Release version 0.58.0

Fix retry logic for HTTP 5xx errors and network failures

CHANGES.rst

@@ -1,9 +1,69 @@
 Changelog
 =========
 
-0.56.0 (2025-12-11)
+0.58.0 (2025-12-16)
 -------------------
 ------------------------
+- Fix retry logic for HTTP 5xx errors and network failures. [Rodos]
+
+  Refactors error handling to retry all 5xx errors (not just 502), network errors (URLError, socket.error, IncompleteRead), and JSON parse errors with exponential backoff and jitter. Respects retry-after and rate limit headers per GitHub API requirements. Consolidates retry logic into make_request_with_retry() wrapper and adds clear logging for retry attempts and failures. Removes dead code from 2016 (errors list, _request_http_error, _request_url_error) that was intentionally disabled in commit 1e5a9048 to fix #29.
+
+  Fixes #140, #110, #138
+- Chore: remove transitive deps from release-requirements.txt. [Rodos]
+- Chore(deps): bump urllib3 in the python-packages group.
+  [dependabot[bot]]
+
+  Bumps the python-packages group with 1 update: [urllib3](https://github.com/urllib3/urllib3).
+
+
+  Updates `urllib3` from 2.6.1 to 2.6.2
+  - [Release notes](https://github.com/urllib3/urllib3/releases)
+  - [Changelog](https://github.com/urllib3/urllib3/blob/main/CHANGES.rst)
+  - [Commits](https://github.com/urllib3/urllib3/compare/2.6.1...2.6.2)
+
+  ---
+  updated-dependencies:
+  - dependency-name: urllib3
+    dependency-version: 2.6.2
+    dependency-type: direct:production
+    update-type: version-update:semver-patch
+    dependency-group: python-packages
+  ...
+
+
+0.57.0 (2025-12-12)
+-------------------
+- Add GitHub Apps documentation and remove outdated header. [Rodos]
+
+  - Add GitHub Apps authentication section with setup steps
+    and CI/CD workflow example using actions/create-github-app-token
+  - Remove outdated machine-man-preview header (graduated 2020)
+
+  Closes #189
+- Docs: add stdin token example to README. [Rodos]
+
+  Add example showing how to pipe a token from stdin using
+  file:///dev/stdin to avoid storing tokens in environment
+  variables or command history.
+
+  Closes #187
+- Add --skip-assets-on flag to skip release asset downloads (#135)
+  [Rodos]
+
+  Allow users to skip downloading release assets for specific repositories
+  while still backing up release metadata. Useful for starred repos with
+  large assets (e.g. syncthing with 27GB+).
+
+  Usage: --skip-assets-on repo1 repo2 owner/repo3
+
+  Features:
+  - Space-separated repos (consistent with --exclude)
+  - Case-insensitive matching
+  - Supports both repo name and owner/repo format
+
+
+0.56.0 (2025-12-11)
+-------------------
 
 Fix
 ~~~

README.rst

@@ -50,8 +50,8 @@ CLI Help output::
                   [--keychain-name OSX_KEYCHAIN_ITEM_NAME]
                   [--keychain-account OSX_KEYCHAIN_ITEM_ACCOUNT]
                   [--releases] [--latest-releases NUMBER_OF_LATEST_RELEASES]
-                  [--skip-prerelease] [--assets] [--attachments]
+                  [--skip-prerelease] [--assets] [--skip-assets-on [REPO ...]]
-                  [--exclude [REPOSITORY [REPOSITORY ...]]
+                  [--attachments] [--exclude [REPOSITORY [REPOSITORY ...]]
                   [--throttle-limit THROTTLE_LIMIT] [--throttle-pause THROTTLE_PAUSE]
                   USER
 
@@ -133,6 +133,9 @@ CLI Help output::
       --skip-prerelease     skip prerelease and draft versions; only applies if including releases
       --assets              include assets alongside release information; only
                             applies if including releases
+      --skip-assets-on [REPO ...]
+                            skip asset downloads for these repositories (e.g.
+                            --skip-assets-on repo1 owner/repo2)
       --attachments         download user-attachments from issues and pull requests
                             to issues/attachments/{issue_number}/ and
                             pulls/attachments/{pull_number}/ directories
@@ -174,6 +177,37 @@ Customise the permissions for your use case, but for a personal account full bac
 **Repository permissions**: Read access to contents, issues, metadata, pull requests, and webhooks.
 
 
+GitHub Apps
+~~~~~~~~~~~
+
+GitHub Apps are ideal for organization backups in CI/CD. Tokens are scoped to specific repositories and expire after 1 hour.
+
+**One-time setup:**
+
+1. Create a GitHub App at *Settings -> Developer Settings -> GitHub Apps -> New GitHub App*
+2. Set a name and homepage URL (can be any URL)
+3. Uncheck "Webhook > Active" (not needed for backups)
+4. Set permissions (same as fine-grained tokens above)
+5. Click "Create GitHub App", then note the **App ID** shown on the next page
+6. Under "Private keys", click "Generate a private key" and save the downloaded file
+7. Go to *Install App* in your app's settings
+8. Select the account/organization and which repositories to back up
+
+**CI/CD usage with GitHub Actions:**
+
+Store the App ID as a repository variable and the private key contents as a secret, then use ``actions/create-github-app-token``::
+
+    - uses: actions/create-github-app-token@v1
+      id: app-token
+      with:
+        app-id: ${{ vars.APP_ID }}
+        private-key: ${{ secrets.APP_PRIVATE_KEY }}
+
+    - run: github-backup myorg -t ${{ steps.app-token.outputs.token }} --as-app -o ./backup --all
+
+Note: Installation tokens expire after 1 hour. For long-running backups, use a fine-grained personal access token instead.
+
+
 Prefer SSH
 ~~~~~~~~~~
 
@@ -361,6 +395,9 @@ Debug an error/block or incomplete backup into a temporary directory. Omit "incr
 
     github-backup -f $FINE_ACCESS_TOKEN -o /tmp/github-backup/ -l debug -P --all-starred --starred --watched --followers --following --issues --issue-comments --issue-events --pulls --pull-comments --pull-commits --labels --milestones --repositories --wikis --releases --assets --pull-details --gists --starred-gists $GH_USER
 
+Pipe a token from stdin to avoid storing it in environment variables or command history (Unix-like systems only)::
+
+    my-secret-manager get github-token | github-backup user -t file:///dev/stdin -o /backup --repositories
 
 Restoring from Backup
 =====================

github_backup/__init__.py

@@ -1 +1 @@
-__version__ = "0.56.0"
+__version__ = "0.58.0"

github_backup/github_backup.py

@@ -12,6 +12,7 @@ import json
 import logging
 import os
 import platform
+import random
 import re
 import select
 import socket
@@ -19,6 +20,7 @@ import ssl
 import subprocess
 import sys
 import time
+from collections.abc import Generator
 from datetime import datetime
 from http.client import IncompleteRead
 from urllib.error import HTTPError, URLError
@@ -74,6 +76,9 @@ else:
                 "  3. Debian/Ubuntu: apt-get install ca-certificates\n\n"
             )
 
+# Retry configuration
+MAX_RETRIES = 5
+
 
 def logging_subprocess(
     popenargs, stdout_log_level=logging.DEBUG, stderr_log_level=logging.ERROR, **kwargs
@@ -440,6 +445,12 @@ def parse_args(args=None):
         dest="include_assets",
         help="include assets alongside release information; only applies if including releases",
     )
+    parser.add_argument(
+        "--skip-assets-on",
+        dest="skip_assets_on",
+        nargs="*",
+        help="skip asset downloads for these repositories",
+    )
     parser.add_argument(
         "--attachments",
         action="store_true",
@@ -561,7 +572,7 @@ def get_github_host(args):
 
 
 def read_file_contents(file_uri):
-    return open(file_uri[len(FILE_URI_PREFIX):], "rt").readline().strip()
+    return open(file_uri[len(FILE_URI_PREFIX) :], "rt").readline().strip()
 
 
 def get_github_repo_url(args, repository):
@@ -597,175 +608,181 @@ def get_github_repo_url(args, repository):
     return repo_url
 
