vectordb-bench 0.0.12__py3-none-any.whl → 0.0.13__py3-none-any.whl

vectordb_bench/backend/clients/pgvectorscale/pgvectorscale.py

@@ -0,0 +1,272 @@
+"""Wrapper around the Pgvectorscale vector database over VectorDB"""
+
+import logging
+import pprint
+from contextlib import contextmanager
+from typing import Any, Generator, Optional, Tuple
+
+import numpy as np
+import psycopg
+from pgvector.psycopg import register_vector
+from psycopg import Connection, Cursor, sql
+
+from ..api import VectorDB
+from .config import PgVectorScaleConfigDict, PgVectorScaleIndexConfig
+
+log = logging.getLogger(__name__)
+
+
+class PgVectorScale(VectorDB):
+    """Use psycopg instructions"""
+
+    conn: psycopg.Connection[Any] | None = None
+    coursor: psycopg.Cursor[Any] | None = None
+
+    def __init__(
+        self,
+        dim: int,
+        db_config: PgVectorScaleConfigDict,
+        db_case_config: PgVectorScaleIndexConfig,
+        collection_name: str = "pg_vectorscale_collection",
+        drop_old: bool = False,
+        **kwargs,
+    ):
+        self.name = "PgVectorScale"
+        self.db_config = db_config
+        self.case_config = db_case_config
+        self.table_name = collection_name
+        self.dim = dim
+
+        self._index_name = "pgvectorscale_index"
+        self._primary_field = "id"
+        self._vector_field = "embedding"
+
+        self.conn, self.cursor = self._create_connection(**self.db_config)        
+
+        log.info(f"{self.name} config values: {self.db_config}\n{self.case_config}")
+        if not any(
+            (
+                self.case_config.create_index_before_load,
+                self.case_config.create_index_after_load,
+            )
+        ):
+            err = f"{self.name} config must create an index using create_index_before_load or create_index_after_load"
+            log.error(err)
+            raise RuntimeError(
+                f"{err}\n{pprint.pformat(self.db_config)}\n{pprint.pformat(self.case_config)}"
+            )
+
+        if drop_old:
+            self._drop_index()
+            self._drop_table()
+            self._create_table(dim)
+            if self.case_config.create_index_before_load:
+                self._create_index()
+
+        self.cursor.close()
+        self.conn.close()
+        self.cursor = None
+        self.conn = None
+
+    @staticmethod
+    def _create_connection(**kwargs) -> Tuple[Connection, Cursor]:
+        conn = psycopg.connect(**kwargs)
+        conn.cursor().execute("CREATE EXTENSION IF NOT EXISTS vectorscale CASCADE")
+        conn.commit()
+        register_vector(conn)
+        conn.autocommit = False
+        cursor = conn.cursor()
+
+        assert conn is not None, "Connection is not initialized"
+        assert cursor is not None, "Cursor is not initialized"
+
+        return conn, cursor
+
+    @contextmanager
+    def init(self) -> Generator[None, None, None]:
+        self.conn, self.cursor = self._create_connection(**self.db_config)
+
+        # index configuration may have commands defined that we should set during each client session
+        session_options: dict[str, Any] = self.case_config.session_param()
+
+        if len(session_options) > 0:
+            for setting_name, setting_val in session_options.items():
+                command = sql.SQL("SET {setting_name} " + "= {setting_val};").format(
+                    setting_name=sql.Identifier(setting_name),
+                    setting_val=sql.Identifier(str(setting_val)),
+                )
+                log.debug(command.as_string(self.cursor))
+                self.cursor.execute(command)
+            self.conn.commit()
+        
+        self._unfiltered_search = sql.Composed(
+            [
+                sql.SQL("SELECT id FROM public.{} ORDER BY embedding ").format(
+                    sql.Identifier(self.table_name)
+                ),
+                sql.SQL(self.case_config.search_param()["metric_fun_op"]),
+                sql.SQL(" %s::vector LIMIT %s::int"),
+            ]
+        )
+
+        try:
+            yield
+        finally:
+            self.cursor.close()
+            self.conn.close()
+            self.cursor = None
+            self.conn = None
+
+    def _drop_table(self):
+        assert self.conn is not None, "Connection is not initialized"
+        assert self.cursor is not None, "Cursor is not initialized"
+        log.info(f"{self.name} client drop table : {self.table_name}")
+
+        self.cursor.execute(
+            sql.SQL("DROP TABLE IF EXISTS public.{table_name}").format(
+                table_name=sql.Identifier(self.table_name)
+            )
+        )
+        self.conn.commit()
+
+    def ready_to_load(self):
+        pass
+
+    def optimize(self):
+        self._post_insert()
+
+    def _post_insert(self):
+        log.info(f"{self.name} post insert before optimize")
+        if self.case_config.create_index_after_load:
+            self._drop_index()
+            self._create_index()
+
+    def _drop_index(self):
