Upgrade to txtai 6.0, closes #70

setup.py

@@ -40,7 +40,7 @@
         "regex>=2020.5.14",
         "rich>=12.0.1",
         "text2digits>=0.1.0",
-        "txtai[api,similarity]>=4.3.1",
+        "txtai[api,similarity]>=6.0.0",
         "txtmarker>=1.0.0",
     ],
     extras_require=extras,

src/python/paperai/index.py

@@ -122,7 +122,7 @@ def embeddings(dbfile, vectors, maxsize):
 
         # Read config and create Embeddings instance
         embeddings = Embeddings(Index.config(vectors))
-        scoring = embeddings.scoring
+        scoring = embeddings.isweighted()
 
         # Build scoring index if scoring method provided
         if scoring:

src/python/paperai/query.py

@@ -57,7 +57,7 @@ def search(embeddings, cur, query, topn, threshold):
         ]
 
         # Tokenize search query, if necessary
-        query = Tokenizer.tokenize(query) if embeddings.scoring else query
+        query = Tokenizer.tokenize(query) if embeddings.isweighted() else query
 
         # Retrieve topn * 5 to account for duplicate matches
         for result in embeddings.search(query, topn * 5):

src/python/paperai/report/common.py

@@ -29,7 +29,6 @@ def __init__(self, embeddings, db, options):
 
         # Store references to embeddings index and open database cursor
         self.embeddings = embeddings
-        self.scoring = self.embeddings.scoring
 
         self.cur = db.cursor()
 
@@ -349,7 +348,7 @@ def sections(self, uid):
         sections = []
         for sid, name, text in self.cur.fetchall():
             if (
-                not self.scoring
+                not self.embeddings.isweighted()
                 or not name
                 or not re.search(Index.SECTION_FILTER, name.lower())
                 or self.options.get("allsections")

test/python/testreport.py

@@ -25,10 +25,10 @@ def testReport1(self):
         Execute.run(Utils.PATH + "/report1.yml", 10, "md", Utils.PATH, None)
 
         hashes = [
-            ("Age.csv", "6a0c4326e5c3136ad1c46606e04e9f62"),
-            ("Heart Disease.csv", "96b144fc1566e2c0aa774d098e203922"),
-            ("Heart Failure.csv", "2e8da24e09b46f71870af9ce146ade76"),
-            ("Report1.md", "3dadaf334251d8d9b8354cd7142c5ec9"),
+            ("Age.csv", "237c2d024f758139833d681c0f7828aa"),
+            ("Heart Disease.csv", "542003f5677b35b73e0b3fff07398789"),
+            ("Heart Failure.csv", "a29c59d63d73c225102d74977efeac34"),
+            ("Report1.md", "6b143c03c154cef1c6d38e0ac1a2eee9"),
         ]
 
         # Check file hashes
@@ -45,11 +45,11 @@ def testReport2(self):
         Execute.run(Utils.PATH + "/report2.yml", 10, "md", Utils.PATH, None)
 
         hashes = [
-            ("Match.csv", "c48e3205aac18e94bac915c9dbaacc89"),
-            ("MatchSurround.csv", "df53a94064c138907d9dee9c3029c542"),
-            ("Section.csv", "e3745af5f23041f5e68bdffa8abb2e56"),
-            ("Surround.csv", "96814dd27a62c28e226cc7505fee9d80"),
-            ("Report2.md", "45c37f77cdfe8e4f213d8b3b18726cee"),
+            ("Match.csv", "4b186a90a7bc9aa7e9f65608acac3235"),
+            ("MatchSurround.csv", "4eae0e75b75f7b7b158094de217362a5"),
+            ("Section.csv", "32e6c53fc7c87595846e00db4fe4b1b8"),
+            ("Surround.csv", "aa2add9eaf4a2e1b2b48753c2172772c"),
+            ("Report2.md", "6133aa0ebcfa5a363e0e187daba8c1de"),
         ]
 
         # Check file hashes
@@ -67,9 +67,9 @@ def testReport3(self):
         Execute.run(Utils.PATH + "/report3.yml", 1, "ant", Utils.PATH, None, Utils.PATH)
 
         hashes = [
-            ("AI.csv", "94f0bead413eb71835c3f27881b29c91"),
+            ("AI.csv", "858bfb2c0026de725cb06417e639237a"),
             ("All.csv", "3bca7a39a541fa68b3ef457625fb0120"),
-            ("Report3.md", "0fc53703dace57e3403294fb8ea7e9d1"),
+            ("Report3.md", "c4cf822444643effbb502baaf44d993c"),
         ]
 
         # Check file hashes