[OJ-54453] Bump black to 26.3.1 to address CVE-2026-32274

.pre-commit-config.yaml

@@ -1,6 +1,6 @@
 repos:
   - repo: https://github.com/ambv/black
-    rev: 24.4.2
+    rev: 26.3.1
     hooks:
       - id: black
         language_version: python3.13

jf_agent/data_manifests/git/adapters/github.py

@@ -1,6 +1,7 @@
-from dateutil import parser as datetime_parser
 from typing import Generator
 
+from dateutil import parser as datetime_parser
+
 from jf_agent.data_manifests.git.adapters.manifest_adapter import ManifestAdapter
 from jf_agent.data_manifests.git.manifest import (
     GitBranchManifest,
@@ -15,11 +16,17 @@
 # TODO: Expand or generalize this to work with things other than github (BBCloud, Gitlab, etc)
 class GithubManifestGenerator(ManifestAdapter):
     '''
-    Basic client for probing a GH instance. 
+    Basic client for probing a GH instance.
     '''
 
     def __init__(
-        self, token: str, base_url: str, company: str, org: str, instance: str, verify=True,
+        self,
+        token: str,
+        base_url: str,
+        company: str,
+        org: str,
+        instance: str,
+        verify=True,
     ) -> None:
         # Super class fields
         self.company = company
@@ -50,12 +57,14 @@
                 user_count=repo['users']['totalCount'],
                 pull_request_count=repo['prs']['totalCount'],
                 branch_count=repo['branches']['totalCount'],
-                commits_on_default_branch=repo['defaultBranch']['target']['history']['totalCount']
-                if repo['defaultBranch']
-                else 0,
-                default_branch_name=repo['defaultBranch']['name']
-                if repo['defaultBranch']
-                else None,
+                commits_on_default_branch=(
+                    repo['defaultBranch']['target']['history']['totalCount']
+                    if repo['defaultBranch']
+                    else 0
+                ),
+                default_branch_name=(
+                    repo['defaultBranch']['name'] if repo['defaultBranch'] else None
+                ),
             )
 
     def get_all_user_data(self) -> Generator[GitUserManifest, None, None]:

jf_agent/data_manifests/git/adapters/manifest_adapter.py

@@ -9,6 +9,7 @@
     GitUserManifest,
 )
 
+
 # TODO: Expand or generalize this to work with things other than github (BBCloud, Gitlab, etc)
 @dataclass
 class ManifestAdapter(ABC):

jf_agent/data_manifests/git/generator.py

@@ -1,18 +1,13 @@
 import logging
-from jf_agent.config_file_reader import GitConfig
 
-from jf_agent.data_manifests.git.adapters.manifest_adapter import ManifestAdapter
+from jf_ingest import logging_helper
 
-from jf_agent.data_manifests.git.manifest import (
-    GitDataManifest,
-    GitRepoManifest,
-    GitUserManifest,
-)
+from jf_agent.config_file_reader import GitConfig
 from jf_agent.data_manifests.git.adapters.github import GithubManifestGenerator
+from jf_agent.data_manifests.git.adapters.manifest_adapter import ManifestAdapter
+from jf_agent.data_manifests.git.manifest import GitDataManifest, GitRepoManifest, GitUserManifest
 from jf_agent.data_manifests.manifest import ManifestSource
 
-from jf_ingest import logging_helper
-
 logger = logging.getLogger(__name__)
 
 
@@ -145,7 +140,11 @@ def create_manifests(
 
 
 def get_manifest_adapter(
-    company_slug: str, git_creds: dict, git_config: GitConfig, instance: str, org: str,
+    company_slug: str,
+    git_creds: dict,
+    git_config: GitConfig,
+    instance: str,
+    org: str,
 ):
     if git_config.git_provider != 'github':
         raise UnsupportedGitProvider(

jf_agent/data_manifests/git/manifest.py

@@ -4,12 +4,12 @@
 
 from jf_agent.data_manifests.manifest import Manifest, ManifestSource
 
-
 IGitDataManifest = TypeVar('IGitDataManifest', bound='GitDataManifest')
 IGitRepoManifest = TypeVar('IGitRepoManifest', bound='GitRepoManifest')
 IGitUserManifest = TypeVar('IGitUserManifest', bound='GitUserManifest')
 IGitPullRequestManifest = TypeVar('IGitPullRequestManifest', bound='GitPullRequestManifest')
 
+
 # This is the parent class for all GitManifest type classes. It inherits
 # from manifests, but ensures that all 'GitManifests' have an instance
 @dataclass

jf_agent/data_manifests/jira/generator.py

@@ -1,12 +1,11 @@
-from concurrent.futures import ThreadPoolExecutor
 import logging
+from concurrent.futures import ThreadPoolExecutor
 