+        assert self.conn is not None, "Connection is not initialized"
+        assert self.cursor is not None, "Cursor is not initialized"
+        log.info(f"{self.name} client drop index : {self._index_name}")
+
+        drop_index_sql = sql.SQL("DROP INDEX IF EXISTS {index_name}").format(
+            index_name=sql.Identifier(self._index_name)
+        )
+        log.debug(drop_index_sql.as_string(self.cursor))
+        self.cursor.execute(drop_index_sql)
+        self.conn.commit()
+
+    def _create_index(self):
+        assert self.conn is not None, "Connection is not initialized"
+        assert self.cursor is not None, "Cursor is not initialized"
+        log.info(f"{self.name} client create index : {self._index_name}")
+
+        index_param: dict[str, Any] = self.case_config.index_param()
+
+        options = []
+        for option_name, option_val in index_param["options"].items():
+            if option_val is not None:
+                options.append(
+                    sql.SQL("{option_name} = {val}").format(
+                        option_name=sql.Identifier(option_name),
+                        val=sql.Identifier(str(option_val)),
+                    )
+                )
+        
+        num_bits_per_dimension = "2" if self.dim < 900 else "1"
+        options.append(
+            sql.SQL("{option_name} = {val}").format(
+                option_name=sql.Identifier("num_bits_per_dimension"),
+                val=sql.Identifier(num_bits_per_dimension),
+            )
+        )
+
+        if any(options):
+            with_clause = sql.SQL("WITH ({});").format(sql.SQL(", ").join(options))
+        else:
+            with_clause = sql.Composed(())
+
+        index_create_sql = sql.SQL(
+            """
+            CREATE INDEX IF NOT EXISTS {index_name} ON public.{table_name} 
+            USING {index_type} (embedding {embedding_metric})
+            """
+        ).format(
+            index_name=sql.Identifier(self._index_name),
+            table_name=sql.Identifier(self.table_name),
+            index_type=sql.Identifier(index_param["index_type"].lower()),
+            embedding_metric=sql.Identifier(index_param["metric"]),
+        )
+        index_create_sql_with_with_clause = (
+            index_create_sql + with_clause
+        ).join(" ")
+        log.debug(index_create_sql_with_with_clause.as_string(self.cursor))
+        self.cursor.execute(index_create_sql_with_with_clause)
+        self.conn.commit()
+
+    def _create_table(self, dim: int):
+        assert self.conn is not None, "Connection is not initialized"
+        assert self.cursor is not None, "Cursor is not initialized"
+
+        try:
+            log.info(f"{self.name} client create table : {self.table_name}")
+
+            self.cursor.execute(
+                sql.SQL(
+                    "CREATE TABLE IF NOT EXISTS public.{table_name} (id BIGINT PRIMARY KEY, embedding vector({dim}));"
+                ).format(table_name=sql.Identifier(self.table_name), dim=dim)
+            )
+            self.conn.commit()
+        except Exception as e:
+            log.warning(
+                f"Failed to create pgvectorscale table: {self.table_name} error: {e}"
+            )
+            raise e from None
+
+    def insert_embeddings(
+        self,
+        embeddings: list[list[float]],
+        metadata: list[int],
+        **kwargs: Any,
+    ) -> Tuple[int, Optional[Exception]]:
+        assert self.conn is not None, "Connection is not initialized"
+        assert self.cursor is not None, "Cursor is not initialized"
+
+        try:
+            metadata_arr = np.array(metadata)
+            embeddings_arr = np.array(embeddings)
+
+            with self.cursor.copy(
+                sql.SQL("COPY public.{table_name} FROM STDIN (FORMAT BINARY)").format(
+                    table_name=sql.Identifier(self.table_name)
+                )
+            ) as copy:
+                copy.set_types(["bigint", "vector"])
+                for i, row in enumerate(metadata_arr):
+                    copy.write_row((row, embeddings_arr[i]))
+            self.conn.commit()
+
+            if kwargs.get("last_batch"):
+                self._post_insert()
+
+            return len(metadata), None
+        except Exception as e:
+            log.warning(
+                f"Failed to insert data into pgvector table ({self.table_name}), error: {e}"
+            )
+            return 0, e
+
+    def search_embedding(
+        self,
+        query: list[float],
+        k: int = 100,
+        filters: dict | None = None,
+        timeout: int | None = None,
+    ) -> list[int]:
+        assert self.conn is not None, "Connection is not initialized"
+        assert self.cursor is not None, "Cursor is not initialized"
+
+        q = np.asarray(query)
+        # TODO add filters support
+        result = self.cursor.execute(
+            self._unfiltered_search, (q, k), prepare=True, binary=True
+        )
+
+        return [int(i[0]) for i in result.fetchall()]

