mirror of
https://github.com/simon987/od-database.git
synced 2025-04-19 10:26:44 +00:00
file lists now deleted after indexing
This commit is contained in:
parent
51a47b3628
commit
f452d0f8b2
@ -119,7 +119,7 @@ class HttpDirectory(RemoteDirectory):
|
|||||||
if self._isdir(anchor):
|
if self._isdir(anchor):
|
||||||
|
|
||||||
directory = File(
|
directory = File(
|
||||||
name=anchor.href,
|
name=anchor.href, # todo handle external links here
|
||||||
mtime=0,
|
mtime=0,
|
||||||
size=0,
|
size=0,
|
||||||
path=path,
|
path=path,
|
||||||
|
@ -1,4 +1,5 @@
|
|||||||
from crawl_server import logger
|
from crawl_server import logger
|
||||||
|
import os
|
||||||
from tasks import TaskResult, Task
|
from tasks import TaskResult, Task
|
||||||
import config
|
import config
|
||||||
import requests
|
import requests
|
||||||
@ -49,14 +50,18 @@ class TaskManager:
|
|||||||
"result": json.dumps(task_result.to_json())
|
"result": json.dumps(task_result.to_json())
|
||||||
}
|
}
|
||||||
|
|
||||||
|
filename = "./crawled/" + str(task_result.website_id) + ".json"
|
||||||
files = {
|
files = {
|
||||||
"file_list": open("./crawled/" + str(task_result.website_id) + ".json")
|
"file_list": open(filename)
|
||||||
}
|
}
|
||||||
|
|
||||||
r = requests.post(config.SERVER_URL + "/task/complete", data=payload, files=files)
|
r = requests.post(config.SERVER_URL + "/task/complete", data=payload, files=files)
|
||||||
|
|
||||||
logger.info("RESPONSE: " + r.text)
|
logger.info("RESPONSE: " + r.text)
|
||||||
|
|
||||||
|
if os.path.exists(filename):
|
||||||
|
os.remove(filename)
|
||||||
|
|
||||||
except Exception as e:
|
except Exception as e:
|
||||||
raise e
|
raise e
|
||||||
|
|
||||||
|
Loading…
x
Reference in New Issue
Block a user