add num_docs (#62)

* add num_docs

add num_docs

* Update test_inmemory_vectordb.py

* Update test_hnswlib_vectordb.py

tests/unit/test_hnswlib_vectordb.py

@@ -169,4 +169,11 @@ def test_hnswlib_vectordb_restore(docs_to_index, tmpdir):
         assert len(res.matches) == 10
         # assert res.id == res.matches[0].id
         # assert res.text == res.matches[0].text
-        # assert res.scores[0] < 0.001 # some precision issues, should be 0
+        # assert res.scores[0] < 0.001 # some precision issues, should be 0
+
+def test_hnswlib_num_dos(tmpdir):
+    db = HNSWVectorDB[MyDoc](workspace=str(tmpdir))
+    doc_list = [MyDoc(text=f'toy doc {i}', embedding=np.random.rand(128)) for i in range(1000)]
+    db.index(inputs=DocList[MyDoc](doc_list))
+    x=db.num_docs()
+    assert x['num_docs']==1000

tests/unit/test_inmemory_vectordb.py

@@ -172,3 +172,10 @@ def test_inmemory_vectordb_restore(docs_to_index, tmpdir):
         assert res.id == res.matches[0].id
         assert res.text == res.matches[0].text
         assert res.scores[0] > 0.99 # some precision issues, should be 1
+        
+def test_inmemory_num_dos(tmpdir):
+    db = InMemoryExactNNVectorDB[MyDoc](workspace=str(tmpdir))
+    doc_list = [MyDoc(text=f'toy doc {i}', embedding=np.random.rand(128)) for i in range(1000)]
+    db.index(inputs=DocList[MyDoc](doc_list))
+    x=db.num_docs()
+    assert x['num_docs']==1000

vectordb/db/base.py

@@ -227,6 +227,9 @@ class VectorDB(Generic[TSchema]):
             ret = asyncio.run(_deploy())
         return ret
 
+    def num_docs(self, **kwargs):
+        return self._executor.num_docs()
+    
     @pass_kwargs_as_params
     @unify_input_output
     def index(self, docs: 'DocList[TSchema]', parameters: Optional[Dict] = None, **kwargs):

vectordb/db/executors/hnsw_indexer.py

@@ -105,7 +105,7 @@ class HNSWLibIndexer(TypedExecutor):
         return self.update(docs, *args, **kwargs)
 
     def num_docs(self, **kwargs):
-        return {'num_docs': self._index.num_docs()}
+        return {'num_docs': self._indexer.num_docs()}
 
     def snapshot(self, snapshot_dir):
         # TODO: Maybe copy the work_dir to workspace if `handle` is False

vectordb/db/executors/inmemory_exact_indexer.py

@@ -71,7 +71,7 @@ class InMemoryExactNNIndexer(TypedExecutor):
         return self._index(docs)
 
     def num_docs(self, *args, **kwargs):
-        return {'num_docs': self._index.num_docs()}
+        return {'num_docs': self._indexer.num_docs()}
 
     def snapshot(self, snapshot_dir):
         snapshot_file = f'{snapshot_dir}/index.bin'