ci: get a list of changed files using git diff instead of GitHub APIv4

.github/workflows/build-notebooks-pr.yaml

@@ -1,5 +1,5 @@
 ---
-"name": "Build Notebooks"
+"name": "Build Notebooks (pr)"
 "on":
   "pull_request":
 
@@ -18,15 +18,18 @@ jobs:
     steps:
       - uses: actions/checkout@v4
 
-      - run: |
+      - name: Determine targets to build based on changed files
+        run: |
+          set -x
+          git fetch --no-tags origin 'pull/${{ github.event.pull_request.number }}/head:${{ github.event.pull_request.head.ref }}'
+          git fetch --no-tags origin '+refs/heads/${{ github.event.pull_request.base.ref }}:refs/remotes/origin/${{ github.event.pull_request.base.ref }}'
           python3 ci/cached-builds/gen_gha_matrix_jobs.py \
-            --owner=${{ github.repository_owner }} \
-            --repo=${{ github.event.pull_request.base.repo.name }} \
-            --pr-number=${{ github.event.pull_request.number }} \
-            --skip-unchanged
+            --from-ref 'origin/${{ github.event.pull_request.base.ref }}' \
+            --to-ref '${{ github.event.pull_request.head.ref }}'
         id: gen
         env:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+        shell: bash
 
   build:
     needs: ["gen"]

.github/workflows/build-notebooks.yaml

@@ -1,7 +1,7 @@
 ---
 # This file is autogenerated by ci/cached-builds/gen_gha_matrix_jobs.py
 {
-    "name": "Build Notebooks",
+    "name": "Build Notebooks (push)",
     "permissions": {
         "packages": "write"
     },

ci/cached-builds/gen_gha_matrix_jobs.py

@@ -98,7 +98,7 @@ def write_github_workflow_file(tree: dict[str, list[str]], path: pathlib.Path) -
         }
 
     workflow = {
-        "name": "Build Notebooks",
+        "name": "Build Notebooks (push)",
         # https://docs.github.com/en/actions/security-guides/automatic-token-authentication#modifying-the-permissions-for-the-github_token
         "permissions": {
             "packages": "write",
@@ -152,14 +152,10 @@ def main() -> None:
     logging.basicConfig(level=logging.DEBUG, stream=sys.stderr)
 
     argparser = argparse.ArgumentParser()
-    argparser.add_argument("--owner", type=str, required=False,
-                           help="GitHub repo owner/org (for the --skip-unchanged feature)")
-    argparser.add_argument("--repo", type=str, required=False,
-                           help="GitHub repo name (for the --skip-unchanged feature)")
-    argparser.add_argument("--pr-number", type=int, required=False,
-                           help="PR number under owner/repo (for the --skip-unchanged feature)")
-    argparser.add_argument("--skip-unchanged", type=bool, required=False, default=False,
-                           action=argparse.BooleanOptionalAction)
+    argparser.add_argument("--from-ref", type=str, required=False,
+                           help="Git ref of the base branch (to determine changed files)")
+    argparser.add_argument("--to-ref", type=str, required=False,
+                           help="Git ref of the PR branch (to determine changed files)")
     args = argparser.parse_args()
 
     # https://www.gnu.org/software/make/manual/make.html#Reading-Makefiles
@@ -170,9 +166,9 @@ def main() -> None:
     write_github_workflow_file(tree, project_dir / ".github" / "workflows" / "build-notebooks.yaml")
 
     leafs = compute_leafs_in_dependency_tree(tree)
-    if args.skip_unchanged:
-        logging.info(f"Skipping targets not modified in PR #{args.pr_number}")
-        changed_files = gha_pr_changed_files.list_changed_files(args.owner, args.repo, args.pr_number)
+    if args.from_ref:
+        logging.info(f"Skipping targets not modified in the PR")
+        changed_files = gha_pr_changed_files.list_changed_files(args.from_ref, args.to_ref)
         leafs = gha_pr_changed_files.filter_out_unchanged(leafs, changed_files)
     output = print_github_actions_pr_matrix(tree, leafs)
 