vectordb_bench/cli/vectordbbench.py

@@ -1,5 +1,7 @@
 from ..backend.clients.pgvector.cli import PgVectorHNSW
+from ..backend.clients.pgvecto_rs.cli import PgVectoRSHNSW, PgVectoRSIVFFlat
 from ..backend.clients.redis.cli import Redis
+from ..backend.clients.memorydb.cli import MemoryDB
 from ..backend.clients.test.cli import Test
 from ..backend.clients.weaviate_cloud.cli import Weaviate
 from ..backend.clients.zilliz_cloud.cli import ZillizAutoIndex
@@ -10,7 +12,10 @@ from ..backend.clients.aws_opensearch.cli import AWSOpenSearch
 from .cli import cli
 
 cli.add_command(PgVectorHNSW)
+cli.add_command(PgVectoRSHNSW)
+cli.add_command(PgVectoRSIVFFlat)
 cli.add_command(Redis)
+cli.add_command(MemoryDB)
 cli.add_command(Weaviate)
 cli.add_command(Test)
 cli.add_command(ZillizAutoIndex)

vectordb_bench/frontend/components/check_results/data.py

@@ -24,7 +24,10 @@ def getFilterTasks(
         task
         for task in tasks
         if task.task_config.db_name in dbNames
-        and task.task_config.case_config.case_id.case_cls(task.task_config.case_config.custom_case).name in caseNames
+        and task.task_config.case_config.case_id.case_cls(
+            task.task_config.case_config.custom_case
+        ).name
+        in caseNames
     ]
     return filterTasks
 
@@ -35,17 +38,20 @@ def mergeTasks(tasks: list[CaseResult]):
         db_name = task.task_config.db_name
         db = task.task_config.db.value
         db_label = task.task_config.db_config.db_label or ""
-        case = task.task_config.case_config.case_id.case_cls(task.task_config.case_config.custom_case)
+        version = task.task_config.db_config.version or ""
+        case = task.task_config.case_config.case_id.case_cls(
+            task.task_config.case_config.custom_case
+        )
         dbCaseMetricsMap[db_name][case.name] = {
             "db": db,
             "db_label": db_label,
+            "version": version,
             "metrics": mergeMetrics(
                 dbCaseMetricsMap[db_name][case.name].get("metrics", {}),
                 asdict(task.metrics),
             ),
             "label": getBetterLabel(
-                dbCaseMetricsMap[db_name][case.name].get(
-                    "label", ResultLabel.FAILED),
+                dbCaseMetricsMap[db_name][case.name].get("label", ResultLabel.FAILED),
                 task.label,
             ),
         }
@@ -57,6 +63,7 @@ def mergeTasks(tasks: list[CaseResult]):
             metrics = metricInfo["metrics"]
             db = metricInfo["db"]
             db_label = metricInfo["db_label"]
+            version = metricInfo["version"]
             label = metricInfo["label"]
             if label == ResultLabel.NORMAL:
                 mergedTasks.append(
@@ -64,6 +71,7 @@ def mergeTasks(tasks: list[CaseResult]):
                         "db_name": db_name,
                         "db": db,
                         "db_label": db_label,
+                        "version": version,
                         "case_name": case_name,
                         "metricsSet": set(metrics.keys()),
                         **metrics,
@@ -79,8 +87,7 @@ def mergeMetrics(metrics_1: dict, metrics_2: dict) -> dict:
     metrics = {**metrics_1}
     for key, value in metrics_2.items():
         metrics[key] = (
-            getBetterMetric(
-                key, value, metrics[key]) if key in metrics else value
+            getBetterMetric(key, value, metrics[key]) if key in metrics else value
         )
 
     return metrics

vectordb_bench/frontend/components/run_test/caseSelector.py

@@ -100,6 +100,16 @@ def caseConfigSetting(st, dbToCaseClusterConfigs, uiCaseItem: UICaseItem, active
                         value=config.inputConfig["value"],
                         help=config.inputHelp,
                     )
+                elif config.inputType == InputType.Float:
+                    caseConfig[config.label] = column.number_input(
+                        config.displayLabel if config.displayLabel else config.label.value,
+                        step=config.inputConfig.get("step", 0.1),
+                        min_value=config.inputConfig["min"],
+                        max_value=config.inputConfig["max"],
+                        key=key,
+                        value=config.inputConfig["value"],
+                        help=config.inputHelp,
+                    )
                 k += 1
         if k == 0:
             columns[1].write("Auto")

vectordb_bench/frontend/components/run_test/dbConfigSetting.py

@@ -1,9 +1,10 @@
 from pydantic import ValidationError
-from vectordb_bench.frontend.config.styles import *
+from vectordb_bench.backend.clients import DB
+from vectordb_bench.frontend.config.styles import DB_CONFIG_SETTING_COLUMNS
 from vectordb_bench.frontend.utils import inputIsPassword
 
 
-def dbConfigSettings(st, activedDbList):
+def dbConfigSettings(st, activedDbList: list[DB]):
     expander = st.expander("Configurations for the selected databases", True)
 
     dbConfigs = {}
@@ -27,7 +28,7 @@ def dbConfigSettings(st, activedDbList):
     return dbConfigs, isAllValid
 
 
-def dbConfigSettingItem(st, activeDb):
+def dbConfigSettingItem(st, activeDb: DB):
     st.markdown(
         f"<div style='font-weight: 600; font-size: 20px; margin-top: 16px;'>{activeDb.value}</div>",
         unsafe_allow_html=True,
@@ -36,20 +37,41 @@ def dbConfigSettingItem(st, activeDb):
 
     dbConfigClass = activeDb.config_cls
     properties = dbConfigClass.schema().get("properties")
-    propertiesItems = list(properties.items())
-    moveDBLabelToLast(propertiesItems)
     dbConfig = {}
-    for j, property in enumerate(propertiesItems):
-        column = columns[j % DB_CONFIG_SETTING_COLUMNS]
-        key, value = property
+    idx = 0
+
+    # db config (unique)
+    for key, property in properties.items():
+        if (
+            key not in dbConfigClass.common_short_configs()
+            and key not in dbConfigClass.common_long_configs()
+        ):
+            column = columns[idx % DB_CONFIG_SETTING_COLUMNS]
+            idx += 1
+            dbConfig[key] = column.text_input(
+                key,
+                key="%s-%s" % (activeDb.name, key),
+                value=property.get("default", ""),
+                type="password" if inputIsPassword(key) else "default",
+            )
+    # db config (common short labels)
+    for key in dbConfigClass.common_short_configs():
+        column = columns[idx % DB_CONFIG_SETTING_COLUMNS]
+        idx += 1
         dbConfig[key] = column.text_input(
             key,
-            key="%s-%s" % (activeDb, key),
-            value=value.get("default", ""),
-            type="password" if inputIsPassword(key) else "default",
+            key="%s-%s" % (activeDb.name, key),
+            value="",
+            type="default",
+            placeholder="optional, for labeling results",
         )
-    return dbConfig
-
 
-def moveDBLabelToLast(propertiesItems):
-    propertiesItems.sort(key=lambda x: 1 if x[0] == "db_label" else 0)
+    # db config (common long text_input)
+    for key in dbConfigClass.common_long_configs():
+        dbConfig[key] = st.text_area(
+            key,
+            key="%s-%s" % (activeDb.name, key),
+            value="",
+            placeholder="optional",
+        )
+    return dbConfig

vectordb_bench/frontend/components/run_test/initStyle.py

@@ -9,6 +9,8 @@ def initStyle(st):
             div[data-testid='stHorizontalBlock'] {gap: 8px;}
             /* check box */
             .stCheckbox p { color: #000; font-size: 18px; font-weight: 600; }
+            /* db selector - db_name should not wrap */
+            div[data-testid="stVerticalBlockBorderWrapper"] div[data-testid="stCheckbox"] div[data-testid="stWidgetLabel"] p { white-space: nowrap; }
         </style>""",
         unsafe_allow_html=True,
-    )
+    )