-from jf_agent.data_manifests.jira.adapter import JiraCloudManifestAdapter
+from jira import JIRAError
 
+from jf_agent.data_manifests.jira.adapter import JiraCloudManifestAdapter
 from jf_agent.data_manifests.jira.manifest import JiraDataManifest, JiraProjectManifest
 from jf_agent.data_manifests.manifest import ManifestSource
-from jira import JIRAError
-
 
 logger = logging.getLogger(__name__)
 

jf_agent/data_manifests/manifest.py

@@ -1,17 +1,15 @@
 import gzip
+import json
 import logging
 from abc import ABC, abstractmethod
 from dataclasses import dataclass, field
 from datetime import datetime
 from enum import Enum
-import json
 from typing import Optional, TypeVar
 
 import requests
-
 from jf_ingest import logging_helper
 
-
 logger = logging.getLogger(__name__)
 
 IManifest = TypeVar('IManifest', bound='Manifest')
@@ -102,7 +100,9 @@ def to_local_file(
         with open(file_name, 'w') as f:
             f.write(self.to_json_str())
 
-    def upload_to_s3(self, jellyfish_api_base: str, jellyfish_api_token: str, skip_ssl_verification: bool = False) -> None:
+    def upload_to_s3(
+        self, jellyfish_api_base: str, jellyfish_api_token: str, skip_ssl_verification: bool = False
+    ) -> None:
         headers = {'Jellyfish-API-Token': jellyfish_api_token, 'content-encoding': 'gzip'}
 
         logger.info(f'Attempting to upload {self.get_unique_key()} manifest to s3...')

jf_agent/git/__init__.py

@@ -373,9 +373,9 @@ def load_and_dump_git(
             else:
                 from jf_agent.git.gitlab_adapter import GitLabAdapter
 
-                GitLabAdapter(config, outdir, compress_output_files, git_connection).load_and_dump_git(
-                    endpoint_git_instance_info
-                )
+                GitLabAdapter(
+                    config, outdir, compress_output_files, git_connection
+                ).load_and_dump_git(endpoint_git_instance_info)
         elif config.git_provider == ADO_PROVIDER:
             for jf_ingest_git_config in jf_ingest_config.git_configs:
                 if jf_ingest_git_config.instance_slug == instance_slug:

jf_agent/git/bitbucket_cloud_adapter.py

@@ -1,29 +1,29 @@
-from jf_agent.git.utils import get_branches_for_standardized_repo
-from tqdm import tqdm
+import logging
 import re
-from dateutil import parser
 from typing import List
-import logging
+
 import requests
+from dateutil import parser
+from jf_ingest import diagnostics, logging_helper
+from tqdm import tqdm
 
+from jf_agent.config_file_reader import GitConfig
 from jf_agent.git import (
     GitAdapter,
-    StandardizedUser,
-    StandardizedProject,
     StandardizedBranch,
-    StandardizedRepository,
     StandardizedCommit,
+    StandardizedProject,
     StandardizedPullRequest,
     StandardizedPullRequestComment,
     StandardizedPullRequestReview,
+    StandardizedRepository,
     StandardizedShortRepository,
+    StandardizedUser,
     pull_since_date_for_repo,
 )
 from jf_agent.git.bitbucket_cloud_client import BitbucketCloudClient
-
+from jf_agent.git.utils import get_branches_for_standardized_repo
 from jf_agent.name_redactor import NameRedactor, sanitize_text
-from jf_agent.config_file_reader import GitConfig
-from jf_ingest import diagnostics, logging_helper
 
 logger = logging.getLogger(__name__)
 
@@ -68,7 +68,8 @@
     @diagnostics.capture_timing()
     @logging_helper.log_entry_exit(logger)
     def get_repos(
-        self, standardized_projects: List[StandardizedProject],
+        self,
+        standardized_projects: List[StandardizedProject],
     ) -> List[StandardizedRepository]:
         logger.info('downloading bitbucket repos...')
 
