Automate chart rendering and screenshot update PRs

.github/workflows/screenshots.yml

@@ -0,0 +1,50 @@
+name: Update chart screenshots
+
+on:
+  workflow_dispatch:
+    inputs:
+      corporate_travel:
+        description: Include corporate-travel (requires the external source to respond)
+        type: boolean
+        default: true
+
+permissions:
+  contents: write
+  pull-requests: write
+
+concurrency:
+  group: chart-screenshots
+  cancel-in-progress: false
+
+jobs:
+  render:
+    runs-on: ubuntu-latest
+    timeout-minutes: 15
+    steps:
+      - uses: actions/checkout@v7
+      - uses: actions/setup-python@v7
+        with:
+          python-version: '3.12'
+          cache: pip
+      - run: python -m pip install -r requirements.txt
+      - name: Render charts
+        env:
+          INCLUDE_CORPORATE_TRAVEL: ${{ inputs.corporate_travel }}
+        run: |
+          if [ "$INCLUDE_CORPORATE_TRAVEL" = "true" ]; then
+            python render.py
+          else
+            python render.py --charts pay food richest-by-age store-sales
+          fi
+      - uses: peter-evans/create-pull-request@v8
+        with:
+          branch: automated/chart-screenshots
+          commit-message: Update chart screenshots
+          title: Update chart screenshots
+          body: |
+            Regenerated with render.py on Python 3.12.
+            Corporate travel included: ${{ inputs.corporate_travel }}.
+            Please review the image changes before merging.
+          add-paths: |
+            */screenshot.png
+            pay/zoomed.png

README.md

@@ -11,3 +11,30 @@ Company Data
 
 <img src="store-sales/screenshot.png" width="49%"></img>
 <img src="corporate-travel/screenshot.png" width="49%"></img>
+
+## Regenerate screenshots
+
+Use Python 3.12, install `requirements.txt`, then run `python render.py`.
+This saves PNGs without opening a window, including the pay detail view.
+The scripts still support interactive use when run individually.
+
+`corporate-travel` fetches 100 company pages. Requests have timeouts and HTTP
+or missing-data failures stop the run before any existing screenshots are
+replaced. The source can deny automated requests (HTTP 403); resolve source
+access before retrying. To explicitly render only local datasets, use:
+
+```
+python render.py --charts pay food richest-by-age store-sales
+```
+
+Use `--output-dir /path/to/previews` to inspect outputs elsewhere first.
+The charts retain their original historical datasets; regeneration does not
+make them current. The pay detail view uses x=85..753 and y=63..248 to approximate
+the existing screenshot's visible region.
+
+In GitHub Actions, manually run **Update chart screenshots**. It renders all
+charts by default; uncheck `corporate_travel` to leave that image unchanged.
+Successful runs create or update a PR on `automated/chart-screenshots` when
+images differ. Enable **Allow GitHub Actions to create and approve pull
+requests** in the repository's Actions settings. This workflow opens PRs;
+it does not approve or merge them.

corporate-travel/main.py

@@ -110,9 +110,14 @@
 for line in s.splitlines():
     if 'McKinsey' in line:
         line += ' '
-    response = SESSION.get("https://www.businesstravelnews.com/Corporate-Travel-100/2018/" + re.sub('\s*&\s*', '-', line).replace(' ', '-').replace('.', '').replace('Intel', 'Intel1')).text
+    response = SESSION.get("https://www.businesstravelnews.com/Corporate-Travel-100/2018/" + re.sub(r'\s*&\s*', '-', line).replace(' ', '-').replace('.', '').replace('Intel', 'Intel1'), timeout=(10, 30))
+    response.raise_for_status()
+    response = response.text
     volumeStrings = [line for line in response.splitlines() if "Volume" in line]
-    volumePerCompany[line] = float(re.search("\$(.*?) million", volumeStrings[0]).group(1))
+    match = re.search(r'\$(.*?) million', volumeStrings[0]) if volumeStrings else None
+    if match is None:
+        raise ValueError(f'No air volume found for {line}')
+    volumePerCompany[line] = float(match.group(1))
 
 plt.figure(figsize=(8, 15))
 plt.barh(*zip(*OrderedDict(sorted(volumePerCompany.items(), key=itemgetter(1))).items()))

render.py

@@ -0,0 +1,46 @@
+"""Render charts without a display; publish files only after every chart succeeds."""
+import argparse
+from pathlib import Path
+import runpy
+import shutil
+import tempfile
+
+import matplotlib
+
+matplotlib.use('Agg')
+import matplotlib.pyplot as plt
+
+ROOT = Path(__file__).resolve().parent
+CHARTS = ('pay', 'food', 'richest-by-age', 'store-sales', 'corporate-travel')
+
+
+def render(charts, output):
+    with tempfile.TemporaryDirectory() as directory:
+        staged = Path(directory)
+        for chart in charts:
+            target = staged / chart
+            target.mkdir(exist_ok=True)
+            try:
+                runpy.run_path(str(ROOT / chart / 'main.py'), run_name='__main__')
+                plt.savefig(target / 'screenshot.png')
+                if chart == 'pay':
+                    # Reproduce the detail view in the existing pay/zoomed.png.
+                    plt.gcf().set_size_inches(20.48, 10.24)
+                    plt.xlim(85, 753)
+                    plt.ylim(63, 248)
+                    plt.savefig(target / 'zoomed.png')
+            finally:
+                plt.close('all')
+        for source in sorted(staged.glob('*/*.png')):
+            destination = output / source.relative_to(staged)
+            destination.parent.mkdir(parents=True, exist_ok=True)
+            shutil.copyfile(source, destination)
+            print(destination)
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument('--charts', nargs='+', choices=CHARTS, default=CHARTS)
+    parser.add_argument('--output-dir', type=Path, default=ROOT)
+    args = parser.parse_args()
+    render(args.charts, args.output_dir)

requirements.txt

@@ -1,2 +1,2 @@
-requests==2.22.0
-matplotlib==3.1.1
+requests==2.34.2
+matplotlib==3.11.1

test_render.py

@@ -0,0 +1,36 @@
+import tempfile
+from pathlib import Path
+from unittest.mock import patch
+
+import matplotlib.pyplot as plt
+
+from render import render
+
+
+def test_failed_chart_preserves_existing_outputs():
+    with tempfile.TemporaryDirectory() as directory:
+        output = Path(directory)
+        previous = output / 'pay' / 'screenshot.png'
+        previous.parent.mkdir()
+        previous.write_bytes(b'previous screenshot')
+
+        def run_chart(path, **kwargs):
+            if 'food' in path:
+                raise ValueError('source unavailable')
+            plt.plot([1, 2], [3, 4])
+
+        with patch('render.runpy.run_path', side_effect=run_chart):
+            try:
+                render(('pay', 'food'), output)
+            except ValueError as error:
+                assert str(error) == 'source unavailable'
+            else:
+                raise AssertionError('Failed chart must fail the rendering run')
+        assert previous.read_bytes() == b'previous screenshot'
+        assert not (output / 'pay' / 'zoomed.png').exists()
+        assert not (output / 'food').exists()
+
+
+if __name__ == '__main__':
+    test_failed_chart_preserves_existing_outputs()
+    print('Failed chart preserves existing outputs: passed')