From 73a04aba3547f19209567df791561c24d640fda9 Mon Sep 17 00:00:00 2001
From: yzqzss <yzqzss@yandex.com>
Date: Wed, 9 Oct 2024 12:04:01 +0800
Subject: [PATCH] move hard_reties config to OtherConfig

---
 wikiteam3/dumpgenerator/api/page_titles.py  | 2 +-
 wikiteam3/dumpgenerator/cli/cli.py          | 7 ++++---
 wikiteam3/dumpgenerator/config.py           | 5 +++--
 wikiteam3/dumpgenerator/dump/image/image.py | 2 +-
 4 files changed, 9 insertions(+), 7 deletions(-)

diff --git a/wikiteam3/dumpgenerator/api/page_titles.py b/wikiteam3/dumpgenerator/api/page_titles.py
index 84575243..c5d8eabb 100644
--- a/wikiteam3/dumpgenerator/api/page_titles.py
+++ b/wikiteam3/dumpgenerator/api/page_titles.py
@@ -27,7 +27,7 @@ def getPageTitlesAPI(config: Config, session: requests.Session):
     delay_session = SessionMonkeyPatch(
             session=session, config=config,
             add_delay=True, delay_msg="Session delay: "+__name__,
-            hard_retries=3
+            hard_retries=3 # TODO: --hard-retries
         )
     delay_session.hijack()
     for namespace in namespaces:
diff --git a/wikiteam3/dumpgenerator/cli/cli.py b/wikiteam3/dumpgenerator/cli/cli.py
index e850e485..37e34647 100644
--- a/wikiteam3/dumpgenerator/cli/cli.py
+++ b/wikiteam3/dumpgenerator/cli/cli.py
@@ -52,7 +52,7 @@ def getArgumentParser():
         "--retries", metavar="5", default=5, help="Maximum number of retries for each request before failing."
     )
     parser.add_argument(
-        "--hard-retries", metavar="3", default=3, help="Maximum number of hard retries for each request before failing."
+        "--hard-retries", metavar="3", default=3, help="Maximum number of hard retries for each request before failing. (for now, this only controls the hard retries during images downloading)"
     )
     parser.add_argument("--path", help="path to store wiki dump at")
     parser.add_argument(
@@ -293,7 +293,7 @@ def get_parameters(params=None) -> Tuple[Config, OtherConfig]:
     # Create session
     mod_requests_text(requests) # monkey patch # type: ignore
     session = requests.Session()
-    patch_sess = SessionMonkeyPatch(session=session, hard_retries=int(args.hard_retries))
+    patch_sess = SessionMonkeyPatch(session=session, hard_retries=1) # hard retry once to avoid spending too much time on initial detection
     patch_sess.hijack()
     def print_request(r: requests.Response, *args, **kwargs):
         # TODO: use logging
@@ -534,7 +534,6 @@ def sleep(self, response=None):
         path = args.path and os.path.normpath(args.path) or "",
         delay = args.delay,
         retries = int(args.retries),
-        hard_retries = int(args.hard_retries),
     )
 
 
@@ -552,6 +551,8 @@ def sleep(self, response=None):
         assert_max_images = args.assert_max_images,
         assert_max_images_bytes = args.assert_max_images_bytes,
 
+        hard_retries = int(args.hard_retries),
+
         upload = args.upload,
         uploader_args = args.uploader_args,
     )
diff --git a/wikiteam3/dumpgenerator/config.py b/wikiteam3/dumpgenerator/config.py
index 93f533e2..517502e3 100644
--- a/wikiteam3/dumpgenerator/config.py
+++ b/wikiteam3/dumpgenerator/config.py
@@ -26,8 +26,6 @@ def asdict(self):
     """ Delay between requests """
     retries: int = 0
     """ Number of retries """
-    hard_retries: int = 0
-    """ Number of hard retries """
     path: str = ''
     """ Path to save the wikidump """
     logs: bool = False
@@ -115,5 +113,8 @@ class OtherConfig:
     assert_max_images: Optional[int] 
     assert_max_images_bytes: Optional[int] 
 
+    hard_retries: int
+    """ Number of hard retries """
+
     upload: bool 
     uploader_args: List[str]
\ No newline at end of file
diff --git a/wikiteam3/dumpgenerator/dump/image/image.py b/wikiteam3/dumpgenerator/dump/image/image.py
index de53fe42..9ecde193 100644
--- a/wikiteam3/dumpgenerator/dump/image/image.py
+++ b/wikiteam3/dumpgenerator/dump/image/image.py
@@ -107,7 +107,7 @@ def modify_headers(headers: Optional[Dict] = None) -> Dict:
             return headers
             
 
-        patch_sess = SessionMonkeyPatch(session=session, config=config, hard_retries=config.hard_retries)
+        patch_sess = SessionMonkeyPatch(session=session, config=config, hard_retries=other.hard_retries)
         patch_sess.hijack()
 
         ia_session = requests.Session()