From 78a4c9ed20e2b08e01d39487bec356da7c191ab8 Mon Sep 17 00:00:00 2001
From: Logan <logan.markewich@live.com>
Date: Mon, 26 Feb 2024 13:12:12 -0600
Subject: [PATCH] fix prompt helper init (#11379)
---
.../core/indices/common/struct_store/base.py | 6 +++---
.../llama_index/core/indices/common_tree/base.py | 7 +++----
.../llama_index/core/indices/prompt_helper.py | 1 +
.../llama_index/core/indices/tree/inserter.py | 6 +++---
.../core/indices/tree/select_leaf_retriever.py | 6 +++---
.../core/response_synthesizers/base.py | 10 +++++++---
.../core/response_synthesizers/factory.py | 15 +++++++++++----
.../core/response_synthesizers/generation.py | 3 +++
.../core/response_synthesizers/refine.py | 3 +++
.../response_synthesizers/simple_summarize.py | 3 +++
.../core/response_synthesizers/tree_summarize.py | 3 +++
llama-index-core/llama_index/core/settings.py | 12 ------------
12 files changed, 43 insertions(+), 32 deletions(-)
diff --git a/llama-index-core/llama_index/core/indices/common/struct_store/base.py b/llama-index-core/llama_index/core/indices/common/struct_store/base.py
index 3333bc0667..d4c4753a99 100644
--- a/llama-index-core/llama_index/core/indices/common/struct_store/base.py
+++ b/llama-index-core/llama_index/core/indices/common/struct_store/base.py
@@ -6,6 +6,7 @@ from typing import Any, Callable, Dict, List, Optional, Sequence, cast
from llama_index.core.callbacks.schema import CBEventType, EventPayload
from llama_index.core.data_structs.table import StructDatapoint
+from llama_index.core.indices.prompt_helper import PromptHelper
from llama_index.core.node_parser.interface import TextSplitter
from llama_index.core.prompts import BasePromptTemplate
from llama_index.core.prompts.default_prompt_selectors import (
@@ -26,7 +27,6 @@ from llama_index.core.settings import (
Settings,
callback_manager_from_settings_or_context,
llm_from_settings_or_context,
- prompt_helper_from_settings_or_context,
)
from llama_index.core.utilities.sql_wrapper import SQLDatabase
from llama_index.core.utils import truncate_text
@@ -67,8 +67,8 @@ class SQLDocumentContextBuilder:
self._sql_database = sql_database
self._text_splitter = text_splitter
self._llm = llm or llm_from_settings_or_context(Settings, service_context)
- self._prompt_helper = prompt_helper_from_settings_or_context(
- Settings, service_context
+ self._prompt_helper = Settings._prompt_helper or PromptHelper.from_llm_metadata(
+ self._llm.metadata,
)
self._callback_manager = callback_manager_from_settings_or_context(
Settings, service_context
diff --git a/llama-index-core/llama_index/core/indices/common_tree/base.py b/llama-index-core/llama_index/core/indices/common_tree/base.py
index 18213f1c7a..ba10592256 100644
--- a/llama-index-core/llama_index/core/indices/common_tree/base.py
+++ b/llama-index-core/llama_index/core/indices/common_tree/base.py
@@ -1,6 +1,5 @@
"""Common classes/functions for tree index operations."""
-
import asyncio
import logging
from typing import Dict, List, Optional, Sequence, Tuple
@@ -8,6 +7,7 @@ from typing import Dict, List, Optional, Sequence, Tuple
from llama_index.core.async_utils import run_async_tasks
from llama_index.core.callbacks.schema import CBEventType, EventPayload
from llama_index.core.data_structs.data_structs import IndexGraph
+from llama_index.core.indices.prompt_helper import PromptHelper
from llama_index.core.indices.utils import get_sorted_node_list, truncate_text
from llama_index.core.llms.llm import LLM
from llama_index.core.prompts import BasePromptTemplate
@@ -17,7 +17,6 @@ from llama_index.core.settings import (
Settings,
callback_manager_from_settings_or_context,
llm_from_settings_or_context,
- prompt_helper_from_settings_or_context,
)
from llama_index.core.storage.docstore import BaseDocumentStore
from llama_index.core.storage.docstore.registry import get_default_docstore
@@ -50,8 +49,8 @@ class GPTTreeIndexBuilder:
self.num_children = num_children
self.summary_prompt = summary_prompt
self._llm = llm or llm_from_settings_or_context(Settings, service_context)
- self._prompt_helper = prompt_helper_from_settings_or_context(
- Settings, service_context