 
-def retrieve_data_gen(args, template, query_args=None, single_request=False):
+def calculate_retry_delay(attempt, headers):
+    """Calculate delay before next retry with exponential backoff."""
+    # Respect retry-after header if present
+    if retry_after := headers.get("retry-after"):
+        return int(retry_after)
+
+    # Respect rate limit reset time
+    if int(headers.get("x-ratelimit-remaining", 1)) < 1:
+        reset_time = int(headers.get("x-ratelimit-reset", 0))
+        return max(10, reset_time - calendar.timegm(time.gmtime()))
+
+    # Exponential backoff with jitter for server errors (1s base, 120s max)
+    delay = min(1.0 * (2**attempt), 120.0)
+    return delay + random.uniform(0, delay * 0.1)
+
+
+def retrieve_data(args, template, query_args=None, paginated=True):
+    """
+    Fetch the data from GitHub API.
+
+    Handle both single requests and pagination with yield of individual dicts.
+    Handles throttling, retries, read errors, and DMCA takedowns.
+    """
+    query_args = query_args or {}
     auth = get_auth(args, encode=not args.as_app)
-    query_args = get_query_args(query_args)
     per_page = 100
-    next_url = None
 
-    while True:
+    def _extract_next_page_url(link_header):
-        if single_request:
+        for link in link_header.split(","):
-            request_per_page = None
+            if 'rel="next"' in link:
-        else:
+                return link[link.find("<") + 1:link.find(">")]
-            request_per_page = per_page
+        return None
 
-        request = _construct_request(
+    def fetch_all() -> Generator[dict, None, None]:
-            request_per_page,
+        next_url = None
-            query_args,
-            next_url or template,
-            auth,
-            as_app=args.as_app,
-            fine=True if args.token_fine is not None else False,
-        )  # noqa
-        r, errors = _get_response(request, auth, next_url or template)
 
-        status_code = int(r.getcode())
+        while True:
+            # FIRST: Fetch response
 
-        # Handle DMCA takedown (HTTP 451) - raise exception to skip entire repository
+            for attempt in range(MAX_RETRIES):
-        if status_code == 451:
+                request = _construct_request(
-            dmca_url = None
+                    per_page=per_page if paginated else None,
-            try:
+                    query_args=query_args,
-                response_data = json.loads(r.read().decode("utf-8"))
+                    template=next_url or template,
-                dmca_url = response_data.get("block", {}).get("html_url")
+                    auth=auth,
-            except Exception:
+                    as_app=args.as_app,
-                pass
+                    fine=args.token_fine is not None,
-            raise RepositoryUnavailableError(
-                "Repository unavailable due to legal reasons (HTTP 451)",
-                dmca_url=dmca_url
-            )
-
-        # Check if we got correct data
-        try:
-            response = json.loads(r.read().decode("utf-8"))
-        except IncompleteRead:
-            logger.warning("Incomplete read error detected")
-            read_error = True
-        except json.decoder.JSONDecodeError:
-            logger.warning("JSON decode error detected")
-            read_error = True
-        except TimeoutError:
-            logger.warning("Tiemout error detected")
-            read_error = True
-        else:
-            read_error = False
-
-        # be gentle with API request limit and throttle requests if remaining requests getting low
-        limit_remaining = int(r.headers.get("x-ratelimit-remaining", 0))
-        if args.throttle_limit and limit_remaining <= args.throttle_limit:
-            logger.info(
-                "API request limit hit: {} requests left, pausing further requests for {}s".format(
-                    limit_remaining, args.throttle_pause
                 )
-            )
+                http_response = make_request_with_retry(request, auth)
-            time.sleep(args.throttle_pause)
 
-        retries = 0
+                match http_response.getcode():
-        while retries < 3 and (status_code == 502 or read_error):
+                    case 200:
-            logger.warning("API request failed. Retrying in 5 seconds")
+                        # Success - Parse JSON response
-            retries += 1
+                        try:
-            time.sleep(5)
+                            response = json.loads(http_response.read().decode("utf-8"))
-            request = _construct_request(
+                            break  # Exit retry loop and handle the data returned
-                request_per_page,
+                        except (
-                query_args,
+                            IncompleteRead,
-                next_url or template,
+                            json.decoder.JSONDecodeError,
-                auth,
+                            TimeoutError,
-                as_app=args.as_app,
+                        ) as e:
-                fine=True if args.token_fine is not None else False,
+                            logger.warning(f"{type(e).__name__} reading response")
-            )  # noqa
+                            if attempt < MAX_RETRIES - 1:
-            r, errors = _get_response(request, auth, next_url or template)
+                                delay = calculate_retry_delay(attempt, {})
+                                logger.warning(
+                                    f"Retrying in {delay:.1f}s (attempt {attempt + 1}/{MAX_RETRIES})"
+                                )
+                                time.sleep(delay)
+                            continue  # Next retry attempt
 
-            status_code = int(r.getcode())
+                    case 451:
-            try:
+                        # DMCA takedown - extract URL if available, then raise
-                response = json.loads(r.read().decode("utf-8"))
+                        dmca_url = None
-                read_error = False
+                        try:
-            except IncompleteRead:
+                            response_data = json.loads(
-                logger.warning("Incomplete read error detected")
+                                http_response.read().decode("utf-8")
-                read_error = True
+                            )
-            except json.decoder.JSONDecodeError:
+                            dmca_url = response_data.get("block", {}).get("html_url")
-                logger.warning("JSON decode error detected")
+                        except Exception:
-                read_error = True
+                            pass
-            except TimeoutError:
+                        raise RepositoryUnavailableError(
-                logger.warning("Tiemout error detected")
+                            "Repository unavailable due to legal reasons (HTTP 451)",
-                read_error = True
+                            dmca_url=dmca_url,
+                        )
 
-        if status_code != 200:
+                    case _:
-            template = "API request returned HTTP {0}: {1}"
+                        raise Exception(
-            errors.append(template.format(status_code, r.reason))
+                            f"API request returned HTTP {http_response.getcode()}: {http_response.reason}"
-            raise Exception(", ".join(errors))
+                        )
+            else:
+                logger.error(
+                    f"Failed to read response after {MAX_RETRIES} attempts for {next_url or template}"
+                )
+                raise Exception(
+                    f"Failed to read response after {MAX_RETRIES} attempts for {next_url or template}"
+                )
 
-        if read_error:
+            # SECOND: Process and paginate
-            template = "API request problem reading response for {0}"
-            errors.append(template.format(request))
-            raise Exception(", ".join(errors))
 
-        if len(errors) == 0:
+            # Pause before next request if rate limit is low
-            if type(response) is list:
+            if (
-                for resp in response:
+                remaining := int(http_response.headers.get("x-ratelimit-remaining", 0))
-                    yield resp
+            ) <= (args.throttle_limit or 0):
-                # Parse Link header for next page URL (cursor-based pagination)
+                if args.throttle_limit:
-                link_header = r.headers.get("Link", "")
+                    logger.info(
-                next_url = None
+                        f"Throttling: {remaining} requests left, pausing {args.throttle_pause}s"
-                if link_header:
+                    )
-                    # Parse Link header: <https://api.github.com/...?per_page=100&after=cursor>; rel="next"
+                    time.sleep(args.throttle_pause)
-                    for link in link_header.split(","):
+
-                        if 'rel="next"' in link:
+            # Yield results
-                            next_url = link[link.find("<") + 1:link.find(">")]
+            if isinstance(response, list):
-                            break
+                yield from response
-                if not next_url:
+            elif isinstance(response, dict):
-                    break
-            elif type(response) is dict and single_request:
                 yield response
 
-        if len(errors) > 0:
+            # Check for more pages
-            raise Exception(", ".join(errors))
+            if not paginated or not (
+                next_url := _extract_next_page_url(
+                    http_response.headers.get("Link", "")
+                )
+            ):
+                break  # No more data
 
-        if single_request:
+    return list(fetch_all())
-            break
 
 
-def retrieve_data(args, template, query_args=None, single_request=False):
+def make_request_with_retry(request, auth):
-    return list(retrieve_data_gen(args, template, query_args, single_request))
+    """Make HTTP request with automatic retry for transient errors."""
 
