Fix typos in tests and contrib directories (#4672)

Summary:
Pull Request resolved: https://github.com/facebookresearch/faiss/pull/4672

Fixed typos in test files and contrib utilities:

tests/:
- Fixed 20 typos across test files (C++ and Python)
- Fixed function/class name typos in Python tests:
  * test_result_order_miltihash → test_result_order_multihash
  * TestOrthognalReconstruct → TestOrthogonalReconstruct
  * test_recons_orthogona_impossible → test_recons_orthogonal_impossible
- Fixed comment typos and error messages in C++ tests

contrib/:
- Fixed 9 typos in Python contrib utilities
- Fixed string literal typos in client_server.py and rpc.py
- Fixed documentation typos in evaluation.py and big_batch_search.py

All changes maintain backward compatibility - no API changes.

Reviewed By: mnorris11

Differential Revision: D86571375

fbshipit-source-id: bd0edf7ca2e004f78e0ccba7f4aac7db6621781a

contrib/big_batch_search.py

@@ -270,7 +270,7 @@ def big_batch_search(
     compute_threads>1: the knn function will get an additional thread_no that
         tells which worker should handle this.
 
-    In threaded mode, the computation is tiled with the bucket perparation and
+    In threaded mode, the computation is tiled with the bucket preparation and
     the writeback of results (useful to maximize GPU utilization).
 
     use_float16: convert all matrices to float16 (faster for GPU gemm)
@@ -279,7 +279,7 @@ def big_batch_search(
 
     checkpointing (only for threaded > 1):
     checkpoint: file where the checkpoints are stored
-    checkpoint_freq: when to perform checkpoinging. Should be a multiple of threaded
+    checkpoint_freq: when to perform checkpointing. Should be a multiple of threaded
 
     start_list, end_list: process only a subset of invlists
     """

contrib/client_server.py

@@ -35,7 +35,7 @@ class SearchServer(rpc.Server):
 
 
 def run_index_server(index: faiss.Index, port: int, v6: bool = False):
-    """ serve requests for that index forerver """
+    """ serve requests for that index forever """
     rpc.run_server(
         lambda s: SearchServer(s, index),
         port, v6=v6)

contrib/evaluation.py

@@ -80,7 +80,7 @@ def range_PR(lims_ref, Iref, lims_new, Inew, mode="overall"):
 
 
 def counts_to_PR(ngt, nres, ninter, mode="overall"):
-    """ computes a  precision-recall for a ser of queries.
+    """ computes a  precision-recall for a set of queries.
     ngt = nb of GT results per query
     nres = nb of found results per query
     ninter = nb of correct results per query (smaller than nres of course)
@@ -313,7 +313,7 @@ class OperatingPoints:
         raise NotImplemented
 
     def do_nothing_key(self):
-        """ parameters to say we do noting, takes 0 time and has 0 performance"""
+        """ parameters to say we do nothing, takes 0 time and has 0 performance"""
         raise NotImplemented
 
     def is_pareto_optimal(self, perf_new, t_new):

contrib/rpc.py

@@ -110,7 +110,7 @@ class Server:
 
 
     def log(self, s):
-        self.logf.write("Sever log %s: %s\n" % (self.log_prefix, s))
+        self.logf.write("Server log %s: %s\n" % (self.log_prefix, s))
 
     def one_function(self):
         """
@@ -179,7 +179,7 @@ class Server:
             traceback.print_exc(50,sys.stderr)
             sys.exit(1)
 
-        LOG.info("exit sever")
+        LOG.info("exit server")
 
     def exec_loop_cleanup(self):
         pass

contrib/torch/quantization.py

@@ -26,7 +26,7 @@ class Quantizer:
 
     def train(self, x):
         """
-        takes a n-by-d array and peforms training
+        takes a n-by-d array and performs training
         """
         pass
 

tests/common_faiss_tests.py

@@ -11,7 +11,7 @@ import numpy as np
 import faiss
 
 # reduce number of threads to avoid excessive nb of threads in opt
-# mode (recuces runtime from 100s to 4s!)
+# mode (reduces runtime from 100s to 4s!)
 faiss.omp_set_num_threads(4)
 
 

tests/test_binary_hashindex.py

@@ -194,7 +194,7 @@ class TestKnn(unittest.TestCase):
     def test_result_order_binhash(self):
         self.subtest_result_order(0)
 
-    def test_result_order_miltihash(self):
+    def test_result_order_multihash(self):
         self.subtest_result_order(3)
 
 

tests/test_build_blocks.py

@@ -80,7 +80,7 @@ class TestMapLong2Long(unittest.TestCase):
         assert m.search(12343) == -1
 
 
-class TestOrthognalReconstruct(unittest.TestCase):
+class TestOrthogonalReconstruct(unittest.TestCase):
 
     def test_recons_orthonormal(self):
         lt = faiss.LinearTransform(20, 10, True)
@@ -103,7 +103,7 @@ class TestOrthognalReconstruct(unittest.TestCase):
 
         self.assertGreater(1e-5, err)
 
-    def test_recons_orthogona_impossible(self):
+    def test_recons_orthogonal_impossible(self):
         lt = faiss.LinearTransform(20, 10, True)
         rs = np.random.RandomState(10)
         A = rs.randn(10 * 20).astype('float32')

tests/test_common_ivf_empty_index.cpp

@@ -36,7 +36,7 @@ std::vector<float> get_random_vectors(size_t n, int seed) {
     return x;
 }
 
-/** InvetedLists implementation that dispatches the search to an InvertedList
+/** InvertedLists implementation that dispatches the search to an InvertedList
  * object that is passed in at query time */
 
 struct DispatchingInvertedLists : faiss::ReadOnlyInvertedLists {
@@ -142,7 +142,7 @@ TEST(COMMON, test_common_trained_index) {
         new_I[i] = I;
     }
 
