aboutsummaryrefslogtreecommitdiff
path: root/backend/hist_data/enwiki/gen_img_data.py
diff options
context:
space:
mode:
authorTerry Truong <terry06890@gmail.com>2023-02-05 13:55:00 +1100
committerTerry Truong <terry06890@gmail.com>2023-02-05 13:55:00 +1100
commitff638e269d439c853b5182b68ff10777f12597f1 (patch)
tree2cc62068a501884821969b6ebeba3686df8f4044 /backend/hist_data/enwiki/gen_img_data.py
parentcc79c17fbc05bddc8b08f2734e721bc241123a4e (diff)
Use relative imports between data generation scriptspackage-imports
Avoids the need for code that modifies sys.path, but requires running of the scripts using 'python -m' with backend/ as cwd. Also expects constants like DB_FILE to be non-relative, due to running from backend/.
Diffstat (limited to 'backend/hist_data/enwiki/gen_img_data.py')
-rwxr-xr-xbackend/hist_data/enwiki/gen_img_data.py9
1 files changed, 5 insertions, 4 deletions
diff --git a/backend/hist_data/enwiki/gen_img_data.py b/backend/hist_data/enwiki/gen_img_data.py
index 05df63d..044e5a0 100755
--- a/backend/hist_data/enwiki/gen_img_data.py
+++ b/backend/hist_data/enwiki/gen_img_data.py
@@ -16,10 +16,11 @@ import html
import urllib.parse
import sqlite3
-DUMP_FILE = 'enwiki-20220501-pages-articles-multistream.xml.bz2'
-INDEX_DB = 'dump_index.db'
-IMG_DB = 'img_data.db' # The database to create
-DB_FILE = os.path.join('..', 'data.db')
+ENWIKI_DIR = os.path.dirname(os.path.realpath(__file__))
+DUMP_FILE = os.path.join(ENWIKI_DIR, 'enwiki-20220501-pages-articles-multistream.xml.bz2')
+INDEX_DB = os.path.join(ENWIKI_DIR, 'dump_index.db')
+IMG_DB = os.path.join(ENWIKI_DIR, 'img_data.db') # The database to create
+DB_FILE = os.path.join(ENWIKI_DIR, '..', 'data.db')
ID_LINE_REGEX = re.compile(r'<id>(.*)</id>')
IMG_LINE_REGEX = re.compile(r'.*\| *image *= *([^|]*)')