-
Notifications
You must be signed in to change notification settings - Fork 34
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
COMMENTS - Empty PR #1256
base: empty-branch
Are you sure you want to change the base?
COMMENTS - Empty PR #1256
Conversation
Co-authored-by: George Burton <[email protected]>
Footnote citations
fix ingest error due to metadata size exceed max size
Feature/streaming with citations
…1130) * diabled gadget tests whilst we are actively working on them * reverted change * formatting
Bumps the pip group with 1 update in the /django_app directory: [werkzeug](https://github.com/pallets/werkzeug). Bumps the pip group with 1 update in the /redbox-core directory: [werkzeug](https://github.com/pallets/werkzeug). Bumps the pip group with 1 update in the / directory: [werkzeug](https://github.com/pallets/werkzeug). Updates `werkzeug` from 3.0.4 to 3.0.6 - [Release notes](https://github.com/pallets/werkzeug/releases) - [Changelog](https://github.com/pallets/werkzeug/blob/main/CHANGES.rst) - [Commits](pallets/werkzeug@3.0.4...3.0.6) Updates `werkzeug` from 3.0.4 to 3.0.6 - [Release notes](https://github.com/pallets/werkzeug/releases) - [Changelog](https://github.com/pallets/werkzeug/blob/main/CHANGES.rst) - [Commits](pallets/werkzeug@3.0.4...3.0.6) Updates `werkzeug` from 3.0.4 to 3.0.6 - [Release notes](https://github.com/pallets/werkzeug/releases) - [Changelog](https://github.com/pallets/werkzeug/blob/main/CHANGES.rst) - [Commits](pallets/werkzeug@3.0.4...3.0.6) --- updated-dependencies: - dependency-name: werkzeug dependency-type: indirect dependency-group: pip - dependency-name: werkzeug dependency-type: indirect dependency-group: pip - dependency-name: werkzeug dependency-type: indirect dependency-group: pip ... Signed-off-by: dependabot[bot] <[email protected]> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
…un_sync to Redbox app for testing
Fixed passing citations through in agentic search
* restored doc name * replaced more original_file_name
…#1138) Co-authored-by: gecBurton <[email protected]>
Calculating tokens from raw LLM response
* unique-citation uris * removed unused arg
remove ecr configuration
* Allow AWS OpenSearch IAM Permissions * Add OpenSearch Permissions
Removed gadget
Co-authored-by: Saisakul Chernbumroong <[email protected]>
* added csr and ecs for lit-ssr * Update infrastructure/aws/ecs.tf * added lit-ssr to Makefile * ecs is working * added service discovery * Add /health endpoint to lit-ssr * updated tf to match lit-ssr health endpoint * using correct port for LIT_SSR --------- Co-authored-by: Kevin Etchells <[email protected]>
chat_map_question_prompt: str = prompts.CHAT_MAP_QUESTION_PROMPT | ||
reduce_system_prompt: str = prompts.REDUCE_SYSTEM_PROMPT | ||
|
||
# Elasticsearch RAG and boost values |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
These can go?
|
||
# Task Prompt Fragments | ||
|
||
chat_system_prompt: str = prompts.CHAT_SYSTEM_PROMPT |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do we just need one system prompt now?
|
||
class RedboxState(BaseModel): | ||
request: RedboxQuery | ||
documents: Annotated[DocumentState, document_reducer] = DocumentState() |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Are we still going to attatch documents to the state like this? Or have it be something that is more linked to the chat history?
documents: Annotated[DocumentState, document_reducer] = DocumentState() | ||
route_name: str | None = None | ||
metadata: Annotated[RequestMetadata | None, metadata_reducer] = None | ||
citations: list[Citation] | None = None |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Remove?
return _merge | ||
|
||
|
||
def build_stuff_pattern( |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I am guessing this and the chat pattern will be the only ones we need?
prompt_set: PromptSet, | ||
output_parser: Runnable = None, | ||
format_instructions: str | None = None, | ||
tools: list[StructuredTool] | None = None, |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Remove tools?
stuff_chunk_context_ratio = models.FloatField(null=True, blank=True) | ||
recursion_limit = models.PositiveIntegerField(null=True, blank=True) | ||
|
||
chat_system_prompt = models.TextField(null=True, blank=True) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Am I right in thinking this is all needed because it was weirdly picking up empty strings and using them for the prompts?
I wonder if this is something we could improve while refactoring? Especially as we are slimming down the number of prompts.
chat_map_question_prompt = models.TextField(null=True, blank=True) | ||
reduce_system_prompt = models.TextField(null=True, blank=True) | ||
|
||
# Elsticsearch RAG and boost values |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
To delete?
document_loader(document_loader=loader, s3_client=s3_client, env=env) | ||
| RunnableLambda(list) | ||
| log_chunks | ||
| RunnableLambda(partial(vectorstore.add_documents, create_index_if_not_exists=False)) # type: ignore[arg-type] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
What does this look like without the vector store?
def _ingest_file(file_name: str, es_index_name: str = alias): | ||
logging.info("Ingesting file: %s", file_name) | ||
|
||
es = env.elasticsearch_client() |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
What does this look like without elasticsearch?
"""Run when chain errors.""" | ||
self.logger.error("Chain error: %s", error) | ||
|
||
def on_text(self, text: str, **kwargs: Any) -> None: # noqa:ARG002 |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is this where message logging happens?
I created an empty branch and look to merge feature/remove-search into it to allow me to comment across the whole repo