+    def is_retryable_status(status_code, headers):
+        # Server errors are always retryable
+        if status_code in (500, 502, 503, 504):
+            return True
+        # Rate limit (403/429) is retryable if limit exhausted
+        if status_code in (403, 429):
+            return int(headers.get("x-ratelimit-remaining", 1)) < 1
+        return False
 
-def get_query_args(query_args=None):
+    for attempt in range(MAX_RETRIES):
-    if not query_args:
-        query_args = {}
-    return query_args
-
-
-def _get_response(request, auth, template):
-    retry_timeout = 3
-    errors = []
-    # We'll make requests in a loop so we can
-    # delay and retry in the case of rate-limiting
-    while True:
-        should_continue = False
         try:
-            r = urlopen(request, context=https_ctx)
+            return urlopen(request, context=https_ctx)
+
         except HTTPError as exc:
-            errors, should_continue = _request_http_error(exc, auth, errors)  # noqa
+            # HTTPError can be used as a response-like object
-            r = exc
+            if not is_retryable_status(exc.code, exc.headers):
-        except URLError as e:
+                raise  # Non-retryable error
-            logger.warning(e.reason)
+
-            should_continue, retry_timeout = _request_url_error(template, retry_timeout)
+            if attempt >= MAX_RETRIES - 1:
-            if not should_continue:
+                logger.error(f"HTTP {exc.code} failed after {MAX_RETRIES} attempts")
-                raise
-        except socket.error as e:
-            logger.warning(e.strerror)
-            should_continue, retry_timeout = _request_url_error(template, retry_timeout)
-            if not should_continue:
                 raise
 
-        if should_continue:
+            delay = calculate_retry_delay(attempt, exc.headers)
-            continue
+            logger.warning(
+                f"HTTP {exc.code}, retrying in {delay:.1f}s "
+                f"(attempt {attempt + 1}/{MAX_RETRIES})"
+            )
+            if auth is None and exc.code in (403, 429):
+                logger.info("Hint: Authenticate to raise your GitHub rate limit")
+            time.sleep(delay)
 
-        break
+        except (URLError, socket.error) as e:
-    return r, errors
+            if attempt >= MAX_RETRIES - 1:
+                logger.error(f"Connection error failed after {MAX_RETRIES} attempts: {e}")
+                raise
+            delay = calculate_retry_delay(attempt, {})
+            logger.warning(
+                f"Connection error: {e}, retrying in {delay:.1f}s "
+                f"(attempt {attempt + 1}/{MAX_RETRIES})"
+            )
+            time.sleep(delay)
+
+    raise Exception(f"Request failed after {MAX_RETRIES} attempts")  # pragma: no cover
 
 
-def _construct_request(
+def _construct_request(per_page, query_args, template, auth, as_app=None, fine=False):
-    per_page, query_args, template, auth, as_app=None, fine=False
-):
     # If template is already a full URL with query params (from Link header), use it directly
     if "?" in template and template.startswith("http"):
         request_url = template
@@ -796,9 +813,6 @@ def _construct_request(
         else:
             auth = auth.encode("ascii")
             request.add_header("Authorization", "token ".encode("ascii") + auth)
-            request.add_header(
-                "Accept", "application/vnd.github.machine-man-preview+json"
-            )
 
     log_url = template if "?" not in template else template.split("?")[0]
     if querystring:
@@ -807,52 +821,6 @@ def _construct_request(
     return request
 
 
-def _request_http_error(exc, auth, errors):
-    # HTTPError behaves like a Response so we can
-    # check the status code and headers to see exactly
-    # what failed.
-
-    should_continue = False
-    headers = exc.headers
-    limit_remaining = int(headers.get("x-ratelimit-remaining", 0))
-
-    if exc.code == 403 and limit_remaining < 1:
-        # The X-RateLimit-Reset header includes a
-        # timestamp telling us when the limit will reset
-        # so we can calculate how long to wait rather
-        # than inefficiently polling:
-        gm_now = calendar.timegm(time.gmtime())
-        reset = int(headers.get("x-ratelimit-reset", 0)) or gm_now
-        # We'll never sleep for less than 10 seconds:
-        delta = max(10, reset - gm_now)
-
-        limit = headers.get("x-ratelimit-limit")
-        logger.warning(
-            "Exceeded rate limit of {} requests; waiting {} seconds to reset".format(
-                limit, delta
-            )
-        )  # noqa
-
-        if auth is None:
-            logger.info("Hint: Authenticate to raise your GitHub rate limit")
-
-        time.sleep(delta)
-        should_continue = True
-    return errors, should_continue
-
-
-def _request_url_error(template, retry_timeout):
-    # In case of a connection timing out, we can retry a few time
-    # But we won't crash and not back-up the rest now
-    logger.info("'{}' timed out".format(template))
-    retry_timeout -= 1
-
-    if retry_timeout >= 0:
-        return True, retry_timeout
-
-    raise Exception("'{}' timed out to much, skipping!".format(template))
-
-
 class S3HTTPRedirectHandler(HTTPRedirectHandler):
     """
     A subclassed redirect handler for downloading Github assets from S3.
@@ -1480,9 +1448,11 @@ def download_attachments(
         manifest = {
             "issue_number": number,
             "issue_type": item_type,
-            "repository": f"{args.user}/{args.repository}"
+            "repository": (
-            if hasattr(args, "repository") and args.repository
+                f"{args.user}/{args.repository}"
-            else args.user,
+                if hasattr(args, "repository") and args.repository
+                else args.user
+            ),
             "manifest_updated_at": datetime.now(timezone.utc).isoformat(),
             "attachments": attachment_metadata_list,
         }
@@ -1500,7 +1470,7 @@ def download_attachments(
 
 def get_authenticated_user(args):
     template = "https://{0}/user".format(get_github_api_host(args))
-    data = retrieve_data(args, template, single_request=True)
+    data = retrieve_data(args, template, paginated=False)
     return data[0]
 
 
@@ -1514,7 +1484,7 @@ def check_git_lfs_install():
 
 def retrieve_repositories(args, authenticated_user):
     logger.info("Retrieving repositories")
-    single_request = False
+    paginated = True
     if args.user == authenticated_user["login"]:
         # we must use the /user/repos API to be able to access private repos
         template = "https://{0}/user/repos".format(get_github_api_host(args))
@@ -1537,18 +1507,16 @@ def retrieve_repositories(args, authenticated_user):
             repo_path = args.repository
         else:
             repo_path = "{0}/{1}".format(args.user, args.repository)
-        single_request = True
+        paginated = False
-        template = "https://{0}/repos/{1}".format(
+        template = "https://{0}/repos/{1}".format(get_github_api_host(args), repo_path)
-            get_github_api_host(args), repo_path
-        )
 
-    repos = retrieve_data(args, template, single_request=single_request)
+    repos = retrieve_data(args, template, paginated=paginated)
 
     if args.all_starred:
         starred_template = "https://{0}/users/{1}/starred".format(
             get_github_api_host(args), args.user
         )
-        starred_repos = retrieve_data(args, starred_template, single_request=False)
+        starred_repos = retrieve_data(args, starred_template)
         # flag each repo as starred for downstream processing
         for item in starred_repos:
             item.update({"is_starred": True})
@@ -1558,14 +1526,17 @@ def retrieve_repositories(args, authenticated_user):
         gists_template = "https://{0}/users/{1}/gists".format(
             get_github_api_host(args), args.user
         )
-        gists = retrieve_data(args, gists_template, single_request=False)
+        gists = retrieve_data(args, gists_template)
         # flag each repo as a gist for downstream processing
         for item in gists:
             item.update({"is_gist": True})
         repos.extend(gists)
 