+ self._prompt_helper = Settings._prompt_helper or PromptHelper.from_llm_metadata(
+ self._llm.metadata,
)
self._callback_manager = callback_manager_from_settings_or_context(
Settings, service_context
diff --git a/llama-index-core/llama_index/core/indices/prompt_helper.py b/llama-index-core/llama_index/core/indices/prompt_helper.py
index 359d194589..6b7f141ed6 100644
--- a/llama-index-core/llama_index/core/indices/prompt_helper.py
+++ b/llama-index-core/llama_index/core/indices/prompt_helper.py
@@ -115,6 +115,7 @@ class PromptHelper(BaseComponent):
"""
context_window = llm_metadata.context_window
+
if llm_metadata.num_output == -1:
num_output = DEFAULT_NUM_OUTPUTS
else:
diff --git a/llama-index-core/llama_index/core/indices/tree/inserter.py b/llama-index-core/llama_index/core/indices/tree/inserter.py
index 58f1d11607..0bb0b54cd4 100644
--- a/llama-index-core/llama_index/core/indices/tree/inserter.py
+++ b/llama-index-core/llama_index/core/indices/tree/inserter.py
@@ -3,6 +3,7 @@
from typing import Optional, Sequence
from llama_index.core.data_structs.data_structs import IndexGraph
+from llama_index.core.indices.prompt_helper import PromptHelper
from llama_index.core.indices.tree.utils import get_numbered_text_from_nodes
from llama_index.core.indices.utils import (
extract_numbers_given_response,
@@ -19,7 +20,6 @@ from llama_index.core.service_context import ServiceContext
from llama_index.core.settings import (
Settings,
llm_from_settings_or_context,
- prompt_helper_from_settings_or_context,
)
from llama_index.core.storage.docstore import BaseDocumentStore
from llama_index.core.storage.docstore.registry import get_default_docstore
@@ -46,8 +46,8 @@ class TreeIndexInserter:
self.insert_prompt = insert_prompt
self.index_graph = index_graph
self._llm = llm or llm_from_settings_or_context(Settings, service_context)
- self._prompt_helper = prompt_helper_from_settings_or_context(
- Settings, service_context
+ self._prompt_helper = Settings._prompt_helper or PromptHelper.from_llm_metadata(
+ self._llm.metadata,
)
self._docstore = docstore or get_default_docstore()
diff --git a/llama-index-core/llama_index/core/indices/tree/select_leaf_retriever.py b/llama-index-core/llama_index/core/indices/tree/select_leaf_retriever.py
index 9483de176c..b8dc73d14e 100644
--- a/llama-index-core/llama_index/core/indices/tree/select_leaf_retriever.py
+++ b/llama-index-core/llama_index/core/indices/tree/select_leaf_retriever.py
@@ -6,6 +6,7 @@ from typing import Any, Dict, List, Optional, cast
from llama_index.core.base.base_retriever import BaseRetriever
from llama_index.core.base.response.schema import Response
from llama_index.core.callbacks.base import CallbackManager
+from llama_index.core.indices.prompt_helper import PromptHelper
from llama_index.core.indices.query.schema import QueryBundle
from llama_index.core.indices.tree.base import TreeIndex
from llama_index.core.indices.tree.utils import get_numbered_text_from_nodes
@@ -32,7 +33,6 @@ from llama_index.core.schema import (
from llama_index.core.settings import (
Settings,
callback_manager_from_settings_or_context,
- prompt_helper_from_settings_or_context,
)
from llama_index.core.utils import print_text, truncate_text
@@ -93,8 +93,8 @@ class TreeSelectLeafRetriever(BaseRetriever):
self._index_struct = index.index_struct
self._docstore = index.docstore
self._service_context = index.service_context
- self._prompt_helper = prompt_helper_from_settings_or_context(
- Settings, index.service_context
+ self._prompt_helper = Settings._prompt_helper or PromptHelper.from_llm_metadata(
+ self._llm.metadata,
)
self._text_qa_template = text_qa_template or DEFAULT_TEXT_QA_PROMPT
diff --git a/llama-index-core/llama_index/core/response_synthesizers/base.py b/llama-index-core/llama_index/core/response_synthesizers/base.py
index 6edc364040..acecef8f11 100644
--- a/llama-index-core/llama_index/core/response_synthesizers/base.py
+++ b/llama-index-core/llama_index/core/response_synthesizers/base.py
@@ -7,6 +7,7 @@ Will support different modes, from 1) stuffing chunks into prompt,
2) create and refine separately over each chunk, 3) tree summarization.
