PYTHON-5126 Resync bson vector spec tests following additions (#2161)

Author: caseyclements

bson/binary.py

@@ -450,6 +450,10 @@ def from_vector(
                 raise ValueError(f"padding does not apply to {dtype=}")
         elif dtype == BinaryVectorDtype.PACKED_BIT:  # pack ints in [0, 255] as unsigned uint8
             format_str = "B"
+            if 0 <= padding > 7:
+                raise ValueError(f"{padding=}. It must be in [0,1, ..7].")
+            if padding and not vector:
+                raise ValueError("Empty vector with non-zero padding.")
         elif dtype == BinaryVectorDtype.FLOAT32:  # pack floats as float32
             format_str = "f"
             if padding:

test/bson_binary_vector/float32.json

@@ -11,6 +11,15 @@
       "padding": 0,
       "canonical_bson": "1C00000005766563746F72000A0000000927000000FE420000E04000"
     },
+    {
+      "description": "Vector with decimals and negative value FLOAT32",
+      "valid": true,
+      "vector": [127.7, -7.7],
+      "dtype_hex": "0x27",
+      "dtype_alias": "FLOAT32",
+      "padding": 0,
+      "canonical_bson": "1C00000005766563746F72000A0000000927006666FF426666F6C000"
+    },
     {
       "description": "Empty Vector FLOAT32",
       "valid": true,
@@ -35,8 +44,22 @@
       "vector": [127.0, 7.0],
       "dtype_hex": "0x27",
       "dtype_alias": "FLOAT32",
-      "padding": 3
+      "padding": 3,
+      "canonical_bson": "1C00000005766563746F72000A0000000927030000FE420000E04000"
+    },
+    {
+      "description": "Insufficient vector data with 3 bytes FLOAT32",
+      "valid": false,
+      "dtype_hex": "0x27",
+      "dtype_alias": "FLOAT32",
+      "canonical_bson": "1700000005766563746F7200050000000927002A2A2A00"
+    },
+    {
+      "description": "Insufficient vector data with 5 bytes FLOAT32",
+      "valid": false,
+      "dtype_hex": "0x27",
+      "dtype_alias": "FLOAT32",
+      "canonical_bson": "1900000005766563746F7200070000000927002A2A2A2A2A00"
     }
   ]
 }
-

test/bson_binary_vector/int8.json

@@ -42,7 +42,8 @@
       "vector": [127, 7],
       "dtype_hex": "0x03",
       "dtype_alias": "INT8",
-      "padding": 3
+      "padding": 3,
+      "canonical_bson": "1600000005766563746F7200040000000903037F0700"
     },
     {
       "description": "INT8 with float inputs",
@@ -54,4 +55,3 @@
     }
   ]
 }
-

test/bson_binary_vector/packed_bit.json

@@ -2,6 +2,15 @@
   "description": "Tests of Binary subtype 9, Vectors, with dtype PACKED_BIT",
   "test_key": "vector",
   "tests": [
+    {
+      "description": "Padding specified with no vector data PACKED_BIT",
+      "valid": false,
+      "vector": [],
+      "dtype_hex": "0x10",
+      "dtype_alias": "PACKED_BIT",
+      "padding": 1,
+      "canonical_bson": "1400000005766563746F72000200000009100100"
+    },
     {
       "description": "Simple Vector PACKED_BIT",
       "valid": true,
@@ -44,7 +53,31 @@
       "dtype_hex": "0x10",
       "dtype_alias": "PACKED_BIT",
       "padding": 0
+    },
+    {
+      "description": "Vector with float values PACKED_BIT",
+      "valid": false,
+      "vector": [127.5],
+      "dtype_hex": "0x10",
+      "dtype_alias": "PACKED_BIT",
+      "padding": 0
+    },
+    {
+      "description": "Exceeding maximum padding PACKED_BIT",
+      "valid": false,
+      "vector": [1],
+      "dtype_hex": "0x10",
+      "dtype_alias": "PACKED_BIT",
+      "padding": 8,
+      "canonical_bson": "1500000005766563746F7200030000000910080100"
+    },
+    {
+      "description": "Negative padding PACKED_BIT",
+      "valid": false,
+      "vector": [1],
+      "dtype_hex": "0x10",
+      "dtype_alias": "PACKED_BIT",
+      "padding": -1
     }
   ]
 }
-

test/test_bson_binary_vector.py

@@ -49,7 +49,7 @@ def create_test(case_spec):
     def run_test(self):
         for test_case in case_spec.get("tests", []):
             description = test_case["description"]
-            vector_exp = test_case["vector"]
+            vector_exp = test_case.get("vector", [])
             dtype_hex_exp = test_case["dtype_hex"]
             dtype_alias_exp = test_case.get("dtype_alias")
             padding_exp = test_case.get("padding", 0)
@@ -76,17 +76,27 @@ def run_test(self):
                     self.assertEqual(
                         vector_obs.dtype, BinaryVectorDtype[dtype_alias_exp], description
                     )
-                self.assertEqual(vector_obs.data, vector_exp, description)
-                self.assertEqual(vector_obs.padding, padding_exp, description)
-
+                if dtype_exp in [BinaryVectorDtype.FLOAT32]:
+                    [
+                        self.assertAlmostEqual(vector_obs.data[i], vector_exp[i], delta=1e-5)
+                        for i in range(len(vector_exp))
+                    ]
+                else:
+                    self.assertEqual(vector_obs.data, vector_exp, description)
                 # Test Binary Vector to BSON
                 vector_exp = Binary.from_vector(vector_exp, dtype_exp, padding_exp)
                 cB_obs = binascii.hexlify(encode({test_key: vector_exp})).decode().upper()
                 self.assertEqual(cB_obs, canonical_bson_exp, description)
 
             else:
                 with self.assertRaises((struct.error, ValueError), msg=description):
+                    # Tests Binary.from_vector
                     Binary.from_vector(vector_exp, dtype_exp, padding_exp)
+                    # Tests Binary.as_vector
+                    cB_exp = binascii.unhexlify(canonical_bson_exp.encode("utf8"))
+                    decoded_doc = decode(cB_exp)
+                    binary_obs = decoded_doc[test_key]
+                    binary_obs.as_vector()
 
     return run_test