     if args.include_starred_gists:
-        if not authenticated_user.get("login") or args.user.lower() != authenticated_user["login"].lower():
+        if (
+            not authenticated_user.get("login")
+            or args.user.lower() != authenticated_user["login"].lower()
+        ):
             logger.warning(
                 "Cannot retrieve starred gists for '%s'. GitHub only allows access to the authenticated user's starred gists.",
                 args.user,
@@ -1574,9 +1545,7 @@ def retrieve_repositories(args, authenticated_user):
             starred_gists_template = "https://{0}/gists/starred".format(
                 get_github_api_host(args)
             )
-            starred_gists = retrieve_data(
+            starred_gists = retrieve_data(args, starred_gists_template)
-                args, starred_gists_template, single_request=False
-            )
             # flag each repo as a starred gist for downstream processing
             for item in starred_gists:
                 item.update({"is_gist": True, "is_starred": True})
@@ -1673,9 +1642,11 @@ def backup_repositories(args, output_directory, repositories):
 
         include_gists = args.include_gists or args.include_starred_gists
         include_starred = args.all_starred and repository.get("is_starred")
-        if (args.include_repository or args.include_everything) or (
+        if (
-            include_gists and repository.get("is_gist")
+            (args.include_repository or args.include_everything)
-        ) or include_starred:
+            or (include_gists and repository.get("is_gist"))
+            or include_starred
+        ):
             repo_name = (
                 repository.get("name")
                 if not repository.get("is_gist")
@@ -1735,7 +1706,9 @@ def backup_repositories(args, output_directory, repositories):
                     include_assets=args.include_assets or args.include_everything,
                 )
         except RepositoryUnavailableError as e:
-            logger.warning(f"Repository {repository['full_name']} is unavailable (HTTP 451)")
+            logger.warning(
+                f"Repository {repository['full_name']} is unavailable (HTTP 451)"
+            )
             if e.dmca_url:
                 logger.warning(f"DMCA notice: {e.dmca_url}")
             logger.info(f"Skipping remaining resources for {repository['full_name']}")
@@ -1795,7 +1768,11 @@ def backup_issues(args, repo_cwd, repository, repos_template):
             modified = os.path.getmtime(issue_file)
             modified = datetime.fromtimestamp(modified).strftime("%Y-%m-%dT%H:%M:%SZ")
             if modified > issue["updated_at"]:
-                logger.info("Skipping issue {0} because it wasn't modified since last backup".format(number))
+                logger.info(
+                    "Skipping issue {0} because it wasn't modified since last backup".format(
+                        number
+                    )
+                )
                 continue
 
         if args.include_issue_comments or args.include_everything:
@@ -1837,14 +1814,14 @@ def backup_pulls(args, repo_cwd, repository, repos_template):
         pull_states = ["open", "closed"]
         for pull_state in pull_states:
             query_args["state"] = pull_state
-            _pulls = retrieve_data_gen(args, _pulls_template, query_args=query_args)
+            _pulls = retrieve_data(args, _pulls_template, query_args=query_args)
             for pull in _pulls:
                 if args.since and pull["updated_at"] < args.since:
                     break
                 if not args.since or pull["updated_at"] >= args.since:
                     pulls[pull["number"]] = pull
     else:
-        _pulls = retrieve_data_gen(args, _pulls_template, query_args=query_args)
+        _pulls = retrieve_data(args, _pulls_template, query_args=query_args)
         for pull in _pulls:
             if args.since and pull["updated_at"] < args.since:
                 break
@@ -1852,7 +1829,7 @@ def backup_pulls(args, repo_cwd, repository, repos_template):
                 pulls[pull["number"]] = retrieve_data(
                     args,
                     _pulls_template + "/{}".format(pull["number"]),
-                    single_request=True,
+                    paginated=False,
                 )[0]
 
     logger.info("Saving {0} pull requests to disk".format(len(list(pulls.keys()))))
@@ -1869,7 +1846,11 @@ def backup_pulls(args, repo_cwd, repository, repos_template):
             modified = os.path.getmtime(pull_file)
             modified = datetime.fromtimestamp(modified).strftime("%Y-%m-%dT%H:%M:%SZ")
             if modified > pull["updated_at"]:
-                logger.info("Skipping pull request {0} because it wasn't modified since last backup".format(number))
+                logger.info(
+                    "Skipping pull request {0} because it wasn't modified since last backup".format(
+                        number
+                    )
+                )
                 continue
         if args.include_pull_comments or args.include_everything:
             template = comments_regular_template.format(number)
@@ -1919,9 +1900,11 @@ def backup_milestones(args, repo_cwd, repository, repos_template):
     elif written_count == 0:
         logger.info("{0} milestones unchanged, skipped write".format(total))
     else:
-        logger.info("Saved {0} of {1} milestones to disk ({2} unchanged)".format(
+        logger.info(
-            written_count, total, total - written_count
+            "Saved {0} of {1} milestones to disk ({2} unchanged)".format(
-        ))
+                written_count, total, total - written_count
+            )
+        )
 
 
 def backup_labels(args, repo_cwd, repository, repos_template):
@@ -1975,6 +1958,20 @@ def backup_releases(args, repo_cwd, repository, repos_template, include_assets=F
         )
         releases = releases[: args.number_of_latest_releases]
 
+    # Check if this repo should skip asset downloads (case-insensitive)
+    skip_assets = False
+    if include_assets:
+        repo_name = repository.get("name", "").lower()
+        repo_full_name = repository.get("full_name", "").lower()
+        skip_repos = [r.lower() for r in (args.skip_assets_on or [])]
+        skip_assets = repo_name in skip_repos or repo_full_name in skip_repos
+        if skip_assets:
+            logger.info(
+                "Skipping assets for {0} ({1} releases) due to --skip-assets-on".format(
+                    repository.get("name"), len(releases)
+                )
+            )
+
     # for each release, store it
     written_count = 0
     for release in releases:
@@ -1986,7 +1983,7 @@ def backup_releases(args, repo_cwd, repository, repos_template, include_assets=F
         if json_dump_if_changed(release, output_filepath):
             written_count += 1
 
-        if include_assets:
+        if include_assets and not skip_assets:
             assets = retrieve_data(args, release["assets_url"])
             if len(assets) > 0:
                 # give release asset files somewhere to live & download them (not including source archives)
@@ -2008,9 +2005,11 @@ def backup_releases(args, repo_cwd, repository, repos_template, include_assets=F
     elif written_count == 0:
         logger.info("{0} releases unchanged, skipped write".format(total))
     else:
-        logger.info("Saved {0} of {1} releases to disk ({2} unchanged)".format(
+        logger.info(
-            written_count, total, total - written_count
+            "Saved {0} of {1} releases to disk ({2} unchanged)".format(
-        ))
+                written_count, total, total - written_count
+            )
+        )
 
 
 def fetch_repository(
@@ -2024,9 +2023,12 @@ def fetch_repository(
 ):
     if bare_clone:
         if os.path.exists(local_dir):
-            clone_exists = subprocess.check_output(
+            clone_exists = (
-                ["git", "rev-parse", "--is-bare-repository"], cwd=local_dir
+                subprocess.check_output(
-            ) == b"true\n"
+                    ["git", "rev-parse", "--is-bare-repository"], cwd=local_dir
+                )
+                == b"true\n"
+            )
         else:
             clone_exists = False
     else:
@@ -2047,7 +2049,9 @@ def fetch_repository(
             )
         else:
             logger.info(
-                "Skipping {0} (repository not accessible - may be empty, private, or credentials invalid)".format(name)
+                "Skipping {0} (repository not accessible - may be empty, private, or credentials invalid)".format(
+                    name
+                )
             )
         return
 

release-requirements.txt

@@ -1,40 +1,15 @@
+# Linting & Formatting
 autopep8==2.3.2
 black==25.12.0
-bleach==6.3.0
-certifi==2025.11.12
-charset-normalizer==3.4.4
-click==8.3.1
-colorama==0.4.6
-docutils==0.22.3
 flake8==7.3.0
-gitchangelog==3.0.4
+
+# Testing
 pytest==9.0.2
-idna==3.11
+
-importlib-metadata==8.7.0
+# Release & Publishing
-jaraco.classes==3.4.0
-keyring==25.7.0
-markdown-it-py==4.0.0
-mccabe==0.7.0
-mdurl==0.1.2
-more-itertools==10.8.0
-mypy-extensions==1.1.0
-packaging==25.0
-pathspec==0.12.1
-pkginfo==1.12.1.2
-platformdirs==4.5.1
-pycodestyle==2.14.0
-pyflakes==3.4.0
-Pygments==2.19.2
-readme-renderer==44.0
-requests==2.32.5
-requests-toolbelt==1.0.0
-restructuredtext-lint==2.0.2
-rfc3986==2.0.0
-rich==14.2.0
-setuptools==80.9.0
-six==1.17.0
-tqdm==4.67.1
 twine==6.2.0
-urllib3==2.6.1
+gitchangelog==3.0.4
-webencodings==0.5.1
+setuptools==80.9.0
-zipp==3.23.0
+
+# Documentation
+restructuredtext-lint==2.0.2

tests/test_http_451.py

@@ -13,7 +13,6 @@ class TestHTTP451Exception:
 
     def test_repository_unavailable_error_raised(self):
         """HTTP 451 should raise RepositoryUnavailableError with DMCA URL."""
-        # Create mock args
         args = Mock()
         args.as_app = False
         args.token_fine = None
@@ -25,7 +24,6 @@ class TestHTTP451Exception:
         args.throttle_limit = None
         args.throttle_pause = 0
 
-        # Mock HTTPError 451 response
         mock_response = Mock()
         mock_response.getcode.return_value = 451
 
@@ -41,14 +39,10 @@ class TestHTTP451Exception:
         mock_response.headers = {"x-ratelimit-remaining": "5000"}
         mock_response.reason = "Unavailable For Legal Reasons"
 
-        def mock_get_response(request, auth, template):
+        with patch("github_backup.github_backup.make_request_with_retry", return_value=mock_response):
-            return mock_response, []
-
-        with patch("github_backup.github_backup._get_response", side_effect=mock_get_response):
             with pytest.raises(github_backup.RepositoryUnavailableError) as exc_info:
-                list(github_backup.retrieve_data_gen(args, "https://api.github.com/repos/test/dmca/issues"))
+                github_backup.retrieve_data(args, "https://api.github.com/repos/test/dmca/issues")
 
-            # Check exception has DMCA URL
             assert exc_info.value.dmca_url == "https://github.com/github/dmca/blob/master/2024/11/2024-11-04-source-code.md"
             assert "451" in str(exc_info.value)
 
@@ -71,14 +65,10 @@ class TestHTTP451Exception:
         mock_response.headers = {"x-ratelimit-remaining": "5000"}
         mock_response.reason = "Unavailable For Legal Reasons"
 
-        def mock_get_response(request, auth, template):
+        with patch("github_backup.github_backup.make_request_with_retry", return_value=mock_response):
-            return mock_response, []
-
-        with patch("github_backup.github_backup._get_response", side_effect=mock_get_response):
             with pytest.raises(github_backup.RepositoryUnavailableError) as exc_info:
-                list(github_backup.retrieve_data_gen(args, "https://api.github.com/repos/test/dmca/issues"))
+                github_backup.retrieve_data(args, "https://api.github.com/repos/test/dmca/issues")
 
-            # Exception raised even without DMCA URL
             assert exc_info.value.dmca_url is None
             assert "451" in str(exc_info.value)
 
@@ -101,42 +91,9 @@ class TestHTTP451Exception:
         mock_response.headers = {"x-ratelimit-remaining": "5000"}
         mock_response.reason = "Unavailable For Legal Reasons"
 
-        def mock_get_response(request, auth, template):
+        with patch("github_backup.github_backup.make_request_with_retry", return_value=mock_response):
-            return mock_response, []
-
-        with patch("github_backup.github_backup._get_response", side_effect=mock_get_response):
             with pytest.raises(github_backup.RepositoryUnavailableError):
-                list(github_backup.retrieve_data_gen(args, "https://api.github.com/repos/test/dmca/issues"))
+                github_backup.retrieve_data(args, "https://api.github.com/repos/test/dmca/issues")
-
-    def test_other_http_errors_unchanged(self):
-        """Other HTTP errors should still raise generic Exception."""
-        args = Mock()
-        args.as_app = False
-        args.token_fine = None
-        args.token_classic = None
-        args.username = None
-        args.password = None
-        args.osx_keychain_item_name = None
-        args.osx_keychain_item_account = None
-        args.throttle_limit = None
-        args.throttle_pause = 0
-
-        mock_response = Mock()
-        mock_response.getcode.return_value = 404
-        mock_response.read.return_value = b'{"message": "Not Found"}'
-        mock_response.headers = {"x-ratelimit-remaining": "5000"}
-        mock_response.reason = "Not Found"
-
-        def mock_get_response(request, auth, template):
-            return mock_response, []
-
-        with patch("github_backup.github_backup._get_response", side_effect=mock_get_response):
-            # Should raise generic Exception, not RepositoryUnavailableError
-            with pytest.raises(Exception) as exc_info:
-                list(github_backup.retrieve_data_gen(args, "https://api.github.com/repos/test/notfound/issues"))
-
-            assert not isinstance(exc_info.value, github_backup.RepositoryUnavailableError)
-            assert "404" in str(exc_info.value)
 
 
 if __name__ == "__main__":

tests/test_pagination.py

@@ -40,7 +40,7 @@ class MockHTTPResponse:
 
 @pytest.fixture
 def mock_args():
-    """Mock args for retrieve_data_gen."""
+    """Mock args for retrieve_data."""
     args = Mock()
     args.as_app = False
     args.token_fine = None
@@ -77,10 +77,8 @@ def test_cursor_based_pagination(mock_args):
         return responses[len(requests_made) - 1]
 
     with patch("github_backup.github_backup.urlopen", side_effect=mock_urlopen):
-        results = list(
+        results = github_backup.retrieve_data(
-            github_backup.retrieve_data_gen(
+            mock_args, "https://api.github.com/repos/owner/repo/issues"
-                mock_args, "https://api.github.com/repos/owner/repo/issues"
-            )
         )
 
     # Verify all items retrieved and cursor was used in second request
@@ -112,10 +110,8 @@ def test_page_based_pagination(mock_args):
         return responses[len(requests_made) - 1]
 
     with patch("github_backup.github_backup.urlopen", side_effect=mock_urlopen):
-        results = list(
+        results = github_backup.retrieve_data(
-            github_backup.retrieve_data_gen(
+            mock_args, "https://api.github.com/repos/owner/repo/pulls"
-                mock_args, "https://api.github.com/repos/owner/repo/pulls"
-            )
         )
 
     # Verify all items retrieved and page parameter was used (not cursor)
@@ -142,10 +138,8 @@ def test_no_link_header_stops_pagination(mock_args):
         return responses[len(requests_made) - 1]
 
     with patch("github_backup.github_backup.urlopen", side_effect=mock_urlopen):
-        results = list(
+        results = github_backup.retrieve_data(
-            github_backup.retrieve_data_gen(
+            mock_args, "https://api.github.com/repos/owner/repo/labels"
-                mock_args, "https://api.github.com/repos/owner/repo/labels"
-            )
         )
 
