openvector_dev 0.1.8__py3-none-any.whl → 0.1.9__py3-none-any.whl

lein_vector/api/facade.py

@@ -1,13 +1,12 @@
 from collections.abc import Sequence
 from datetime import UTC, datetime
-from uuid import uuid4
+from uuid import UUID, uuid4
 
 import redis.asyncio as aioredis
 
 from lein_vector import MemoryManagerQdrant, QdrantAdapter, RedisShortTermMemory
 from lein_vector.schemas.chunk import Chunk, ChunkPayload
 from lein_vector.sentence_transformer import EmbeddingProviderGemini
-import warnings
 
 
 class Memory:
@@ -117,10 +116,12 @@ class Memory:
             await self.long.upsert_chunk_with_vector(new_chunk, vector)
             # self.short.clear_until(block_size)
 
-        if curr_no % 40 == 0:
-            await self.long.merge_old_chunks(user_id, "type0", n=self.merge_n)
+        if curr_no % (block_size * self.merge_n) == 0:
+            return await self.long.merge_old_chunks(user_id=user_id, bot=bot, chunk_type="type0", n=self.merge_n)
 
-    async def get_long_memories(self, user_id: int, bot: str, search_terms: list[str], topk: int = 3) -> list:
+    async def get_long_memories(
+        self, user_id: int, bot: str, search_terms: list[str], topk: int = 3
+    ) -> list:
         """
         Возвращает ТОЛЬКО длительную память по списку тем.
         :param bot: Кодовое имя бота
@@ -139,7 +140,9 @@ class Memory:
             filter_={"chunk_type": "type1"},
         )
 
-    async def get_short_memories(self, user_id: int, bot: str, n_memories: int = 20) -> list:
+    async def get_short_memories(
+        self, user_id: int, bot: str, n_memories: int = 20
+    ) -> list:
         """
         Возвращает ТОЛЬКО кратковременную память по списку тем
         :param bot: Кодовое имя бота
@@ -150,7 +153,9 @@ class Memory:
         data = await self.short.window(user_id, bot, n_memories)
         return self._to_openai(data)
 
-    async def add_short_msg(self, user_id: int, bot: str, text: str, *, role: str = "user"):
+    async def add_short_msg(
+        self, user_id: int, bot: str, text: str, *, role: str = "user"
+    ):
         """
         Добавляет в "короткую память" новое сообщение
         :param user_id: Идентификатор пользователя
@@ -162,7 +167,54 @@ class Memory:
         """
         await self.short.add(bot=bot, user_id=user_id, role=role, text=text)
 
-    async def delete_memory(self, user_id: int) -> None:
+    async def add_summary_chunk(
+        self,
+        user_id: int,
+        bot: str,
+        text: str,
+        old_chunks: list[UUID],
+        *,
+        chunk_type: str = "type1",
+    ) -> None:
+        """
+        Ручное добавление summary-чанка:
+        1) upsert нового чанка-саммари;
+        2) удаление использованных чанков.
+        :param user_id:     id пользователя
+        :param bot:         кодовое имя бота
+        :param text:        текст саммари
+        :param old_chunks:  список UUID чанков, вошедших в саммари
+        :param chunk_type:  тип нового чанка (по умолчанию 'type1')
+        """
+        # Эмбеддинг для саммари
+        embedding = await self.embed.get_embedding(text)
+
+        # Создаём новый Chunk
+        summary_chunk = Chunk(
+            chunk_id=uuid4(),
+            bot=bot,
+            user_id=user_id,
+            chunk_type=chunk_type,
+            created_at=datetime.now(UTC),
+            last_hit=datetime.now(UTC),
+            hit_count=0,
+            text=text,
+            persistent=False,
+            summary_of=old_chunks,
+        )
+
+        # Записываем в Qdrant
+        await self.long.upsert_chunk_with_vector(
+            user_id=user_id,
+            chunk=summary_chunk,
+            embedding=embedding,
+        )
+
+        # Удаляем исходные чанки
+        if old_chunks:
+            await self.long.delete_chunks(user_id, old_chunks)
+
+    async def delete_memory(self, user_id: int, bot: str) -> None:
         """
         Удаляет
         :param user_id: Идентификатор пользователя.

lein_vector/bases/memory_manager_abc.py

@@ -1,9 +1,13 @@
 from uuid import UUID
+
 from lein_vector.schemas.chunk import Chunk
 
