Added support for histogram creation.
Added argument for histogram creation.
Added function to streamline Hirstogram generatoin.
Added function to generate all histograms.
BUG=chromium:1103853
TEST=None
Change-Id: Id489ecd9e9b3faaa7960cbae630b690a1400f6aa
Reviewed-on: https://chromium-review.googlesource.com/c/chromiumos/chromite/+/2374125
Commit-Queue: Kevin Ruvalcaba <kevinruvalcaba@google.com>
Tested-by: Kevin Ruvalcaba <kevinruvalcaba@google.com>
Reviewed-by: Alex Klein <saklein@chromium.org>
diff --git a/contrib/depgraph_visualization/README.md b/contrib/depgraph_visualization/README.md
index 1bf452c..1b8fc7b 100644
--- a/contrib/depgraph_visualization/README.md
+++ b/contrib/depgraph_visualization/README.md
@@ -1,4 +1,4 @@
-# Depgraph visualizer
+# Depgraph Visualizer
## Installation
Enter your SDK, go to this modules location,
@@ -10,6 +10,7 @@
cd ~/trunk/chromite/contrib/depgraph_visualization
./install.sh
source my_visualizations/bin/activate
+cd my_visualizations
```
This will put a script called `visualize_depgraph` on your PATH. From there you
@@ -41,6 +42,16 @@
```bash
visualize_depgraph net-fs/samba -b=amd64-generic --output-path=bar/foo --output-name=SambaGraph
```
+## Secondary usage
+With the argument `--include-histograms` you can also generate four png files
+with histograms for dependency and reverse dependency distribution.
+These plots are partition in two because the number of packages Y with a range
+X of (reverse)dependencies goes from 600 to 1.
+```bash
+visualize_depgraph --include-Histograms=True --output-path=foo/bar -b=arm64-generic
+```
+These files are saved in the same directory as the main output file.
+
## Important notes
If you were to use a package with no dependencies
diff --git a/contrib/depgraph_visualization/depgraph_visualization/depgraph_viz.py b/contrib/depgraph_visualization/depgraph_visualization/depgraph_viz.py
index 6f03c7d..581ca2f 100644
--- a/contrib/depgraph_visualization/depgraph_visualization/depgraph_viz.py
+++ b/contrib/depgraph_visualization/depgraph_visualization/depgraph_viz.py
@@ -27,6 +27,8 @@
help='Write output to the given path.')
parser.add_argument('--output-name', default='DepGraph',
help='Write output file name.')
+ parser.add_argument('--include-histograms', default=False,
+ help='Create and save histograms about dependencies.')
parser.add_argument('pkgs', nargs='*', default=_DEFAULT_PACKAGES)
opts = parser.parse_args(argv)
opts.Freeze()
@@ -77,3 +79,5 @@
runtime_tree = CreateRuntimeTree(sysroot, opts.pkgs)
dep_vis = visualize.DepVisualizer(runtime_tree)
dep_vis.VisualizeGraph(output_name=out_name, output_dir=out_dir)
+ if opts.include_histograms:
+ dep_vis.GenerateHistograms(opts.build_target, out_dir)
diff --git a/contrib/depgraph_visualization/depgraph_visualization/visualize.py b/contrib/depgraph_visualization/depgraph_visualization/visualize.py
index 16059de..01ff291 100644
--- a/contrib/depgraph_visualization/depgraph_visualization/visualize.py
+++ b/contrib/depgraph_visualization/depgraph_visualization/visualize.py
@@ -10,6 +10,7 @@
"""
from typing import Dict, Iterator, List, Set, Tuple
+import matplotlib.pyplot as plt # pylint: disable=import-error
class PackageNode(object):
@@ -145,6 +146,97 @@
# Writes an HTML file with the graph on it.
net.write_html(f'{output_dir}/{output_name}.html')
+ def GenerateHistograms(self, build_name: str, path: str):
+ """Creates 4 histograms with dependency and rvs dependency distribution.
+
+ The amount of packages with a certain range of dependencies and
+ reverse dependencies ranges from 600 to 1 so we split the histograms
+ of both into two; giving us four in total.
+
+ Args:
+ build_name: Name of the target build
+ path: Path to output files.
+ """
+ # Prepare data to plot.
+ dep_count = [len(n.dependencies) for n in self.pkg_dict.values()]
+ # There isn't a good explanation for the values of the bins other than
+ # they yield a good result.
+ dep_bins_low = [0, 1, 3, 5, 9, 13, 17, 20]
+ # The bigger bins start from 21 and go on multiples of 50
+ # until the highest value of the data points.
+ highest_dep = max(dep_count)
+ top = (highest_dep // 50)+1
+ dep_bins_high = [21] + [i*50 for i in range(1, top)] + [highest_dep]
+
+ # The histogram gets created and saved.
+ _SaveHistogram(dep_count,
+ dep_bins_low,
+ f'({build_name}): Dependency_distribution_low',
+ path,
+ '#205973')
+
+ _SaveHistogram(dep_count,
+ dep_bins_high,
+ f'({build_name}): Dependency_distribution_high',
+ path,
+ '#205973')
+
+ # Do the same for the reverse dependencies.
+ rvs_count = [len(n.rvs_dependencies) for n in self.pkg_dict.values()]
+ rvs_bins_low = [1, 2, 10, 20]
+ highest_rvs = max(rvs_count)
+ top = (highest_rvs // 50)+1
+ rvs_bins_high = [21] + [i*50 for i in range(1, top)] + [highest_rvs]
+
+ _SaveHistogram(rvs_count,
+ rvs_bins_low,
+ f'({build_name}): Reverse_Dependency_distribution_low',
+ path,
+ '#ef7e56')
+
+ _SaveHistogram(rvs_count,
+ rvs_bins_high,
+ f'({build_name}): Reverse_Dependency_distribution_high',
+ path,
+ '#ef7e56')
+
+
+def _SaveHistogram(data: List[int],
+ bins: List[int],
+ name: str,
+ path: str,
+ color: str):
+ """Streamline the process of plotting histograms.
+
+ Plots and saves a histogram as a png file.
+
+ Args:
+ data: List with data points.
+ bins: List with ranges for the histogram.
+ name: Name of the output file.
+ path: Path of the output file.
+ color: Color in either rgb or hexadecimal format.
+ """
+
+ plt.hist(data,
+ bins=bins,
+ edgecolor='black',
+ color=[color])
+
+ # Location for the labels in the x-ticks.
+ xplace = [(a+b) // 2 for a, b in zip(bins[:-1], bins[1:])]
+ # Create labels accurately portraying bin ranges.
+ xlabels = [f'{a}-{b-1}' for a, b in zip(bins[:-2], bins[1:-1])]
+ # Last range is inclusive.
+ xlabels += [f'{bins[-2]}-{bins[-1]}']
+ plt.xticks(xplace, xlabels)
+
+ plt.ylabel('Number of Packages')
+ plt.title(name)
+ plt.savefig(f'{path}/{name}.png')
+ # The plot needs to be cleaned otherwise all the graphs clump together.
+ plt.clf()
+
def _BfsColoring(net,
queue: List[Iterator[PackageNode]],
diff --git a/contrib/depgraph_visualization/requirements.txt b/contrib/depgraph_visualization/requirements.txt
index be25b9e..ebbf13c 100644
--- a/contrib/depgraph_visualization/requirements.txt
+++ b/contrib/depgraph_visualization/requirements.txt
@@ -1,3 +1,4 @@
--find-links file:///mnt/host/source/infra_virtualenv/pip_packages
portage>=2.3
+matplotlib==3.3.0
pyvis==0.1.*
