aboutsummaryrefslogtreecommitdiffhomepage
path: root/src/sec_certs/utils/sanitization.py
diff options
context:
space:
mode:
Diffstat (limited to 'src/sec_certs/utils/sanitization.py')
-rw-r--r--src/sec_certs/utils/sanitization.py9
1 files changed, 9 insertions, 0 deletions
diff --git a/src/sec_certs/utils/sanitization.py b/src/sec_certs/utils/sanitization.py
index fd477746..4a7326eb 100644
--- a/src/sec_certs/utils/sanitization.py
+++ b/src/sec_certs/utils/sanitization.py
@@ -3,6 +3,8 @@ from __future__ import annotations
import html
import logging
from datetime import date
+from pathlib import Path
+from urllib.parse import urlparse
import numpy as np
import pandas as pd
@@ -23,6 +25,13 @@ def sanitize_link(record: str | None) -> str | None:
return record.replace(":443", "").replace(" ", "%20").replace("http://", "https://")
+def sanitize_link_fname(record: str | None) -> str | None:
+ if not record:
+ return None
+ parsed = urlparse(record)
+ return Path(parsed.path).name
+
+
def sanitize_cc_link(record: str | None) -> str | None:
record = sanitize_link(record)
if not record: