# LlamaIndex Managed Integration: Vectara
The Vectara Index provides a simple implementation to Vectara's end-to-end RAG pipeline,
including data ingestion, document retrieval, reranking results, summary generation, and hallucination evaluation.
## Setup
First, make sure you have the latest LlamaIndex version installed.
Next, install the Vectara Index:
```
pip install -U llama-index-indices-managed-vectara
```
Finally, set up your Vectara corpus. If you don't have a Vectara account, you can [sign up](https://vectara.com/integrations/llamaindex) and follow our [Quick Start](https://docs.vectara.com/docs/quickstart) guide to create a corpus and an API key (make sure it has both indexing and query permissions).
## Usage
First let's initialize the index with some sample documents.
```python
import os
os.environ["VECTARA_API_KEY"] = "<YOUR_VECTARA_API_KEY>"
os.environ["VECTARA_CORPUS_ID"] = "<YOUR_VECTARA_CORPUS_ID>"
os.environ["VECTARA_CUSTOMER_ID"] = "<YOUR_VECTARA_CUSTOMER_ID>"
from llama_index.indices.managed.vectara import VectaraIndex
from llama_index.core.schema import Document
docs = [
Document(
text="""
This is test text for Vectara integration with LlamaIndex.
Users should love their experience with this integration
""",
),
Document(
text="""
The Vectara index integration with LlamaIndex implements Vectara's RAG pipeline.
It can be used both as a retriever and query engine.
""",
),
]
index = VectaraIndex.from_documents(docs)
```
You can now use this index to retrieve documents.
```python
# Retrieves the top search result
retriever = index.as_retriever(similarity_top_k=1)
results = retriever.retrieve("How will users feel about this new tool?")
print(results[0])
```
You can also use it as a query engine to get a generated summary from the retrieved results.
```python
query_engine = index.as_query_engine()
results = query_engine.query(
"Which company has partnered with Vectara to implement their RAG pipeline as an index?"
)
print(f"Generated summary: {results.response}\n")
print("Top sources:")
for node in results.source_nodes[:2]:
print(node)
```
If you want to see the full features and capabilities of `VectaraIndex`, check out this Jupyter [notebook](https://github.com/vectara/example-notebooks/blob/main/notebooks/using-vectara-with-llamaindex.ipynb).
Raw data
{
"_id": null,
"home_page": null,
"name": "llama-index-indices-managed-vectara",
"maintainer": null,
"docs_url": null,
"requires_python": "<4.0,>=3.9",
"maintainer_email": null,
"keywords": null,
"author": "David Oplatka",
"author_email": "david.oplatka@vectara.com",
"download_url": "https://files.pythonhosted.org/packages/cd/2a/5122cee7fec604d696990a25bc165f30a6da11c5219069cff93a17e6709d/llama_index_indices_managed_vectara-0.3.1.tar.gz",
"platform": null,
"description": "# LlamaIndex Managed Integration: Vectara\n\nThe Vectara Index provides a simple implementation to Vectara's end-to-end RAG pipeline,\nincluding data ingestion, document retrieval, reranking results, summary generation, and hallucination evaluation.\n\n## Setup\n\nFirst, make sure you have the latest LlamaIndex version installed.\n\nNext, install the Vectara Index:\n\n```\npip install -U llama-index-indices-managed-vectara\n```\n\nFinally, set up your Vectara corpus. If you don't have a Vectara account, you can [sign up](https://vectara.com/integrations/llamaindex) and follow our [Quick Start](https://docs.vectara.com/docs/quickstart) guide to create a corpus and an API key (make sure it has both indexing and query permissions).\n\n## Usage\n\nFirst let's initialize the index with some sample documents.