Make Slack CI reporting stronger (#21823)

* Use token

* Avoid failure

* better error

* Fix

* fix style

---------

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

utils/get_ci_error_statistics.py

@@ -4,17 +4,22 @@ import math
 import os
 import subprocess
 import time
+import traceback
 import zipfile
 from collections import Counter
 
 import requests
 
 
-def get_job_links(workflow_run_id):
+def get_job_links(workflow_run_id, token=None):
     """Extract job names and their job links in a GitHub Actions workflow run"""
 
+    headers = None
+    if token is not None:
+        headers = {"Accept": "application/vnd.github+json", "Authorization": f"Bearer {token}"}
+
     url = f"https://api.github.com/repos/huggingface/transformers/actions/runs/{workflow_run_id}/jobs?per_page=100"
-    result = requests.get(url).json()
+    result = requests.get(url, headers=headers).json()
     job_links = {}
 
     try:
@@ -22,21 +27,25 @@ def get_job_links(workflow_run_id):
         pages_to_iterate_over = math.ceil((result["total_count"] - 100) / 100)
 
         for i in range(pages_to_iterate_over):
-            result = requests.get(url + f"&page={i + 2}").json()
+            result = requests.get(url + f"&page={i + 2}", headers=headers).json()
             job_links.update({job["name"]: job["html_url"] for job in result["jobs"]})
 
         return job_links
-    except Exception as e:
-        print("Unknown error, could not fetch links.", e)
+    except Exception:
+        print(f"Unknown error, could not fetch links:\n{traceback.format_exc()}")
 
     return {}
 
 
-def get_artifacts_links(worflow_run_id):
+def get_artifacts_links(worflow_run_id, token=None):
     """Get all artifact links from a workflow run"""
 
+    headers = None
+    if token is not None:
+        headers = {"Accept": "application/vnd.github+json", "Authorization": f"Bearer {token}"}
+
     url = f"https://api.github.com/repos/huggingface/transformers/actions/runs/{worflow_run_id}/artifacts?per_page=100"
-    result = requests.get(url).json()
+    result = requests.get(url, headers=headers).json()
     artifacts = {}
 
     try:
@@ -44,12 +53,12 @@ def get_artifacts_links(worflow_run_id):
         pages_to_iterate_over = math.ceil((result["total_count"] - 100) / 100)
 
         for i in range(pages_to_iterate_over):
-            result = requests.get(url + f"&page={i + 2}").json()
+            result = requests.get(url + f"&page={i + 2}", headers=headers).json()
             artifacts.update({artifact["name"]: artifact["archive_download_url"] for artifact in result["artifacts"]})
 
         return artifacts
-    except Exception as e:
-        print("Unknown error, could not fetch links.", e)
+    except Exception:
+        print(f"Unknown error, could not fetch links:\n{traceback.format_exc()}")
 
     return {}
 
@@ -211,24 +220,19 @@ def make_github_table_per_model(reduced_by_model):
 if __name__ == "__main__":
     parser = argparse.ArgumentParser()
     # Required parameters
-    parser.add_argument(
-        "--workflow_run_id", default=None, type=str, required=True, help="A GitHub Actions workflow run id."
-    )
+    parser.add_argument("--workflow_run_id", type=str, required=True, help="A GitHub Actions workflow run id.")
     parser.add_argument(
         "--output_dir",
-        default=None,
         type=str,
         required=True,
         help="Where to store the downloaded artifacts and other result files.",
     )
-    parser.add_argument(
-        "--token", default=None, type=str, required=True, help="A token that has actions:read permission."
-    )
+    parser.add_argument("--token", default=None, type=str, help="A token that has actions:read permission.")
     args = parser.parse_args()
 
     os.makedirs(args.output_dir, exist_ok=True)
 
-    _job_links = get_job_links(args.workflow_run_id)
+    _job_links = get_job_links(args.workflow_run_id, token=args.token)
     job_links = {}
     # To deal with `workflow_call` event, where a job name is the combination of the job names in the caller and callee.
     # For example, `PyTorch 1.11 / Model tests (models/albert, single-gpu)`.
@@ -242,7 +246,7 @@ if __name__ == "__main__":
     with open(os.path.join(args.output_dir, "job_links.json"), "w", encoding="UTF-8") as fp:
         json.dump(job_links, fp, ensure_ascii=False, indent=4)
 
-    artifacts = get_artifacts_links(args.workflow_run_id)
+    artifacts = get_artifacts_links(args.workflow_run_id, token=args.token)
     with open(os.path.join(args.output_dir, "artifacts.json"), "w", encoding="UTF-8") as fp:
         json.dump(artifacts, fp, ensure_ascii=False, indent=4)