plot full range of completeness - contamination

CAMI-challenge · May 23, 2019 · 089ea20 · 089ea20
1 parent f9b118f
commit 089ea20
Show file tree

Hide file tree

Showing 4 changed files with 17 additions and 17 deletions.
diff --git a/.circleci/config.yml b/.circleci/config.yml
@@ -9,9 +9,6 @@
 version: 2
 jobs:
   build:
-    branches:
-      ignore:
-      - amber20
     working_directory: ~/CAMI-challenge/AMBER
     parallelism: 1
     shell: /bin/bash --login

diff --git a/cami_i_hc/index.html b/cami_i_hc/index.html
diff --git a/src/amber_html.py b/src/amber_html.py
@@ -468,17 +468,19 @@ def create_precision_recall_all_genomes_scatter(pd_genome_bins, tools):
 
 
 def create_contamination_plot(pd_bins, tools, title, xlabel, ylabel, create_column_function):
+    pd_bins_copy = pd_bins[[utils_labels.TOOL, 'purity_bp', 'completeness_bp']].copy().dropna(subset=['purity_bp'])
+    create_column_function(pd_bins_copy)
+
     colors_list = plots.create_colors_list()
     bokeh_colors = [matplotlib.colors.to_hex(c) for c in colors_list]
 
-    p = figure(title=title, plot_width=580, plot_height=400, y_range=(0, 1), toolbar_location="below")
+    p = figure(title=title, plot_width=580, plot_height=400, toolbar_location="below")
     p.x_range.start = 0
+    p.y_range.start = pd_bins_copy['newcolumn'].min()
+    p.y_range.end = 1
     legend_it = []
     for color, tool in zip(bokeh_colors, tools):
-        pd_tool_bins = pd_bins[pd_bins[utils_labels.TOOL] == tool]
-        pd_tool_bins = pd_tool_bins.dropna(subset=['purity_bp'])
-
-        create_column_function(pd_tool_bins)
+        pd_tool_bins = pd_bins_copy[pd_bins_copy[utils_labels.TOOL] == tool]
         pd_tool_bins = pd_tool_bins.sort_values(by='newcolumn', ascending=False).reset_index()
         pd_tool_bins = pd_tool_bins.drop(['index'], axis=1)
 

diff --git a/src/plots.py b/src/plots.py
@@ -351,24 +351,25 @@ def plot_contamination(pd_bins, binning_type, title, xlabel, ylabel, create_colu
     if len(pd_bins) == 0:
         return
 
+    pd_bins_copy = pd_bins[[utils_labels.TOOL, 'purity_bp', 'completeness_bp']].copy().dropna(subset=['purity_bp'])
+    create_column_function(pd_bins_copy)
+
     colors_list = create_colors_list()
 
     fig, axs = plt.subplots(figsize=(6, 5))
 
-    tools = pd_bins[utils_labels.TOOL].unique().tolist()
+    tools = pd_bins_copy[utils_labels.TOOL].unique().tolist()
 
     for color, tool in zip(colors_list, tools):
-        pd_tool_bins = pd_bins[pd_bins[utils_labels.TOOL] == tool]
-        pd_tool_bins = pd_tool_bins.dropna(subset=['purity_bp'])
-
-        create_column_function(pd_tool_bins)
+        pd_tool_bins = pd_bins_copy[pd_bins_copy[utils_labels.TOOL] == tool]
         pd_tool_bins = pd_tool_bins.sort_values(by='newcolumn', ascending=False).reset_index()
         pd_tool_bins = pd_tool_bins.drop(['index'], axis=1)
 
         axs.plot(list(range(1, len(pd_tool_bins) + 1)), pd_tool_bins['newcolumn'], color=color)
 
-    axs.set_ylim(0.0, 1.0)
+    axs.set_ylim(pd_bins_copy['newcolumn'].min(), 1.0)
     axs.set_xlim(1, None)
+    axs.grid(which='major', linestyle='-', linewidth='0.5', color='lightgrey')
 
     # transform plot_labels to percentages
     vals = axs.get_yticks()