Merge pull request #8 from afeld/check-change

check the change for every notebook in CI

Author: afeld

.github/workflows/main.yml

@@ -31,7 +31,7 @@ jobs:
           - lecture_2.ipynb
           - lecture_3.ipynb
           - lecture_4.ipynb
-          - lecture_5.ipynb
+          # - lecture_5.ipynb # gets a different .similarity() score locally vs. CI
           # - lecture_5_exercise.ipynb # has incomplete code
           - lecture_6.ipynb
     steps:
@@ -55,4 +55,4 @@ jobs:
         run: pipenv install
 
       - name: Confirm notebook hasn't changed
-        run: pipenv run ./scripts/update.sh ${{ matrix.notebook }}
+        run: pipenv run ./scripts/check_change.sh ${{ matrix.notebook }}

scripts/diffable.py

@@ -2,14 +2,46 @@
 import re
 import sys
 
+
+def is_pip_upgrade_msg(line):
+    return isinstance(line, str) and re.match(r"WARNING.+pip version|upgrade pip", line)
+
+
+def is_vid(cell):
+    try:
+        text = cell["outputs"][0]["data"]["text/plain"][0]
+    except (IndexError, KeyError, TypeError):
+        return False
+
+    return text == "<IPython.core.display.Video object>"
+
+
 input_str = sys.stdin.read()
 notebook = json.loads(input_str)
 
+# nbconvert wants to embed videos, so skip them
+notebook["cells"] = [cell for cell in notebook["cells"] if not is_vid(cell)]
+
 for cell in notebook["cells"]:
+    if "execution_count" in cell:
+        # ignore all the execution count numbers
+        cell["execution_count"] = None
+
     if cell["cell_type"] != "code":
         continue
 
+    # ignore any system command output
+    if cell["source"][0].startswith("!"):
+        cell["outputs"] = []
+
+    # filter out pip upgrade warnings
+    cell["outputs"] = [line for line in cell["outputs"] if not is_pip_upgrade_msg(line)]
+
     for output in cell["outputs"]:
+        if "execution_count" in output:
+            # ignore all the execution count numbers
+            output["execution_count"] = 1
+
         # clear HTML output, since it often has generated IDs (from displacy, plotly, etc.) that change with each execution
         if "data" in output and "text/html" in output["data"]:
             cell["outputs"] = []