+
 class MemoryManagerABC:
     async def upsert_chunk(self, user_id: int, bot: str, chunk: Chunk) -> None: ...
-    async def upsert_chunks(self, user_id: int, bot: str, chunks: list[Chunk]) -> None: ...
+    async def upsert_chunks(
+        self, user_id: int, bot: str, chunks: list[Chunk]
+    ) -> None: ...
     async def retrieve_by_embedding(
         self, user_id: int, bot: str, embedding: list[float], topk: int = 3
     ) -> list[Chunk]: ...
@@ -18,14 +22,20 @@ class MemoryManagerABC:
     async def retrieve_by_text(
         self, user_id: int, bot: str, query: str, topk: int = 3
     ) -> list[Chunk]: ...
-    async def merge_old_chunks(self, user_id: int, bot: str, chunk_type: str) -> None: ...
+    async def merge_old_chunks(
+        self, user_id: int, bot: str, chunk_type: str
+    ) -> None: ...
     async def archive_user(self, user_id: int, bot: str) -> None: ...
     async def restore_user(self, user_id: int, bot: str) -> None: ...
     async def increment_hit(self, user_id: int, bot: str, chunk_id: UUID) -> None: ...
     async def pop_first_n(
         self, user_id: int, bot: str, chunk_type: str, n: int
     ) -> list[Chunk]: ...
-    async def delete_oldest_nonpersistent(self, user_id: int, bot: str, keep: int) -> None: ...
+    async def delete_oldest_nonpersistent(
+        self, user_id: int, bot: str, keep: int
+    ) -> None: ...
     async def delete_chunk(self, user_id: int, bot: str, chunk_id: UUID) -> None: ...
-    async def delete_chunks(self, user_id: int, bot: str, chunk_ids: list[UUID]) -> None: ...
+    async def delete_chunks(
+        self, user_id: int, bot: str, chunk_ids: list[UUID]
+    ) -> None: ...
     async def delete_all(self, user_id: int, bot: str) -> None: ...

lein_vector/memory_manager_qdrant.py

@@ -1,6 +1,6 @@
 import asyncio
 from datetime import UTC, datetime
-from typing import Any
+from typing import Any, Tuple, List
 from uuid import UUID
 
 from lein_vector.bases.memory_manager_abc import MemoryManagerABC
@@ -77,13 +77,11 @@ class MemoryManagerQdrant(MemoryManagerABC):
             )
 
         async def _one(e):
-            return await self.qdrant.search(
-                e, q_filter, topk, score_threshold
-            )
+            return await self.qdrant.search(e, q_filter, topk, score_threshold)
 
         return await asyncio.gather(*[_one(e) for e in embeddings])
 
-    #ToDO: filter
+    # ToDO: filter
     async def retrieve_by_type(
         self, user_id: int, bot: str, chunk_type: str, topk: int = 3
     ) -> list[ChunkPayload]:
@@ -91,30 +89,34 @@ class MemoryManagerQdrant(MemoryManagerABC):
         filter_ = {"user_id": user_id, "bot": bot, "chunk_type": chunk_type}
         return await self.qdrant.get_all_chunks_with_filter(filter_)
 
-    async def merge_old_chunks(self, user_id: int, bot: str, chunk_type: str, n: int = 5) -> None:
+    async def merge_old_chunks(
+            self,
+            user_id: int,
+            bot: str,
+            chunk_type: str,
+            n: int = 5,
+    ) -> Tuple[str | None, List[UUID]]:
+        """
+        Собрать n старых чанков и вернуть:
+          • merged_text — склеенные сообщения c ролями,
+          • used_ids    — UUID этих чанков.
+
+        Если чанков меньше n — вернётся (None, []).
+        """
         chunks = await self.qdrant.get_n_oldest_chunks(user_id, bot, chunk_type, n)
         if len(chunks) < n:
-            return
-        merged_text = " | ".join([c.text for c in chunks])
-        from datetime import datetime
-        from uuid import uuid4
-
-        # ToDo: LLM summary
-        summary_chunk = Chunk(
-            chunk_id=uuid4(),
-            user_id=user_id,
-            bot=bot,
-            chunk_type=self._next_type(chunk_type),
-            created_at=datetime.now(UTC),
-            last_hit=datetime.now(UTC),
-            hit_count=0,
-            text=merged_text,
-            persistent=False,
-            summary_of=[c.chunk_id for c in chunks],
-        )
-        await self.upsert_chunk(user_id, bot, summary_chunk)
+            return None, []
+
+        def _ensure_role(txt: str, default_role: str = "gf") -> str:
+            # если строка уже начинается с 'role: ', оставляем как есть
+            if txt.split(":", 1)[0] in {"user", "gf", "assistant"}:
+                return txt
+            return f"{default_role}: {txt}"
 
-        await self.delete_chunks(user_id, bot, [c.chunk_id for c in chunks])
+        merged_text = "\n".join(_ensure_role(c.text) for c in chunks)
+        used_ids = [c.chunk_id for c in chunks]
+
+        return merged_text, used_ids
 
     async def archive_user(self, user_id: int, bot: str) -> None:
         all_chunks = await self.qdrant.get_all_chunks(user_id, bot)
@@ -134,7 +136,9 @@ class MemoryManagerQdrant(MemoryManagerABC):
     async def delete_chunk(self, user_id: int, bot: str, chunk_id: UUID) -> None:
         await self.qdrant.delete(chunk_id)
 
-    async def delete_chunks(self, user_id: int, bot: str, chunk_ids: list[UUID]) -> None:
+    async def delete_chunks(
+        self, user_id: int, bot: str, chunk_ids: list[UUID]
+    ) -> None:
         await self.qdrant.delete_batch(chunk_ids)
 
     async def delete_all(self, user_id: int, bot: str) -> None:

lein_vector/schemas/chunk.py

@@ -36,7 +36,7 @@ class ChunkPayload(BaseModel):
     source_chunk_id: UUID | None = None
     extra: dict | None = Field(default_factory=dict)
 
-    def to_chunk(self, last_hit: datetime = None, hit_count: int = 0) -> Chunk:
+    def to_chunk(self, last_hit: datetime | None = None, hit_count: int = 0) -> Chunk:
         return Chunk(
             **self.model_dump(),
             last_hit=last_hit or datetime.now(UTC),

lein_vector/sentence_transformer.py

@@ -22,6 +22,7 @@ class EmbeddingProviderGemini(EmbeddingProviderABC):
     async def get_embedding(self, text: str) -> list[float]:
         # В Gemini SDK обычно нет async, значит — обёртка через run_in_executor:
         import asyncio
+
         if not isinstance(text, str):
             return
         loop = asyncio.get_running_loop()

{openvector_dev-0.1.8.dist-info → openvector_dev-0.1.9.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: openvector_dev
-Version: 0.1.8
+Version: 0.1.9
 Summary: 
 Author: p00ler
 Author-email: liveitspain@gmail.com

{openvector_dev-0.1.8.dist-info → openvector_dev-0.1.9.dist-info}/RECORD

@@ -1,17 +1,17 @@
 lein_vector/__init__.py,sha256=ieXy65X4eVzUtSHNvlh00s4aMwKReafNni4s2q68z3Q,496
 lein_vector/api/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-lein_vector/api/facade.py,sha256=URxGEad8QrhYY6QJTrz552mk_9z8TWi7RB1veykyD-4,7886
+lein_vector/api/facade.py,sha256=OC94eCQnkTgq0BycEEaEMX5YUs5EUtACpiLspYcqusk,9651
 lein_vector/bases/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 lein_vector/bases/embeding_provider_abc.py,sha256=WBpVC6ra-SYqhJeZs8R8U679wfGebXgwOOVPIii-IvY,265
-lein_vector/bases/memory_manager_abc.py,sha256=pWuIw_wYlvUlt40jbiE92lWzdYaowZKg3XK0Mq0tQrw,1687
-lein_vector/memory_manager_qdrant.py,sha256=DFhvtpDl3OxLSpCQiNXnKVSoqQ3475cSDcfE5wF8Njw,6175
+lein_vector/bases/memory_manager_abc.py,sha256=NE13Lz-3aX3RjkRhsd5FsezPNk0xju34fFpO3Fa1n9A,1755
+lein_vector/memory_manager_qdrant.py,sha256=jzb04gnESVRH62QSYhnwsl-HpS6Kci721rPA3_6LSDQ,6382
 lein_vector/memory_manager_ram.py,sha256=lUATu6U-cZuRHnR6U1L8aJQB8FNK0Wi40vXsZ5fjDgI,6002
 lein_vector/qdrant_adapter.py,sha256=uPhYAp0KmYQhY3DF29JMO1JTbM__xDWGlnZZPor3VKI,4193
 lein_vector/redis_short_term.py,sha256=uCcncjgXAlNAY46b3_Pjvrn_G0NEI16kRRR9hxayzak,3155
 lein_vector/schemas/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-lein_vector/schemas/chunk.py,sha256=FSUwaFUSZk2ptoxupsdIfj5HYAMnHnu3eBfNpXCUMpM,1195
-lein_vector/sentence_transformer.py,sha256=gbWtkT01QFFA0IR_Ta-xsfBfb6owDzGnZDKY1g9fvBQ,1851
+lein_vector/schemas/chunk.py,sha256=qq2J7UM94AhKXKg_gnErac5GQF9DCoaEeEHyxZg_xuk,1202
+lein_vector/sentence_transformer.py,sha256=SjjIa43Y7JxcU9VkpT2Ml-CVXb8tqmrbMutZN6O_I6s,1853
 lein_vector/short_term.py,sha256=KFot9r26d9TZYLscVT42V0NwufIQUkx-hftOe2p7qLU,1933
-openvector_dev-0.1.8.dist-info/METADATA,sha256=GOE9yKHgwxWZ5jmu8sOpqnz1XBcmKh0dSuP3zoGWOwQ,3167
-openvector_dev-0.1.8.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
-openvector_dev-0.1.8.dist-info/RECORD,,
+openvector_dev-0.1.9.dist-info/METADATA,sha256=ZnabZl7TnXLZXj7xFqQk5awRGPzFpuyWnDjMyUAx_Z0,3167
+openvector_dev-0.1.9.dist-info/WHEEL,sha256=XbeZDeTWKc1w7CSIyre5aMDU_-PohRwTQceYnisIYYY,88
+openvector_dev-0.1.9.dist-info/RECORD,,