"""
+
import logging
from abc import abstractmethod
from typing import Any, Dict, Generator, List, Optional, Sequence, Union
@@ -41,7 +42,6 @@ from llama_index.core.settings import (
Settings,
callback_manager_from_settings_or_context,
llm_from_settings_or_context,
- prompt_helper_from_settings_or_context,
)
from llama_index.core.types import RESPONSE_TEXT_TYPE
@@ -69,8 +69,12 @@ class BaseSynthesizer(ChainableMixin, PromptMixin):
callback_manager
or callback_manager_from_settings_or_context(Settings, service_context)
)
- self._prompt_helper = prompt_helper or prompt_helper_from_settings_or_context(
- Settings, service_context
+ self._prompt_helper = (
+ prompt_helper
+ or Settings._prompt_helper
+ or PromptHelper.from_llm_metadata(
+ self._llm.metadata,
+ )
)
self._streaming = streaming
diff --git a/llama-index-core/llama_index/core/response_synthesizers/factory.py b/llama-index-core/llama_index/core/response_synthesizers/factory.py
index f8543667c3..aa191e0b01 100644
--- a/llama-index-core/llama_index/core/response_synthesizers/factory.py
+++ b/llama-index-core/llama_index/core/response_synthesizers/factory.py
@@ -31,7 +31,6 @@ from llama_index.core.settings import (
Settings,
callback_manager_from_settings_or_context,
llm_from_settings_or_context,
- prompt_helper_from_settings_or_context,
)
from llama_index.core.types import BasePydanticProgram
@@ -63,9 +62,17 @@ def get_response_synthesizer(
Settings, service_context
)
llm = llm or llm_from_settings_or_context(Settings, service_context)
- prompt_helper = prompt_helper or prompt_helper_from_settings_or_context(
- Settings, service_context
- )
+
+ if service_context is not None:
+ prompt_helper = service_context.prompt_helper
+ else:
+ prompt_helper = (
+ prompt_helper
+ or Settings._prompt_helper
+ or PromptHelper.from_llm_metadata(
+ llm.metadata,
+ )
+ )
if response_mode == ResponseMode.REFINE:
return Refine(
diff --git a/llama-index-core/llama_index/core/response_synthesizers/generation.py b/llama-index-core/llama_index/core/response_synthesizers/generation.py
index 396cc561b3..e6f176017e 100644
--- a/llama-index-core/llama_index/core/response_synthesizers/generation.py
+++ b/llama-index-core/llama_index/core/response_synthesizers/generation.py
@@ -22,6 +22,9 @@ class Generation(BaseSynthesizer):
# deprecated
service_context: Optional[ServiceContext] = None,
) -> None:
+ if service_context is not None:
+ prompt_helper = service_context.prompt_helper
+
super().__init__(
llm=llm,
callback_manager=callback_manager,
diff --git a/llama-index-core/llama_index/core/response_synthesizers/refine.py b/llama-index-core/llama_index/core/response_synthesizers/refine.py
index b162923661..8dd0fcc509 100644
--- a/llama-index-core/llama_index/core/response_synthesizers/refine.py
+++ b/llama-index-core/llama_index/core/response_synthesizers/refine.py
@@ -107,6 +107,9 @@ class Refine(BaseSynthesizer):
# deprecated
service_context: Optional[ServiceContext] = None,
) -> None:
+ if service_context is not None:
+ prompt_helper = service_context.prompt_helper
+
super().__init__(
llm=llm,
callback_manager=callback_manager,
diff --git a/llama-index-core/llama_index/core/response_synthesizers/simple_summarize.py b/llama-index-core/llama_index/core/response_synthesizers/simple_summarize.py
index e0c81be5f3..152459df1c 100644
--- a/llama-index-core/llama_index/core/response_synthesizers/simple_summarize.py
+++ b/llama-index-core/llama_index/core/response_synthesizers/simple_summarize.py
@@ -24,6 +24,9 @@ class SimpleSummarize(BaseSynthesizer):
# deprecated
service_context: Optional[ServiceContext] = None,
) -> None:
+ if service_context is not None:
+ prompt_helper = service_context.prompt_helper
+
super().__init__(
llm=llm,
callback_manager=callback_manager,
diff --git a/llama-index-core/llama_index/core/response_synthesizers/tree_summarize.py b/llama-index-core/llama_index/core/response_synthesizers/tree_summarize.py
index 066163a51f..356e5dfee0 100644
--- a/llama-index-core/llama_index/core/response_synthesizers/tree_summarize.py
+++ b/llama-index-core/llama_index/core/response_synthesizers/tree_summarize.py
@@ -41,6 +41,9 @@ class TreeSummarize(BaseSynthesizer):
# deprecated
service_context: Optional[ServiceContext] = None,
) -> None:
+ if service_context is not None:
+ prompt_helper = service_context.prompt_helper
+
super().__init__(
llm=llm,
callback_manager=callback_manager,
diff --git a/llama-index-core/llama_index/core/settings.py b/llama-index-core/llama_index/core/settings.py
index d7875e566f..c280f8a88e 100644
--- a/llama-index-core/llama_index/core/settings.py
+++ b/llama-index-core/llama_index/core/settings.py
@@ -197,8 +197,6 @@ class _Settings:
"""Set the text splitter."""
self.node_parser = text_splitter
- # ---- Prompt helper ----
-
@property
def prompt_helper(self) -> PromptHelper:
"""Get the prompt helper."""
@@ -296,16 +294,6 @@ def node_parser_from_settings_or_context(
return settings.node_parser
-def prompt_helper_from_settings_or_context(
- settings: _Settings, context: Optional["ServiceContext"]
-) -> PromptHelper:
- """Get settings from either settings or context."""
- if context is not None:
- return context.prompt_helper
-
- return settings.prompt_helper
-
-
def transformations_from_settings_or_context(
settings: _Settings, context: Optional["ServiceContext"]
) -> List[TransformComponent]:
--
GitLab