     # Verify pagination stopped after first request

tests/test_retrieve_data.py

@@ -0,0 +1,365 @@
+"""Tests for retrieve_data function."""
+
+import json
+import socket
+from unittest.mock import Mock, patch
+from urllib.error import HTTPError, URLError
+
+import pytest
+
+from github_backup import github_backup
+from github_backup.github_backup import (
+    MAX_RETRIES,
+    calculate_retry_delay,
+    make_request_with_retry,
+)
+
+
+class TestCalculateRetryDelay:
+    def test_respects_retry_after_header(self):
+        headers = {'retry-after': '30'}
+        assert calculate_retry_delay(0, headers) == 30
+
+    def test_respects_rate_limit_reset(self):
+        import time
+        import calendar
+        # Set reset time 60 seconds in the future
+        future_reset = calendar.timegm(time.gmtime()) + 60
+        headers = {
+            'x-ratelimit-remaining': '0',
+            'x-ratelimit-reset': str(future_reset)
+        }
+        delay = calculate_retry_delay(0, headers)
+        # Should be approximately 60 seconds (with some tolerance for execution time)
+        assert 55 <= delay <= 65
+
+    def test_exponential_backoff(self):
+        delay_0 = calculate_retry_delay(0, {})
+        delay_1 = calculate_retry_delay(1, {})
+        delay_2 = calculate_retry_delay(2, {})
+        # Base delay is 1s, so delays should be roughly 1, 2, 4 (plus jitter)
+        assert 0.9 <= delay_0 <= 1.2  # ~1s + up to 10% jitter
+        assert 1.8 <= delay_1 <= 2.4  # ~2s + up to 10% jitter
+        assert 3.6 <= delay_2 <= 4.8  # ~4s + up to 10% jitter
+
+    def test_max_delay_cap(self):
+        # Very high attempt number should not exceed 120s + jitter
+        delay = calculate_retry_delay(100, {})
+        assert delay <= 120 * 1.1  # 120s max + 10% jitter
+
+    def test_minimum_rate_limit_delay(self):
+        import time
+        import calendar
+        # Set reset time in the past (already reset)
+        past_reset = calendar.timegm(time.gmtime()) - 100
+        headers = {
+            'x-ratelimit-remaining': '0',
+            'x-ratelimit-reset': str(past_reset)
+        }
+        delay = calculate_retry_delay(0, headers)
+        # Should be minimum 10 seconds even if reset time is in past
+        assert delay >= 10
+
+
+class TestRetrieveDataRetry:
+    """Tests for retry behavior in retrieve_data."""
+
+    @pytest.fixture
+    def mock_args(self):
+        args = Mock()
+        args.as_app = False
+        args.token_fine = None
+        args.token_classic = "fake_token"
+        args.username = None
+        args.password = None
+        args.osx_keychain_item_name = None
+        args.osx_keychain_item_account = None
+        args.throttle_limit = None
+        args.throttle_pause = 0
+        return args
+
+    def test_json_parse_error_retries_and_fails(self, mock_args):
+        """HTTP 200 with invalid JSON should retry and eventually fail."""
+        mock_response = Mock()
+        mock_response.getcode.return_value = 200
+        mock_response.read.return_value = b"not valid json {"
+        mock_response.headers = {"x-ratelimit-remaining": "5000"}
+
+        call_count = 0
+
+        def mock_make_request(*args, **kwargs):
+            nonlocal call_count
+            call_count += 1
+            return mock_response
+
+        with patch("github_backup.github_backup.make_request_with_retry", side_effect=mock_make_request):
+            with patch("github_backup.github_backup.calculate_retry_delay", return_value=0):  # No delay in tests
+                with pytest.raises(Exception) as exc_info:
+                    github_backup.retrieve_data(mock_args, "https://api.github.com/repos/test/repo/issues")
+
+        assert "Failed to read response after" in str(exc_info.value)
+        assert call_count == MAX_RETRIES
+
+    def test_json_parse_error_recovers_on_retry(self, mock_args):
+        """HTTP 200 with invalid JSON should succeed if retry returns valid JSON."""
+        bad_response = Mock()
+        bad_response.getcode.return_value = 200
+        bad_response.read.return_value = b"not valid json {"
+        bad_response.headers = {"x-ratelimit-remaining": "5000"}
+
+        good_response = Mock()
+        good_response.getcode.return_value = 200
+        good_response.read.return_value = json.dumps([{"id": 1}]).encode("utf-8")
+        good_response.headers = {"x-ratelimit-remaining": "5000", "Link": ""}
+
+        responses = [bad_response, bad_response, good_response]
+        call_count = 0
+
+        def mock_make_request(*args, **kwargs):
+            nonlocal call_count
+            result = responses[call_count]
+            call_count += 1
+            return result
+
+        with patch("github_backup.github_backup.make_request_with_retry", side_effect=mock_make_request):
+            with patch("github_backup.github_backup.calculate_retry_delay", return_value=0):
+                result = github_backup.retrieve_data(mock_args, "https://api.github.com/repos/test/repo/issues")
+
+        assert result == [{"id": 1}]
+        assert call_count == 3  # Failed twice, succeeded on third
+
+    def test_http_error_raises_exception(self, mock_args):
+        """Non-success HTTP status codes should raise Exception."""
+        mock_response = Mock()
+        mock_response.getcode.return_value = 404
+        mock_response.read.return_value = b'{"message": "Not Found"}'
+        mock_response.headers = {"x-ratelimit-remaining": "5000"}
+        mock_response.reason = "Not Found"
+
+        with patch("github_backup.github_backup.make_request_with_retry", return_value=mock_response):
+            with pytest.raises(Exception) as exc_info:
+                github_backup.retrieve_data(mock_args, "https://api.github.com/repos/test/notfound/issues")
+
+            assert not isinstance(exc_info.value, github_backup.RepositoryUnavailableError)
+            assert "404" in str(exc_info.value)
+
+
+class TestMakeRequestWithRetry:
+    """Tests for HTTP error retry behavior in make_request_with_retry."""
+
+    def test_502_error_retries_and_succeeds(self):
+        """HTTP 502 should retry and succeed if subsequent request works."""
+        good_response = Mock()
+        good_response.read.return_value = b'{"ok": true}'
+
+        call_count = 0
+        fail_count = MAX_RETRIES - 1  # Fail all but last attempt
+
+        def mock_urlopen(*args, **kwargs):
+            nonlocal call_count
+            call_count += 1
+            if call_count <= fail_count:
+                raise HTTPError(
+                    url="https://api.github.com/test",
+                    code=502,
+                    msg="Bad Gateway",
+                    hdrs={"x-ratelimit-remaining": "5000"},
+                    fp=None,
+                )
+            return good_response
+
+        with patch("github_backup.github_backup.urlopen", side_effect=mock_urlopen):
+            with patch("github_backup.github_backup.calculate_retry_delay", return_value=0):
+                result = make_request_with_retry(Mock(), None)
+
+        assert result == good_response
+        assert call_count == MAX_RETRIES
+
+    def test_503_error_retries_until_exhausted(self):
+        """HTTP 503 should retry MAX_RETRIES times then raise."""
+        call_count = 0
+
