Pebblo enables developers to safely load data and promote their Gen AI app to deployment without worrying about the organization’s compliance and security requirements. The project identifies semantic topics and entities found in the loaded data and summarizes them on the UI or a PDF report.
Pebblo has two components.
This document describes how to augment your existing Langchain DocumentLoader with Pebblo Safe DocumentLoader to get deep data visibility on the types of Topics and Entities ingested into the Gen-AI Langchain application. For details on Pebblo Server see this pebblo server document.
Pebblo Safeloader enables safe data ingestion for Langchain DocumentLoader. This is done by wrapping the document loader call with Pebblo Safe DocumentLoader.
Note: To configure pebblo server on some url other that pebblo's default (localhost:8000) url, put the correct URL in PEBBLO_CLASSIFIER_URL env variable. This is configurable using the classifier_url keyword argument as well. Ref: server-configurations
Assume a Langchain RAG application snippet using CSVLoader to read a CSV document for inference.
Here is the snippet of Document loading using CSVLoader.
from langchain_community.document_loaders import CSVLoader
loader = CSVLoader("data/corp_sens_data.csv")
documents = loader.load()
print(documents)
The Pebblo SafeLoader can be enabled with few lines of code change to the above snippet.
from langchain_community.document_loaders import CSVLoader, PebbloSafeLoader
loader = PebbloSafeLoader(
CSVLoader("data/corp_sens_data.csv"),
name="acme-corp-rag-1",
owner="Joe Smith",
description="Support productivity RAG application",
)
documents = loader.load()
print(documents)
To send semantic data to pebblo-cloud, pass api-key to PebbloSafeLoader as an argument or alternatively, put the api-key in PEBBLO_API_KEY environment variable.
from langchain_community.document_loaders import CSVLoader, PebbloSafeLoader
loader = PebbloSafeLoader(
CSVLoader("data/corp_sens_data.csv"),
name="acme-corp-rag-1",
owner="Joe Smith",
description="Support productivity RAG application",
api_key="my-api-key",
)
documents = loader.load()
print(documents)
To add semantic topics and sematic entities to metadata of loaded documents, set load_semantic to True as an argument or alternatively, define a new environment variable PEBBLO_LOAD_SEMANTIC, and setting it to True.
from langchain_community.document_loaders import CSVLoader, PebbloSafeLoader
loader = PebbloSafeLoader(
CSVLoader("data/corp_sens_data.csv"),
name="acme-corp-rag-1",
owner="Joe Smith",
description="Support productivity RAG application",
api_key="my-api-key",
load_semantic=True,
)
documents = loader.load()
print(documents[0].metadata)
Set anonymize_snippets to True to anonymize all personally identifiable information (PII) from the snippets going into VectorDB and the generated reports.
Note: The Pebblo Entity Classifier effectively identifies personally identifiable information (PII) and is continuously evolving. While its recall is not yet 100%, it is steadily improving. For more details, please refer to the Pebblo Entity Classifier docs
from langchain_community.document_loaders import CSVLoader, PebbloSafeLoader
loader = PebbloSafeLoader(
CSVLoader("data/corp_sens_data.csv"),
name="acme-corp-rag-1",
owner="Joe Smith",
description="Support productivity RAG application",
anonymize_snippets=True,
)
documents = loader.load()
print(documents[0].metadata)
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4