Merge pull request #788 from ajstanley/secure_ssl_only

allowed i7 harvesting without checking certs
mjordan · Jun 7, 2024 · 986e8ba · 986e8ba
2 parents 066f567 + 7cdc8e6
commit 986e8ba
Show file tree

Hide file tree

Showing 2 changed files with 18 additions and 5 deletions.
diff --git a/i7Import/get_islandora_7_content.py b/i7Import/get_islandora_7_content.py
@@ -22,7 +22,9 @@
 parser = argparse.ArgumentParser()
 parser.add_argument("--config", required=True, help="Configuration file to use.")
 parser.add_argument(
-    "--metadata_solr_request", required=False, help="Option to solr metadata request."
+    "--metadata_solr_request",
+    required=False,
+    help="Option to supply solr metadata request.",
 )
 args = parser.parse_args()
 utils = i7ImportUtilities(args.config)
@@ -44,6 +46,8 @@
     metadata_solr_request = utils.get_metadata_solr_request(args.metadata_solr_request)
 else:
     metadata_solr_request = utils.get_default_metadata_solr_request()
+if config["secure_ssl_only"] is False:
+    requests.packages.urllib3.disable_warnings()
 if config["debug"]:
     pretty_print = metadata_solr_request.replace("&", "\n&")
     print(f"Solr request: {pretty_print}")
@@ -83,7 +87,7 @@
 row_count = 0
 pbar = InitBar()
 num_csv_rows = len(rows)
-print(f"Processing {num_csv_rows -1}.")
+print(f"Processing {num_csv_rows - 1}.")
 with open(config["csv_output_path"], "w", newline="") as csvfile:
     writer = csv.DictWriter(csvfile, fieldnames=headers)
     writer.writeheader()

diff --git a/i7Import/i7ImportUtilities.py b/i7Import/i7ImportUtilities.py
@@ -55,6 +55,7 @@ def __init__(self, config_location):
         "solr_filters": False,
         "start": 0,
         "rows": 100000,
+        "secure_ssl_only": True,
     }
 
     def get_config(self):
@@ -97,7 +98,9 @@ def parse_rels_ext(self, pid):
             print(f"\n{rels_ext_url}")
         try:
             rels_ext_download_response = requests.get(
-                url=rels_ext_url, allow_redirects=True
+                verify=self.config["secure_ssl_only"],
+                url=rels_ext_url,
+                allow_redirects=True,
             )
             if rels_ext_download_response.ok:
                 rel_ext = {}
@@ -136,7 +139,9 @@ def get_default_metadata_solr_request(self):
         # then used in another query to get the populated CSV data.
         try:
             field_list_response = requests.get(
-                url=fields_solr_url, allow_redirects=True
+                verify=self.config["secure_ssl_only"],
+                url=fields_solr_url,
+                allow_redirects=True,
             )
             raw_field_list = field_list_response.content.decode()
         except requests.exceptions.RequestException as e:
@@ -196,7 +201,11 @@ def get_i7_asset(self, pid, datastream):
             if self.config["get_file_url"]:
                 obj_download_response = requests.head(url=obj_url, allow_redirects=True)
             else:
-                obj_download_response = requests.get(url=obj_url, allow_redirects=True)
+                obj_download_response = requests.get(
+                    verify=self.config["secure_ssl_only"],
+                    url=obj_url,
+                    allow_redirects=True,
+                )
             if obj_download_response.status_code == 200:
                 # Get MIMETYPE from 'Content-Type' header
                 obj_mimetype = obj_download_response.headers["content-type"]