+        def mock_urlopen(*args, **kwargs):
+            nonlocal call_count
+            call_count += 1
+            raise HTTPError(
+                url="https://api.github.com/test",
+                code=503,
+                msg="Service Unavailable",
+                hdrs={"x-ratelimit-remaining": "5000"},
+                fp=None,
+            )
+
+        with patch("github_backup.github_backup.urlopen", side_effect=mock_urlopen):
+            with patch("github_backup.github_backup.calculate_retry_delay", return_value=0):
+                with pytest.raises(HTTPError) as exc_info:
+                    make_request_with_retry(Mock(), None)
+
+        assert exc_info.value.code == 503
+        assert call_count == MAX_RETRIES
+
+    def test_404_error_not_retried(self):
+        """HTTP 404 should not be retried - raise immediately."""
+        call_count = 0
+
+        def mock_urlopen(*args, **kwargs):
+            nonlocal call_count
+            call_count += 1
+            raise HTTPError(
+                url="https://api.github.com/test",
+                code=404,
+                msg="Not Found",
+                hdrs={"x-ratelimit-remaining": "5000"},
+                fp=None,
+            )
+
+        with patch("github_backup.github_backup.urlopen", side_effect=mock_urlopen):
+            with pytest.raises(HTTPError) as exc_info:
+                make_request_with_retry(Mock(), None)
+
+        assert exc_info.value.code == 404
+        assert call_count == 1  # No retries
+
+    def test_rate_limit_403_retried_when_remaining_zero(self):
+        """HTTP 403 with x-ratelimit-remaining=0 should retry."""
+        good_response = Mock()
+        call_count = 0
+
+        def mock_urlopen(*args, **kwargs):
+            nonlocal call_count
+            call_count += 1
+            if call_count == 1:
+                raise HTTPError(
+                    url="https://api.github.com/test",
+                    code=403,
+                    msg="Forbidden",
+                    hdrs={"x-ratelimit-remaining": "0"},
+                    fp=None,
+                )
+            return good_response
+
+        with patch("github_backup.github_backup.urlopen", side_effect=mock_urlopen):
+            with patch("github_backup.github_backup.calculate_retry_delay", return_value=0):
+                result = make_request_with_retry(Mock(), None)
+
+        assert result == good_response
+        assert call_count == 2
+
+    def test_403_not_retried_when_remaining_nonzero(self):
+        """HTTP 403 with x-ratelimit-remaining>0 should not retry (permission error)."""
+        call_count = 0
+
+        def mock_urlopen(*args, **kwargs):
+            nonlocal call_count
+            call_count += 1
+            raise HTTPError(
+                url="https://api.github.com/test",
+                code=403,
+                msg="Forbidden",
+                hdrs={"x-ratelimit-remaining": "5000"},
+                fp=None,
+            )
+
+        with patch("github_backup.github_backup.urlopen", side_effect=mock_urlopen):
+            with pytest.raises(HTTPError) as exc_info:
+                make_request_with_retry(Mock(), None)
+
+        assert exc_info.value.code == 403
+        assert call_count == 1  # No retries
+
+    def test_connection_error_retries_and_succeeds(self):
+        """URLError (connection error) should retry and succeed if subsequent request works."""
+        good_response = Mock()
+        call_count = 0
+        fail_count = MAX_RETRIES - 1  # Fail all but last attempt
+
+        def mock_urlopen(*args, **kwargs):
+            nonlocal call_count
+            call_count += 1
+            if call_count <= fail_count:
+                raise URLError("Connection refused")
+            return good_response
+
+        with patch("github_backup.github_backup.urlopen", side_effect=mock_urlopen):
+            with patch("github_backup.github_backup.calculate_retry_delay", return_value=0):
+                result = make_request_with_retry(Mock(), None)
+
+        assert result == good_response
+        assert call_count == MAX_RETRIES
+
+    def test_socket_error_retries_until_exhausted(self):
+        """socket.error should retry MAX_RETRIES times then raise."""
+        call_count = 0
+
+        def mock_urlopen(*args, **kwargs):
+            nonlocal call_count
+            call_count += 1
+            raise socket.error("Connection reset by peer")
+
+        with patch("github_backup.github_backup.urlopen", side_effect=mock_urlopen):
+            with patch("github_backup.github_backup.calculate_retry_delay", return_value=0):
+                with pytest.raises(socket.error):
+                    make_request_with_retry(Mock(), None)
+
+        assert call_count == MAX_RETRIES
+
+
+class TestRetrieveDataThrottling:
+    """Tests for throttling behavior in retrieve_data."""
+
+    @pytest.fixture
+    def mock_args(self):
+        args = Mock()
+        args.as_app = False
+        args.token_fine = None
+        args.token_classic = "fake_token"
+        args.username = None
+        args.password = None
+        args.osx_keychain_item_name = None
+        args.osx_keychain_item_account = None
+        args.throttle_limit = 10  # Throttle when remaining <= 10
+        args.throttle_pause = 5  # Pause 5 seconds
+        return args
+
+    def test_throttling_pauses_when_rate_limit_low(self, mock_args):
+        """Should pause when x-ratelimit-remaining is at or below throttle_limit."""
+        mock_response = Mock()
+        mock_response.getcode.return_value = 200
+        mock_response.read.return_value = json.dumps([{"id": 1}]).encode("utf-8")
+        mock_response.headers = {"x-ratelimit-remaining": "5", "Link": ""}  # Below throttle_limit
+
+        with patch("github_backup.github_backup.make_request_with_retry", return_value=mock_response):
+            with patch("github_backup.github_backup.time.sleep") as mock_sleep:
+                github_backup.retrieve_data(mock_args, "https://api.github.com/repos/test/repo/issues")
+
+        mock_sleep.assert_called_once_with(5)  # throttle_pause value
+
+
+class TestRetrieveDataSingleItem:
+    """Tests for single item (dict) responses in retrieve_data."""
+
+    @pytest.fixture
+    def mock_args(self):
+        args = Mock()
+        args.as_app = False
+        args.token_fine = None
+        args.token_classic = "fake_token"
+        args.username = None
+        args.password = None
+        args.osx_keychain_item_name = None
+        args.osx_keychain_item_account = None
+        args.throttle_limit = None
+        args.throttle_pause = 0
+        return args
+
+    def test_dict_response_returned_as_list(self, mock_args):
+        """Single dict response should be returned as a list with one item."""
+        mock_response = Mock()
+        mock_response.getcode.return_value = 200
+        mock_response.read.return_value = json.dumps({"login": "testuser", "id": 123}).encode("utf-8")
+        mock_response.headers = {"x-ratelimit-remaining": "5000", "Link": ""}
+
+        with patch("github_backup.github_backup.make_request_with_retry", return_value=mock_response):
+            result = github_backup.retrieve_data(mock_args, "https://api.github.com/user")
+
+        assert result == [{"login": "testuser", "id": 123}]

tests/test_skip_assets_on.py

@@ -0,0 +1,320 @@
+"""Tests for --skip-assets-on flag behavior (issue #135)."""
+
+import pytest
+from unittest.mock import Mock, patch
+
+from github_backup import github_backup
+
+
+class TestSkipAssetsOn:
+    """Test suite for --skip-assets-on flag.
+
+    Issue #135: Allow skipping asset downloads for specific repositories
+    while still backing up release metadata.
+    """
+
+    def _create_mock_args(self, **overrides):
+        """Create a mock args object with sensible defaults."""
+        args = Mock()
+        args.user = "testuser"
+        args.output_directory = "/tmp/backup"
+        args.include_repository = False
+        args.include_everything = False
+        args.include_gists = False
+        args.include_starred_gists = False
+        args.all_starred = False
+        args.skip_existing = False
+        args.bare_clone = False
+        args.lfs_clone = False
+        args.no_prune = False
+        args.include_wiki = False
+        args.include_issues = False
+        args.include_issue_comments = False
+        args.include_issue_events = False
+        args.include_pulls = False
+        args.include_pull_comments = False
+        args.include_pull_commits = False
+        args.include_pull_details = False
+        args.include_labels = False
+        args.include_hooks = False
+        args.include_milestones = False
+        args.include_releases = True
+        args.include_assets = True
+        args.skip_assets_on = []
+        args.include_attachments = False
+        args.incremental = False
+        args.incremental_by_files = False
+        args.github_host = None
+        args.prefer_ssh = False
+        args.token_classic = "test-token"
+        args.token_fine = None
+        args.username = None
+        args.password = None
+        args.as_app = False
+        args.osx_keychain_item_name = None
+        args.osx_keychain_item_account = None
+        args.skip_prerelease = False
+        args.number_of_latest_releases = None
+
+        for key, value in overrides.items():
+            setattr(args, key, value)
+
+        return args
+
+    def _create_mock_repository(self, name="test-repo", owner="testuser"):
+        """Create a mock repository object."""
+        return {
+            "name": name,
+            "full_name": f"{owner}/{name}",
+            "owner": {"login": owner},
+            "private": False,
+            "fork": False,
+            "has_wiki": False,
+        }
+
+    def _create_mock_release(self, tag="v1.0.0"):
+        """Create a mock release object."""
+        return {
+            "tag_name": tag,
+            "name": tag,
+            "prerelease": False,
+            "draft": False,
+            "assets_url": f"https://api.github.com/repos/testuser/test-repo/releases/{tag}/assets",
+        }
+
+    def _create_mock_asset(self, name="asset.zip"):
+        """Create a mock asset object."""
+        return {
+            "name": name,
+            "url": f"https://api.github.com/repos/testuser/test-repo/releases/assets/{name}",
+        }
+
+
+class TestSkipAssetsOnArgumentParsing(TestSkipAssetsOn):
+    """Tests for --skip-assets-on argument parsing."""
+
+    def test_skip_assets_on_not_set_defaults_to_none(self):
+        """When --skip-assets-on is not specified, it should default to None."""
+        args = github_backup.parse_args(["testuser"])
+        assert args.skip_assets_on is None
+
+    def test_skip_assets_on_single_repo(self):
+        """Single --skip-assets-on should create list with one item."""
+        args = github_backup.parse_args(["testuser", "--skip-assets-on", "big-repo"])
+        assert args.skip_assets_on == ["big-repo"]
+
+    def test_skip_assets_on_multiple_repos(self):
+        """Multiple repos can be specified space-separated (like --exclude)."""
+        args = github_backup.parse_args(
+            [
+                "testuser",
+                "--skip-assets-on",
+                "big-repo",
+                "another-repo",
+                "owner/third-repo",
+            ]
+        )
+        assert args.skip_assets_on == ["big-repo", "another-repo", "owner/third-repo"]
+
+
+class TestSkipAssetsOnBehavior(TestSkipAssetsOn):
+    """Tests for --skip-assets-on behavior in backup_releases."""
+
+    @patch("github_backup.github_backup.download_file")
+    @patch("github_backup.github_backup.retrieve_data")
+    @patch("github_backup.github_backup.mkdir_p")
+    @patch("github_backup.github_backup.json_dump_if_changed")
+    def test_assets_downloaded_when_not_skipped(
+        self, mock_json_dump, mock_mkdir, mock_retrieve, mock_download
+    ):
+        """Assets should be downloaded when repo is not in skip list."""
+        args = self._create_mock_args(skip_assets_on=[])
+        repository = self._create_mock_repository(name="normal-repo")
+        release = self._create_mock_release()
+        asset = self._create_mock_asset()
+
+        mock_json_dump.return_value = True
+        mock_retrieve.side_effect = [
+            [release],  # First call: get releases
+            [asset],  # Second call: get assets
+        ]
+
+        with patch("os.path.join", side_effect=lambda *args: "/".join(args)):
+            github_backup.backup_releases(
+                args,
+                "/tmp/backup/repositories/normal-repo",
+                repository,
+                "https://api.github.com/repos/{owner}/{repo}",
+                include_assets=True,
+            )
+
+        # download_file should have been called for the asset
+        mock_download.assert_called_once()
+
+    @patch("github_backup.github_backup.download_file")
+    @patch("github_backup.github_backup.retrieve_data")
+    @patch("github_backup.github_backup.mkdir_p")
+    @patch("github_backup.github_backup.json_dump_if_changed")
+    def test_assets_skipped_when_repo_name_matches(
+        self, mock_json_dump, mock_mkdir, mock_retrieve, mock_download
+    ):
+        """Assets should be skipped when repo name is in skip list."""
+        args = self._create_mock_args(skip_assets_on=["big-repo"])
+        repository = self._create_mock_repository(name="big-repo")
+        release = self._create_mock_release()
+
+        mock_json_dump.return_value = True
+        mock_retrieve.return_value = [release]
+
+        github_backup.backup_releases(
+            args,
+            "/tmp/backup/repositories/big-repo",
+            repository,
+            "https://api.github.com/repos/{owner}/{repo}",
+            include_assets=True,
+        )
+
+        # download_file should NOT have been called
+        mock_download.assert_not_called()
+
+    @patch("github_backup.github_backup.download_file")
+    @patch("github_backup.github_backup.retrieve_data")
+    @patch("github_backup.github_backup.mkdir_p")
+    @patch("github_backup.github_backup.json_dump_if_changed")
+    def test_assets_skipped_when_full_name_matches(
+        self, mock_json_dump, mock_mkdir, mock_retrieve, mock_download
+    ):
+        """Assets should be skipped when owner/repo format matches."""
+        args = self._create_mock_args(skip_assets_on=["otheruser/big-repo"])
+        repository = self._create_mock_repository(name="big-repo", owner="otheruser")
+        release = self._create_mock_release()
+
+        mock_json_dump.return_value = True
+        mock_retrieve.return_value = [release]
+
+        github_backup.backup_releases(
+            args,
+            "/tmp/backup/repositories/big-repo",
+            repository,
+            "https://api.github.com/repos/{owner}/{repo}",
+            include_assets=True,
+        )
+
+        # download_file should NOT have been called
+        mock_download.assert_not_called()
+
+    @patch("github_backup.github_backup.download_file")
+    @patch("github_backup.github_backup.retrieve_data")
+    @patch("github_backup.github_backup.mkdir_p")
+    @patch("github_backup.github_backup.json_dump_if_changed")
+    def test_case_insensitive_matching(
+        self, mock_json_dump, mock_mkdir, mock_retrieve, mock_download
+    ):
+        """Skip matching should be case-insensitive."""
+        # User types uppercase, repo name is lowercase
+        args = self._create_mock_args(skip_assets_on=["BIG-REPO"])
+        repository = self._create_mock_repository(name="big-repo")
+        release = self._create_mock_release()
+
+        mock_json_dump.return_value = True
+        mock_retrieve.return_value = [release]
+
+        github_backup.backup_releases(
+            args,
+            "/tmp/backup/repositories/big-repo",
+            repository,
+            "https://api.github.com/repos/{owner}/{repo}",
+            include_assets=True,
+        )
+
+        # download_file should NOT have been called (case-insensitive match)
+        assert not mock_download.called
+
+    @patch("github_backup.github_backup.download_file")
+    @patch("github_backup.github_backup.retrieve_data")
+    @patch("github_backup.github_backup.mkdir_p")
+    @patch("github_backup.github_backup.json_dump_if_changed")
+    def test_multiple_skip_repos(
+        self, mock_json_dump, mock_mkdir, mock_retrieve, mock_download
+    ):
+        """Multiple repos in skip list should all be skipped."""
+        args = self._create_mock_args(skip_assets_on=["repo1", "repo2", "repo3"])
+        repository = self._create_mock_repository(name="repo2")
+        release = self._create_mock_release()
+
+        mock_json_dump.return_value = True
+        mock_retrieve.return_value = [release]
+
+        github_backup.backup_releases(
+            args,
+            "/tmp/backup/repositories/repo2",
+            repository,
+            "https://api.github.com/repos/{owner}/{repo}",
+            include_assets=True,
+        )
+
+        # download_file should NOT have been called
+        mock_download.assert_not_called()
+
+    @patch("github_backup.github_backup.download_file")
+    @patch("github_backup.github_backup.retrieve_data")
+    @patch("github_backup.github_backup.mkdir_p")
+    @patch("github_backup.github_backup.json_dump_if_changed")
+    def test_release_metadata_still_saved_when_assets_skipped(
+        self, mock_json_dump, mock_mkdir, mock_retrieve, mock_download
+    ):
+        """Release JSON should still be saved even when assets are skipped."""
+        args = self._create_mock_args(skip_assets_on=["big-repo"])
+        repository = self._create_mock_repository(name="big-repo")
+        release = self._create_mock_release()
+
+        mock_json_dump.return_value = True
+        mock_retrieve.return_value = [release]
+
+        github_backup.backup_releases(
+            args,
+            "/tmp/backup/repositories/big-repo",
+            repository,
+            "https://api.github.com/repos/{owner}/{repo}",
+            include_assets=True,
+        )
+
+        # json_dump_if_changed should have been called for release metadata
+        mock_json_dump.assert_called_once()
+        # But download_file should NOT have been called
+        mock_download.assert_not_called()
+
+    @patch("github_backup.github_backup.download_file")
+    @patch("github_backup.github_backup.retrieve_data")
+    @patch("github_backup.github_backup.mkdir_p")
+    @patch("github_backup.github_backup.json_dump_if_changed")
+    def test_non_matching_repo_still_downloads_assets(
+        self, mock_json_dump, mock_mkdir, mock_retrieve, mock_download
+    ):
+        """Repos not in skip list should still download assets."""
+        args = self._create_mock_args(skip_assets_on=["other-repo"])
+        repository = self._create_mock_repository(name="normal-repo")
+        release = self._create_mock_release()
+        asset = self._create_mock_asset()
+
+        mock_json_dump.return_value = True
+        mock_retrieve.side_effect = [
+            [release],  # First call: get releases
+            [asset],  # Second call: get assets
+        ]
+
+        with patch("os.path.join", side_effect=lambda *args: "/".join(args)):
+            github_backup.backup_releases(
+                args,
+                "/tmp/backup/repositories/normal-repo",
+                repository,
+                "https://api.github.com/repos/{owner}/{repo}",
+                include_assets=True,
+            )
+
+        # download_file SHOULD have been called
+        mock_download.assert_called_once()
+
+
+if __name__ == "__main__":
+    pytest.main([__file__, "-v"])