@@ -197,6 +193,8 @@ def test_select_changed_targets(self):
         changed_files = ["jupyter/datascience/ubi9-python-3.9/Dockerfile"]
 
         leafs = gha_pr_changed_files.filter_out_unchanged(leafs, changed_files)
-        assert set(leafs) == {'cuda-jupyter-tensorflow-ubi9-python-3.9',
+        assert set(leafs) == {'amd-jupyter-pytorch-c9s-python-3.9',
+                              'amd-jupyter-tensorflow-c9s-python-3.9',
+                              'cuda-jupyter-tensorflow-ubi9-python-3.9',
                               'jupyter-trustyai-ubi9-python-3.9',
                               'jupyter-pytorch-ubi9-python-3.9'}

ci/cached-builds/gha_pr_changed_files.py

@@ -1,11 +1,9 @@
-import json
 import logging
 import os
 import pathlib
 import re
 import subprocess
 import unittest
-import urllib.request
 
 PROJECT_ROOT = pathlib.Path(__file__).parent.parent.parent.resolve()
 
@@ -15,57 +13,13 @@ def get_github_token() -> str:
     return github_token
 
 
-# https://docs.github.com/en/graphql/guides/forming-calls-with-graphql
-def compose_gh_api_request(pull_number: int, owner="opendatahub-io", repo="notebooks", per_page=100,
-                           cursor="") -> urllib.request.Request:
-    github_token = get_github_token()
-
-    return urllib.request.Request(
-        url="https://api.github.com/graphql",
-        method="POST",
-        headers={
-            "Authorization": f"bearer {github_token}",
-        },
-        # https://docs.github.com/en/graphql/guides/using-the-explorer
-        data=json.dumps({"query": f"""
-{{
-  repository(owner:"{owner}", name:"{repo}") {{
-    pullRequest(number:{pull_number}) {{
-      files(first:{per_page}, after:"{cursor}") {{
-        edges {{
-          node {{
-            path
-          }}
-          cursor
-        }}
-      }}
-    }}
-  }}
-}}
-    """}).encode("utf-8"),
-    )
-
-
-def list_changed_files(owner: str, repo: str, pr_number: int, per_page=100) -> list[str]:
-    files = []
-
-    logging.debug("Getting list of changed files from GitHub API")
-
-    CURSOR = ""
-    while CURSOR is not None:
-        request = compose_gh_api_request(pull_number=pr_number, owner=owner, repo=repo, per_page=per_page,
-                                         cursor=CURSOR)
-        response = urllib.request.urlopen(request)
-        data = json.loads(response.read().decode("utf-8"))
-        response.close()
-        edges = data["data"]["repository"]["pullRequest"]["files"]["edges"]
-
-        CURSOR = None
-        for edge in edges:
-            files.append(edge["node"]["path"])
-            CURSOR = edge["cursor"]
-
-    logging.debug(f"Determined {len(files)} changed files: {files[:5]} (..., printing up to 5)")
+def list_changed_files(from_ref: str, to_ref: str) -> list[str]:
+    logging.debug("Getting list of changed files from git diff")
+
+    files = subprocess.check_output(["git", "diff", "--name-only", from_ref, to_ref],
+                                    encoding='utf-8').splitlines()
+
+    logging.debug(f"Determined {len(files)} changed files: {files[:100]} (..., printing up to 100 files)")
     return files
 
 
@@ -110,12 +64,9 @@ def filter_out_unchanged(targets: list[str], changed_files: list[str]) -> list[s
 
 
 class SelfTests(unittest.TestCase):
-    def test_compose_gh_api_request__call_without_asserting(self):
-        request = compose_gh_api_request(pull_number=556, per_page=100, cursor="")
-        print(request.data)
-
-    def test_list_changed_files__pagination_works(self):
-        changed_files = list_changed_files(owner="opendatahub-io", repo="notebooks", pr_number=556, per_page=1)
+    def test_list_changed_files(self):
+        """This is PR #556 in opendatahub-io/notebooks"""
+        changed_files = list_changed_files(from_ref="4d4841f", to_ref="2c36c11")
         assert set(changed_files) == {'codeserver/ubi9-python-3.9/Dockerfile',
                                       'codeserver/ubi9-python-3.9/run-code-server.sh'}