Skip to content

Faiss integration #351

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.

Sign up for GitHub

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

Merged
merged 6 commits into from
Jun 11, 2024
Merged

Faiss integration #351

Show file tree
Hide file tree
Changes from 3 commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
7a13a68
feat: refactoring of rag node
VinciGit00 Jun 4, 2024
7ed2fe8
feat: add dynamic caching
VinciGit00 Jun 4, 2024
d790361
feat: add caching
VinciGit00 Jun 6, 2024
543b487
add default folder for the cache
VinciGit00 Jun 7, 2024
c881f64
fix(cache): correctly pass the node arguments and logging
PeriniM Jun 11, 2024
edddb68
docs(cache): added cache_path param
PeriniM Jun 11, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions .gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,6 +23,7 @@ docs/source/_static/
venv/
.venv/
.vscode/
.conda/

# exclude pdf, mp3
*.pdf
Expand All @@ -38,3 +39,6 @@ lib/
*.html
.idea

# extras
cache/
run_smart_scraper.py
15 changes: 14 additions & 1 deletion scrapegraphai/nodes/rag_node.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3,6 +3,7 @@
"""

from typing import List, Optional
import os

from langchain.docstore.document import Document
from langchain.retrievers import ContextualCompressionRetriever
Expand Down Expand Up @@ -98,7 +99,19 @@ def execute(self, state: dict) -> dict:
)
embeddings = self.embedder_model

retriever = FAISS.from_documents(chunked_docs, embeddings).as_retriever()
folder_name = "cache"
Copy link
Collaborator

@DiTo97 DiTo97 Jun 6, 2024 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

that's the only thing I'd change.

as it stands, each RAG node with the cache config would overwrite the very same folder, effectively nullifying the cache, unless a RAG node on the very same input state is invoked subsequently.

I would allow for more flexibility by either 1): allowing the user to specify the cache folder in the config for that specific RAG node or 2) automatically compute a unique identifier for that specific input state (e.g., the filename, the endpoint or a some hashing function on the documents) so that the user may choose what to keep and what not to keep in the cache


if self.node_config.get("cache", False) and not os.path.exists(folder_name):
index = FAISS.from_documents(chunked_docs, embeddings)
os.makedirs(folder_name)

index.save_local(folder_name)
if self.node_config.get("cache", False) and os.path.exists(folder_name):
index = FAISS.load_local(folder_path=folder_name, embeddings=embeddings)
else:
index = FAISS.from_documents(chunked_docs, embeddings)

retriever = index.as_retriever()

redundant_filter = EmbeddingsRedundantFilter(embeddings=embeddings)
# similarity_threshold could be set, now k=20
Expand Down