DEV-1185: fix AWG download of related files

Previously, requests for file metadata in the AWG environment were
unauthenticated. However, the metadata endpoint in AWG requires proper
authenticated requests. This change allows for requests for file
metadata in the AWG environment to be properly authenticated.

.pre-commit-config.yaml

@@ -1,7 +1,16 @@
----
 repos:
-  - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v2.5.0
+- repo: https://github.com/pre-commit/pre-commit-hooks
+  rev: v2.5.0
+  hooks:
+    - id: end-of-file-fixer
+    - id: no-commit-to-branch
+      args: [--branch, develop, --branch, master, --pattern, release/.*]
+- repo: https://github.com/psf/black
+  rev: 23.12.1
+  hooks:
+    - id: black
+-   repo: [email protected]:Yelp/detect-secrets
+    rev: v0.13.0
     hooks:
       - id: end-of-file-fixer
       - id: no-commit-to-branch

gdc_client/download/parser.py

@@ -80,7 +80,9 @@ def download(parser, args):
 
     ids = ids_map.values() if args.latest else ids_map.keys()
 
-    index_client = GDCIndexClient(args.server, not args.no_verify)
+    index_client = GDCIndexClient(
+        uri=args.server, token=args.token_file, verify=not args.no_verify
+    )
     client = get_client(args, index_client)
 
     # separate the smaller files from the larger files
@@ -163,7 +165,6 @@ def download(parser, args):
 
 
 def retry_download(client, url, retry_amount, no_auto_retry, wait_time):
-
     log.debug("Retrying download {0}".format(url))
 
     error = True

gdc_client/query/index.py

@@ -8,12 +8,13 @@
 
 
 class GDCIndexClient(object):
-    def __init__(self, uri, verify=True):
+    def __init__(self, uri, token=None, verify=True):
         self.uri = uri
         self.active_meta_endpoint = "/v0/files"
         self.legacy_meta_endpoint = "/v0/legacy/files"
         self.metadata = dict()
         self.verify = verify
+        self.token = token
 
     def get_related_files(self, uuid):
         # type: (str) -> list[str]
@@ -54,7 +55,8 @@ def _get_hits(self, url, metadata_query):
         """
         json_response = {}
         # using a POST request lets us avoid the MAX URL character length limit
-        r = requests.post(url, json=metadata_query, verify=self.verify)
+        headers = {"X-Auth-Token": self.token}
+        r = requests.post(url, json=metadata_query, headers=headers, verify=self.verify)
 
         if r is None:
             return []