alex/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2026-04-05 15:27:53 +10:00
Code Issues Packages Projects Releases Wiki Activity

Add log print for url indexing

Browse Source
This commit is contained in:
JDC
2020-11-23 17:23:26 -05:00
committed by Nick Sweeting
parent 0acf479b70
commit db9c2edccc
2 changed files with 9 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
archivebox/search/utils.py
View File

@@ -1,6 +1,11 @@
from django.db.models import QuerySet
from archivebox.util import enforce_types
from archivebox.config import ANSI
def log_index_started(url):
print('{green}[*] Indexing url: {} in the search index {reset}'.format(url, **ANSI))
print( )
def get_file_result_content(res, extra_path, use_pwd=False):
if use_pwd:
@@ -12,7 +17,7 @@ def get_file_result_content(res, extra_path, use_pwd=False):
fpath = f'{fpath}/{extra_path}'
with open(fpath, 'r') as file:
data = file.read().replace('\n', '')
data = file.read()
if data:
return [data]
return []
@@ -28,6 +33,7 @@ def get_indexable_content(results: QuerySet):
if method not in ('readability', 'singlefile', 'dom', 'wget'):
return []
# This should come from a plugin interface
if method == 'readability':
return get_file_result_content(res, 'content.txt')
elif method == 'singlefile':