From 4f551ce41490c16f0f29249b9c391a73646721e4 Mon Sep 17 00:00:00 2001
From: mudler <mudler@mocaccino.org>
Date: Sun, 7 May 2023 09:05:24 +0200
Subject: [PATCH] examples: add update index example, update README

---
 examples/query_data/README.md | 27 ++++++++++++++++++++++++---
 examples/query_data/query.py  |  6 ++++--
 examples/query_data/update.py | 32 ++++++++++++++++++++++++++++++++
 3 files changed, 60 insertions(+), 5 deletions(-)
 create mode 100644 examples/query_data/update.py

diff --git a/examples/query_data/README.md b/examples/query_data/README.md
index fb32442..9185709 100644
--- a/examples/query_data/README.md
+++ b/examples/query_data/README.md
@@ -4,11 +4,17 @@ This example makes use of [Llama-Index](https://gpt-index.readthedocs.io/en/stab
 
 It loosely follows [the quickstart](https://gpt-index.readthedocs.io/en/stable/guides/primer/usage_pattern.html).
 
+Summary of the steps:
+
+- prepare the dataset (and store it into `data`)
+- prepare a vector index database to run queries on
+- run queries
+
 ## Requirements
 
-For this in order to work, you will need a model compatible with the `llama.cpp` backend. This is will not work with gpt4all.
+For this in order to work, you will need LocalAI and a model compatible with the `llama.cpp` backend. This is will not work with gpt4all, however you can mix models (use a llama.cpp one to build the index database, and gpt4all to query it).
 
-The example uses `WizardLM`. Edit the config files in `models/` accordingly to specify the model you use (change `HERE`).
+The example uses `WizardLM` for both embeddings and Q&A. Edit the config files in `models/` accordingly to specify the model you use (change `HERE` in the configuration files).
 
 You will also need a training data set. Copy that over `data`.
 
@@ -28,7 +34,9 @@ cd LocalAI/examples/query_data
 docker-compose up -d --build
 ```
 
-### Create a storage:
+### Create a storage
+
+In this step we will create a local vector database from our document set, so later we can ask questions on it with the LLM.
 
 ```bash
 export OPENAI_API_BASE=http://localhost:8080/v1
@@ -41,9 +49,22 @@ After it finishes, a directory "storage" will be created with the vector index d
 
 ## Query
 
+We can now query the dataset. 
+
 ```bash
 export OPENAI_API_BASE=http://localhost:8080/v1
 export OPENAI_API_KEY=sk-
 
 python query.py
+```
+
+## Update
+
+To update our vector database, run `update.py`
+
+```bash
+export OPENAI_API_BASE=http://localhost:8080/v1
+export OPENAI_API_KEY=sk-
+
+python update.py
 ```
\ No newline at end of file
diff --git a/examples/query_data/query.py b/examples/query_data/query.py
index bf268c5..0f1408e 100644
--- a/examples/query_data/query.py
+++ b/examples/query_data/query.py
@@ -29,5 +29,7 @@ storage_context = StorageContext.from_defaults(persist_dir='./storage')
 index = load_index_from_storage(storage_context,     service_context=service_context,    )
 
 query_engine = index.as_query_engine()
-response = query_engine.query("XXXXXX your question here XXXXX")
-print(response)
\ No newline at end of file
+
+data = input("Question: ")
+response = query_engine.query(data)
+print(response)
diff --git a/examples/query_data/update.py b/examples/query_data/update.py
new file mode 100644
index 0000000..9ad283c
--- /dev/null
+++ b/examples/query_data/update.py
@@ -0,0 +1,32 @@
+import os
+
+# Uncomment to specify your OpenAI API key here (local testing only, not in production!), or add corresponding environment variable (recommended)
+# os.environ['OPENAI_API_KEY']= ""
+
+from llama_index import   LLMPredictor, PromptHelper, SimpleDirectoryReader, ServiceContext
+from langchain.llms.openai import OpenAI
+from llama_index import StorageContext, load_index_from_storage
+
+base_path = os.environ.get('OPENAI_API_BASE', 'http://localhost:8080/v1')
+
+# This example uses text-davinci-003 by default; feel free to change if desired
+llm_predictor = LLMPredictor(llm=OpenAI(temperature=0, model_name="gpt-3.5-turbo", openai_api_base=base_path))
+
+# Configure prompt parameters and initialise helper
+max_input_size = 1024
+num_output = 256
+max_chunk_overlap = 20
+
+prompt_helper = PromptHelper(max_input_size, num_output, max_chunk_overlap)
+
+# Load documents from the 'data' directory
+service_context = ServiceContext.from_defaults(llm_predictor=llm_predictor, prompt_helper=prompt_helper)
+
+# rebuild storage context
+storage_context = StorageContext.from_defaults(persist_dir='./storage')
+
+# load index
+index = load_index_from_storage(storage_context,     service_context=service_context,    )
+documents = SimpleDirectoryReader('data').load_data()
+index.refresh(documents)
+index.storage_context.persist(persist_dir="./storage")
\ No newline at end of file