host clirmatrix downloads json file on jhu server

Author: ssun32

ir_datasets/datasets/clirmatrix.py

@@ -29,8 +29,10 @@ def _init():
     base_path = ir_datasets.util.home_path()/NAME
 
     def _dlc_init():
+        import json, gzip
         dlc = DownloadConfig.context(NAME, base_path)
-        clirmatrix_dlc = _DownloadConfig(dlc['downloads'].path(), parser='json')
+        with gzip.open(dlc['downloads'].path(), 'rb') as f:
+            clirmatrix_dlc = _DownloadConfig(contents = json.load(f))
         return clirmatrix_dlc
 
     _dlc = ir_datasets.util.Lazy(_dlc_init)

ir_datasets/etc/downloads.yaml

@@ -107,9 +107,9 @@ clinicaltrials:
 
 clirmatrix:
   downloads:
-    url: 'https://macavaney.us/clirmatrix_downloads.json' # TODO: move this to JHU server?
-    expected_md5: '9e70cd85ec45caa8c16061c42d1ce9b8'
-    cache_path: 'clirmatrix_downloads.json'
+    url: 'http://www.cs.jhu.edu/~shuosun/clirmatrix/data/downloads.json.gz'
+    expected_md5: '371cc532aca236759bd3602eb6ce2181'
+    cache_path: 'clirmatrix_downloads.json.gz'
 
 
 clueweb09: