fix LFS leftovers and expired https cert

capreolus-ir · Jun 9, 2020 · 8627404 · 8627404
1 parent 4380c87
commit 8627404
Show file tree

Hide file tree

Showing 4 changed files with 23 additions and 14 deletions.
diff --git a/MANIFEST.in b/MANIFEST.in
@@ -1,12 +1,16 @@
-include capreolus/utils/cache.capnp
-include capreolus/collection/dummy.yaml
-include capreolus/collection/dummy_folds.json
-include capreolus/collection/qrels.dummy.txt
-include capreolus/collection/topics.dummy.txt
-include capreolus/collection/dummy/data/dummy_trec_doc
-include capreolus/collection/qrels.robust2004.txt
-include capreolus/collection/rob04_cedr_folds.json
-include capreolus/collection/rob04_yang19_folds.json
-include capreolus/collection/rob04_yang19_rm3.run
-include capreolus/collection/robust04.yaml
-include capreolus/collection/topics.robust04.301-450.601-700.txt
+include capreolus/data/antique.json
+include capreolus/data/dummy/data/dummy_trec_doc
+include capreolus/data/dummy_folds.json
+include capreolus/data/dummy.yaml
+include capreolus/data/msmarcopassage.folds.json
+include capreolus/data/qrels.antique.txt
+include capreolus/data/qrels.dummy.txt
+include capreolus/data/qrels.msmarcopassage.txt
+include capreolus/data/qrels.robust2004.txt
+include capreolus/data/rob04_cedr_folds.json
+include capreolus/data/rob04_yang19_folds.json
+include capreolus/data/rob04_yang19_rm3.run
+include capreolus/data/robust04.yaml
+include capreolus/data/topics.antique.txt
+include capreolus/data/topics.dummy.txt
+include capreolus/data/topics.robust04.301-450.601-700.txt
diff --git a/capreolus/collection/__init__.py b/capreolus/collection/__init__.py
@@ -195,7 +195,7 @@ class ANTIQUE(Collection):
     generator_type = "DefaultLuceneDocumentGenerator"
 
     def download_if_missing(self):
-        url = "https://ciir.cs.umass.edu/downloads/Antique/antique-collection.txt"
+        url = "http://ciir.cs.umass.edu/downloads/Antique/antique-collection.txt"
         cachedir = self.get_cache_path()
         document_dir = os.path.join(cachedir, "documents")
         coll_filename = os.path.join(document_dir, "antique-collection.txt")
@@ -232,7 +232,10 @@ def _convert_to_trec(self, inp_path, outp_path):
 
     def _validate_document_path(self, path):
         """ Checks that the sha256sum is correct """
-        return hash_file(path) == "409e0960f918970977ceab9e5b1d372f45395af25d53b95644bdc9ccbbf973da"
+        return (
+            hash_file(os.path.join(path, "antique-collection.txt"))
+            == "409e0960f918970977ceab9e5b1d372f45395af25d53b95644bdc9ccbbf973da"
+        )
 
 
 @Collection.register

diff --git a/requirements.txt b/requirements.txt
@@ -27,6 +27,7 @@ fasteners==0.15
 pytest-mock==1.10.4
 mock
 pyserini==0.9.3.0
+numpy
 scipy
 sphinx
 sphinxcontrib-apidoc

diff --git a/setup.py b/setup.py
@@ -61,6 +61,7 @@ def run(self):
         "pytest-mock==1.10.4",
         "mock",
         "pyserini==0.9.3.0",
+        "numpy",
         "scipy",
         "keras",
         "google-api-python-client",
-Original file line number
+Diff line change
@@ Expand Up / @@ -27,6 +27,7 @@ fasteners==0.15 @@
     pytest-mock==1.10.4
     mock
     pyserini==0.9.3.0
+    numpy
     scipy
     sphinx
     sphinxcontrib-apidoc
@@ Expand Down @@