyplot, ds,
-                             x=gene_x_plot, y=gene_y_plot)
+                             x=gene_x_plot, y=gene_y_plot, s=1, colors=ds.obs['color'].values)
 wot.graphics.legend_figure(pyplot,
                            [["#316DA2", "Ancestors of {}".format(target_cell_set)]])
 pyplot.autoscale(enable=True, tight=True)

docs/notebook/05_plotting_shared_ancestry.py → docs/examples/code/05_plotting_shared_ancestry.py

@@ -11,7 +11,7 @@
 cell_set_1 = "Red blood cells"
 cell_set_2 = "Granulocytes"
 cell_sets_file = 'cell_sets.gmt'
-target_timepoint = 50
+target_timepoint = 7
 destination_file = "shared_ancestry.png"
 # --------------------------------------
 
@@ -38,7 +38,7 @@
           for i in range(len(t))]
 ds.obs.loc[:, 'color'] = colors
 
-wot.graphics.plot_2d_dataset(pyplot, ds)
+wot.graphics.plot_2d_dataset(pyplot, ds, s=1, colors=ds.obs['color'].values)
 wot.graphics.legend_figure(pyplot,
                            [["#A00000", "Ancestors of " + cell_set_1],
                             ["#0000A0", "Ancestors of " + cell_set_2]],

docs/notebook/06_plotting_trajectory_trends.py → docs/examples/code/06_plotting_trajectory_trends.py

@@ -8,7 +8,7 @@
 days_file = 'days.txt'
 cell_sets_file = 'cell_sets.gmt'
 target_cell_set = 'Red blood cells'
-target_timepoint = 50
+target_timepoint = 7
 skip_first_n_genes = 2
 destination_file = "trajectory_trends.png"
 # --------------------------------------
@@ -22,7 +22,8 @@
 
 # timepoints, means, variances = tmap_model.compute_trajectory_trends(ds, population)
 trajectory_ds = tmap_model.compute_trajectories({target_cell_set: all_populations[target_cell_set]})
-results = wot.ot.compute_trajectory_trends_from_trajectory(trajectory_ds, ds)
+
+results = wot.tmap.compute_trajectory_trends_from_trajectory(trajectory_ds, ds)
 means, variances = results[0]
 timepoints = means.obs.index.values
 pyplot.figure(figsize=(5, 5))

docs/notebook/07_plotting_ancestor_census.py → docs/examples/code/07_plotting_ancestor_census.py

@@ -6,7 +6,7 @@
 matrix_file = 'matrix.txt'
 cell_sets_file = 'cell_sets.gmt'
 target_cell_set = 'Red blood cells'
-target_timepoint = 50
+target_timepoint = 7
 destination_file = "ancestor_census.png"
 # --------------------------------------
 

docs/notebook/08_plotting_validation_summary.py → docs/examples/code/08_plotting_validation_summary.py

@@ -3,6 +3,7 @@
 
 import wot.commands
 import wot.graphics
+
 # ------ Configuration variables -------
 matrix_file = 'matrix.txt'
 days_file = 'days.txt'
@@ -11,11 +12,9 @@
 # --------------------------------------
 
 ot_model = wot.ot.initialize_ot_model(matrix_file, days_file,
-                                      covariate=covariate_file, growth_iters=1, tmap_prefix='val')
+                                      covariate=covariate_file, growth_iters=1, tmap_out='val')
 vs = wot.commands.compute_validation_summary(ot_model)
-vs['time'] = (vs['interval_start'] + vs['interval_end']) / 2
-vs['type'] = vs['pair0'].astype(str).str[0]
-res = vs.groupby(['time', 'type'])['distance'] \
+res = vs.groupby(['interval_mid', 'name'])['distance'] \
     .agg([np.mean, np.std])
 
 legend = {