@@ -195,7 +196,9 @@
     @diagnostics.capture_timing()
     @logging_helper.log_entry_exit(logger)
     def get_pull_requests(
-        self, standardized_repos: List[StandardizedRepository], server_git_instance_info,
+        self,
+        standardized_repos: List[StandardizedRepository],
+        server_git_instance_info,
     ) -> List[StandardizedPullRequest]:
         logger.info('downloading bitbucket prs...')
         for i, repo in enumerate(
@@ -225,7 +228,9 @@
                                 or not api_pr['destination']['repository']
                             ):
                                 logging_helper.log_standard_error(
-                                    logging.WARNING, msg_args=[api_pr['id']], error_code=3030,
+                                    logging.WARNING,
+                                    msg_args=[api_pr['id']],
+                                    error_code=3030,
                                 )
                                 continue
 
@@ -257,7 +262,10 @@
                 except Exception:
                     # if something happens when pulling PRs for a repo, just keep going.
                     logging_helper.log_standard_error(
-                        logging.ERROR, msg_args=[repo.id], error_code=3021, exc_info=True,
+                        logging.ERROR,
+                        msg_args=[repo.id],
+                        error_code=3021,
+                        exc_info=True,
                     )
 
         logger.info('Done downloading PRs!')
@@ -345,9 +353,9 @@
         message=sanitize_text(api_commit['message'], strip_text_content),
         is_merge=len(api_commit['parents']) > 1,
         repo=standardized_repo.short(),  # use short form of repo
-        branch_name=branch_name
-        if not redact_names_and_urls
-        else _branch_redactor.redact_name(branch_name),
+        branch_name=(
+            branch_name if not redact_names_and_urls else _branch_redactor.redact_name(branch_name)
+        ),
     )
 
 
@@ -372,13 +380,26 @@
         username, email = m.groups()
         username = username.strip()
         email = email.strip()
-        return StandardizedUser(id=raw_name, login=email, name=username, email=email,)
+        return StandardizedUser(
+            id=raw_name,
+            login=email,
+            name=username,
+            email=email,
+        )
 
-    return StandardizedUser(id=raw_name, name=raw_name, login=raw_name,)
+    return StandardizedUser(
+        id=raw_name,
+        name=raw_name,
+        login=raw_name,
+    )
 
 
 def _standardize_pr(
-    client, repo, api_pr, strip_text_content: bool, redact_names_and_urls: bool,
+    client,
+    repo,
+    api_pr,
+    strip_text_content: bool,
+    redact_names_and_urls: bool,
 ):
     # Process the PR's diff to get additions, deletions, changed_files
     additions, deletions, changed_files = None, None, None
@@ -387,7 +408,9 @@
         additions, deletions, changed_files = _calculate_diff_counts(diff_str)
         if additions is None:
             logging_helper.log_standard_error(
-                logging.WARNING, msg_args=[api_pr["id"], repo.id], error_code=3031,
+                logging.WARNING,
+                msg_args=[api_pr["id"], repo.id],
+                error_code=3031,
             )
     except requests.exceptions.RetryError:
         # Server threw a 500 on the request for the diff and we started retrying;
@@ -401,15 +424,19 @@
         elif e.response.status_code == 401:
             # Server threw a 401 on the request for the diff; not sure why this would be, but it seems rare
             logging_helper.log_standard_error(
-                logging.WARNING, msg_args=[api_pr["id"], repo.id], error_code=3041,
+                logging.WARNING,
+                msg_args=[api_pr["id"], repo.id],
+                error_code=3041,
             )
         else:
             # Some other HTTP error happened; Re-raise
             raise
     except UnicodeDecodeError:
         # Occasional diffs seem to be invalid UTF-8
         logging_helper.log_standard_error(
-            logging.WARNING, msg_args=[api_pr["id"], repo.id], error_code=3051,
+            logging.WARNING,
+            msg_args=[api_pr["id"], repo.id],
+            error_code=3051,
         )
 
     # Comments
@@ -436,7 +463,8 @@
                 review_state='APPROVED',
             )
             for i, approval in enumerate(
-                (a['approval'] for a in activity if 'approval' in a), start=1,
+                (a['approval'] for a in activity if 'approval' in a),
+                start=1,
             )
         ]
 

jf_agent/git/bitbucket_cloud_client.py

@@ -1,14 +1,14 @@
-from collections import deque
-from datetime import datetime, timedelta
 import logging
 import time
+from collections import deque
+from datetime import datetime, timedelta
 from typing import Optional
 
 import requests
+from jf_ingest import logging_helper
 from requests.utils import default_user_agent
 
 from jf_agent.ratelimit import RateLimiter, RateLimitRealmConfig
-from jf_ingest import logging_helper
 
 logger = logging.getLogger(__name__)
 
@@ -146,7 +146,9 @@ def get_all_pages(self, url, rate_limit_realm=None, ignore404=False):
                 except requests.exceptions.HTTPError as e:
                     if e.response.status_code == 404 and ignore404:
                         # URLs are potentially sensitive data, so print instead of log!
-                        print(f'Caught a 404 for {url} - ignoring',)
+                        print(
+                            f'Caught a 404 for {url} - ignoring',
+                        )
                         return
                     raise