mirrors/pferd
1
0
Fork 0
mirror of https://github.com/Garmelon/PFERD.git synced 2023-12-21 10:23:01 +01:00
Code Issues Packages Projects Releases Wiki Activity

Update local crawler to new crawler structure

Browse Source
This commit is contained in:
Joscha 2021-05-22 21:46:05 +02:00
parent e21795ee35
commit ae3d80664c
1 changed files with 20 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
PFERD/crawlers/local.py
View File

@ -56,10 +56,8 @@ class LocalCrawler(Crawler):
else:
self._block_size = 1024**2 # 1 MiB
async def crawl(self) -> None:
async def _run(self) -> None:
await self._crawl_path(self._target, PurePath())
if self.error_free:
await self.cleanup()
@anoncritical
async def _crawl_path(self, path: Path, pure: PurePath) -> None:
@ -69,9 +67,13 @@ class LocalCrawler(Crawler):
await self._crawl_file(path, pure)
async def _crawl_dir(self, path: Path, pure: PurePath) -> None:
cl = await self.crawl(pure)
if not cl:
return
tasks = []
async with self.crawl_bar(pure):
async with cl:
await asyncio.sleep(random.uniform(
0.5 * self._crawl_delay,
self._crawl_delay,
@ -79,7 +81,6 @@ class LocalCrawler(Crawler):
for child in path.iterdir():
pure_child = pure / child.name
if self.should_crawl(child):
tasks.append(self._crawl_path(child, pure_child))
await asyncio.gather(*tasks)
@ -91,7 +92,7 @@ class LocalCrawler(Crawler):
if not dl:
return
async with self.download_bar(pure) as bar:
async with dl as (bar, sink):
await asyncio.sleep(random.uniform(
0.5 * self._download_delay,
self._download_delay,
@ -99,7 +100,6 @@ class LocalCrawler(Crawler):
bar.set_total(stat.st_size)
async with dl as sink:
with open(path, "rb") as f:
while True:
data = f.read(self._block_size)