-    // compare with reference reslt
+    // compare with reference result
     for (int i = 0; i < N; i++) {
         ASSERT_EQ(ref_I[i], new_I[i]);
     }

tests/test_distances_simd.cpp

@@ -234,7 +234,7 @@ TEST(TestFvecL1, manhattan_distance) {
         auto distance = faiss::fvec_L1(x.data(), y.data(), x.size());
 
         ASSERT_EQ(distance, true_distance)
-                << "Mismatching fvec_Linf results for nrows = " << nrows;
+                << "Mismatching fvec_L1 results for nrows = " << nrows;
     }
 }
 

tests/test_fast_scan_ivf.py

@@ -393,7 +393,7 @@ class TestIVFImplem12(unittest.TestCase):
     def test_by_residual_odd_dim(self):
         self.do_test(True, d=30)
 
-    # testin single query
+    # testing single query
     def test_no_residual_single_query(self):
         self.do_test(False, nq=1)
 
@@ -594,11 +594,11 @@ class TestReconstruct(unittest.TestCase):
     def test_by_residual(self):
         self.do_test(by_residual=True)
 
-    def do_test_generic(self, factory_string, 
-                        by_residual=False, metric=faiss.METRIC_L2): 
+    def do_test_generic(self, factory_string,
+                        by_residual=False, metric=faiss.METRIC_L2):
         d = 32
         ds = datasets.SyntheticDataset(d, 250, 200, 10)
-        index = faiss.index_factory(ds.d, factory_string, metric) 
+        index = faiss.index_factory(ds.d, factory_string, metric)
         if "IVF" in factory_string:
             index.by_residual = by_residual
             index.make_direct_map(True)
@@ -618,7 +618,7 @@ class TestReconstruct(unittest.TestCase):
         index2 = faiss.deserialize_index(faiss.serialize_index(index))
         codes2 = index2.sa_encode(ds.get_database()[120:130])
         np.testing.assert_array_equal(codes, codes2)
-        
+
 
     def test_ivfpq_residual(self):
         self.do_test_generic("IVF20,PQ16x4fs", by_residual=True)

tests/test_graph_based.py

@@ -52,13 +52,13 @@ class TestHNSW(unittest.TestCase):
         lims, D, I = index.range_search(self.xq, radius)
 
         nmiss = 0
-        # check if returned resutls are a subset of the reference results
+        # check if returned results are a subset of the reference results
         for i in range(len(self.xq)):
             ref = Iref[lims_ref[i]: lims_ref[i + 1]]
             new = I[lims[i]: lims[i + 1]]
             self.assertLessEqual(set(new), set(ref))
             nmiss += len(ref) - len(new)
-        # currenly we miss 405 / 6019 neighbors
+        # currently we miss 405 / 6019 neighbors
         self.assertLessEqual(nmiss, lims_ref[-1] * 0.1)
 
     def test_hnsw_unbounded_queue(self):

tests/test_ivf_index.cpp

@@ -192,7 +192,7 @@ TEST(IVF, list_context) {
         }
         std::vector<faiss::idx_t> coarse_idx(nb);
         index.quantizer->assign(nb, database.data(), coarse_idx.data());
-        // pass dummy ids, the acutal ids are assigned in TextContext object
+        // pass dummy ids, the actual ids are assigned in TextContext object
         std::vector<faiss::idx_t> xids(nb, 42);
         index.add_core(
                 nb, database.data(), xids.data(), coarse_idx.data(), &context);

tests/test_ivfpq_indexing.cpp

@@ -28,7 +28,7 @@ TEST(IVFPQ, accuracy) {
     // make the index object and train it
     faiss::IndexFlatL2 coarse_quantizer(d);
 
-    // a reasonable number of cetroids to index nb vectors
+    // a reasonable number of centroids to index nb vectors
     int ncentroids = 25;
 
     faiss::IndexIVFPQ index(&coarse_quantizer, d, ncentroids, 16, 8);

tests/test_local_search_quantizer.py

@@ -318,7 +318,7 @@ class TestComponents(unittest.TestCase):
             n,
             1)
 
-        # do icm encoding without pre-computed unary and bianry terms in Python
+        # do icm encoding without pre-computed unary and binary terms in Python
         codebooks = faiss.vector_float_to_array(lsq.codebooks)
         codebooks = codebooks.reshape(M, K, d).copy()
         ref_codes = icm_encode_ref(x, codebooks, codes)

tests/test_ondisk_ivf.cpp

@@ -156,7 +156,7 @@ TEST(ONDISK, test_add) {
     }
 }
 
-// WARN this thest will run multithreaded only in opt mode
+// WARN this test will run multithreaded only in opt mode
 TEST(ONDISK, make_invlists_threaded) {
     int nlist = 100;
     int code_size = 32;

tests/test_residual_quantizer.py

@@ -974,7 +974,7 @@ def precomp_codebooks(codebooks):
 
 
 ############################################################
-# Reference imelementation of table-based beam search (use_beam_LUT=1)
+# Reference implementation of table-based beam search (use_beam_LUT=1)
 ############################################################
 
 def beam_search_encode_step_tab(codes, L, distances, codebook_cross_prods_i,

tests/test_search_params.py

@@ -379,7 +379,7 @@ class TestSearchParams(unittest.TestCase):
 
         Dnew, Inew = index.search(ds.get_queries(), 10, params=params)
 
-        # make sure rhe parameter does indeed change the result...
+        # make sure the parameter does indeed change the result...
         self.assertFalse(np.all(Inew == I0))
 
         for param_name, value in ps_params.items():