\n\n```python\nimport os\n\nos.environ[\"VECTARA_API_KEY\"] = \"<YOUR_VECTARA_API_KEY>\"\nos.environ[\"VECTARA_CORPUS_ID\"] = \"<YOUR_VECTARA_CORPUS_ID>\"\nos.environ[\"VECTARA_CUSTOMER_ID\"] = \"<YOUR_VECTARA_CUSTOMER_ID>\"\n\nfrom llama_index.indices.managed.vectara import VectaraIndex\nfrom llama_index.core.schema import Document\n\ndocs = [\n Document(\n text=\"\"\"\n This is test text for Vectara integration with LlamaIndex.\n Users should love their experience with this integration\n \"\"\",\n ),\n Document(\n text=\"\"\"\n The Vectara index integration with LlamaIndex implements Vectara's RAG pipeline.\n It can be used both as a retriever and query engine.\n \"\"\",\n ),\n]\n\nindex = VectaraIndex.from_documents(docs)\n```\n\nYou can now use this index to retrieve documents.\n\n```python\n# Retrieves the top search result\nretriever = index.as_retriever(similarity_top_k=1)\n\nresults = retriever.retrieve(\"How will users feel about this new tool?\")\nprint(results[0])\n```\n\nYou can also use it as a query engine to get a generated summary from the retrieved results.\n\n```python\nquery_engine = index.as_query_engine()\n\nresults = query_engine.query(\n \"Which company has partnered with Vectara to implement their RAG pipeline as an index?\"\n)\nprint(f\"Generated summary: {results.response}\\n\")\nprint(\"Top sources:\")\nfor node in results.source_nodes[:2]:\n print(node)\n```\n\nIf you want to see the full features and capabilities of `VectaraIndex`, check out this Jupyter [notebook](https://github.com/vectara/example-notebooks/blob/main/notebooks/using-vectara-with-llamaindex.ipynb).\n",
"bugtrack_url": null,
"license": "MIT",
"summary": "llama-index managed vectara integration",
"version": "0.3.1",
"project_urls": null,
"split_keywords": [],
"urls": [
{
"comment_text": "",
"digests": {
"blake2b_256": "ba8f89739ba5716050c88fe0422b392c978c4daa41a17f2839bb9cb3773f4089",
"md5": "fe8bfb591b59f68495f366406270c6ac",
"sha256": "4631d133f49025d2f62f19d8219574b4074c51132afec35ffb880dfb22f52ebd"
},
"downloads": -1,
"filename": "llama_index_indices_managed_vectara-0.3.1-py3-none-any.whl",
"has_sig": false,
"md5_digest": "fe8bfb591b59f68495f366406270c6ac",
"packagetype": "bdist_wheel",
"python_version": "py3",
"requires_python": "<4.0,>=3.9",
"size": 16838,
"upload_time": "2024-12-07T05:57:49",
"upload_time_iso_8601": "2024-12-07T05:57:49.778024Z",
"url": "https://files.pythonhosted.org/packages/ba/8f/89739ba5716050c88fe0422b392c978c4daa41a17f2839bb9cb3773f4089/llama_index_indices_managed_vectara-0.3.1-py3-none-any.whl",
"yanked": false,
"yanked_reason": null
},
{
"comment_text": "",
"digests": {
"blake2b_256": "cd2a5122cee7fec604d696990a25bc165f30a6da11c5219069cff93a17e6709d",
"md5": "72071d230e2c26543b5df94e879cf502",
"sha256": "1dd7bf9b5029e69d721734a4c220b6f28652904650c7234b736a23318a0e2bd5"
},
"downloads": -1,
"filename": "llama_index_indices_managed_vectara-0.3.1.tar.gz",
"has_sig": false,
"md5_digest": "72071d230e2c26543b5df94e879cf502",
"packagetype": "sdist",
"python_version": "source",
"requires_python": "<4.0,>=3.9",
"size": 15501,
"upload_time": "2024-12-07T05:57:51",
"upload_time_iso_8601": "2024-12-07T05:57:51.286812Z",
"url": "https://files.pythonhosted.org/packages/cd/2a/5122cee7fec604d696990a25bc165f30a6da11c5219069cff93a17e6709d/llama_index_indices_managed_vectara-0.3.1.tar.gz",
"yanked": false,
"yanked_reason": null
}
],
"upload_time": "2024-12-07 05:57:51",
"github": false,
"gitlab": false,
"bitbucket": false,
"codeberg": false,
"lcname": "llama-index-indices-managed-vectara"
}
JSON