From 986e0049e0974556787a5207fef415b6a21b2844 Mon Sep 17 00:00:00 2001
From: Andrey Fedorov <andrey.fedorov@gmail.com>
Date: Thu, 8 Aug 2024 10:34:30 -0400
Subject: [PATCH] ENH: add clinical_index

also added checks for existence of the URLs containing remote indices
---
 idc_index/index.py |  9 ++++++++-
 tests/idcindex.py  | 18 ++++++++++++++++++
 2 files changed, 26 insertions(+), 1 deletion(-)

diff --git a/idc_index/index.py b/idc_index/index.py
index 4fff30d7..8b47d9a6 100644
--- a/idc_index/index.py
+++ b/idc_index/index.py
@@ -41,7 +41,7 @@ class IDCClient:
     CITATION_FORMAT_BIBTEX = "application/x-bibtex"
 
     # Singleton pattern
-    # NOTE: In the future, one may want to use multiple clients e.g. for sub-datasets so a attribute-singleton as shown bewlo seems a better option.
+    # NOTE: In the future, one may want to use multiple clients e.g. for sub-datasets so a attribute-singleton as shown below seems a better option.
     # _instance: IDCClient
     # def __new__(cls):
     #     if not hasattr(cls, "_instance") or getattr(cls, "_instance") is None:
@@ -74,6 +74,8 @@ def __init__(self):
             {"Modality": pd.Series.unique, "series_size_MB": "sum"}
         )
 
+        idc_version = f"v{Version(idc_index_data.__version__).major}"
+
         self.indices_overview = {
             "index": {
                 "description": "Main index containing one row per DICOM series.",
@@ -95,6 +97,11 @@ def __init__(self):
                 "installed": False,
                 "url": f"{asset_endpoint_url}/sm_instance_index.parquet",
             },
+            "clinical_index": {
+                "description": "Index of clinical data accompanying the available images.",
+                "installed": False,
+                "url": f"https://idc-open-metadata.s3.amazonaws.com/bigquery_export/idc_{idc_version}_clinical/column_metadata/000000000000.parquet",
+            },
         }
 
         # Lookup s5cmd
diff --git a/tests/idcindex.py b/tests/idcindex.py
index 1ce45acb..7752b968 100644
--- a/tests/idcindex.py
+++ b/tests/idcindex.py
@@ -9,6 +9,7 @@
 
 import pandas as pd
 import pytest
+import requests
 from click.testing import CliRunner
 from idc_index import IDCClient, cli
 
@@ -18,6 +19,17 @@
 logging.basicConfig(level=logging.DEBUG)
 
 
+def remote_file_exists(url):
+    try:
+        response = requests.head(url, allow_redirects=True)
+        # Check if the status code indicates success
+        return response.status_code == 200
+    except requests.RequestException as e:
+        # Handle any exceptions (e.g., network issues)
+        print(f"An error occurred: {e}")
+        return False
+
+
 @pytest.fixture(autouse=True)
 def _change_test_dir(request, monkeypatch):
     monkeypatch.chdir(request.fspath.dirname)
@@ -494,6 +506,12 @@ def test_fetch_index(self):
         assert i.indices_overview["sm_index"]["installed"] is True
         assert hasattr(i, "sm_index")
 
+    def test_indices_urls(self):
+        i = IDCClient()
+        for index in i.indices_overview:
+            if i.indices_overview[index]["url"] is not None:
+                assert remote_file_exists(i.indices_overview[index]["url"])
+
 
 if __name__ == "__main__":
     unittest.main()