issue #4: llama index deployment design

.devcontainer/devcontainer.json

@@ -1,7 +1,7 @@
 {
   "name": "Python 3 with Jupyter",
   "image": "mcr.microsoft.com/devcontainers/python:3.11",
-  "postCreateCommand": "pip3 install --user -r requirements.txt && pip3 install --user -r document-search/requirements.txt",
+  "postCreateCommand": "pip3 install --user -r requirements.txt",
   "customizations": {
     "vscode": {
       "extensions": [

.env.template

@@ -1,32 +1,44 @@
-# API Key for Azure OpenAI
-API_KEY=
+# Storage Account Connection String. Obtain this from your Azure Storage account
+# in the Azure Portal under "Access keys".
+STORAGE_ACCOUNT_CONN_STRING=
 
-# Endpoint URL for Azure OpenAI
+# API Key for Azure OpenAI. Same value for both. Get this from the Azure portal
+# under your Azure OpenAI Service resource.
+OPENAI_API_KEY=
+AZURE_OPENAI_API_KEY=
+
+# Endpoint URL for Azure OpenAI. Find this in the Overview section of your Azure
+# OpenAI Service resource in the Azure portal.
 AZURE_ENDPOINT=
 
-# API Version for Azure OpenAI
-API_VERSION=
+# API Version for Azure OpenAI Check the Azure OpenAI documentation for the
+# current API version
+# OPENAI_API_VERSION=2023-07-01-preview
 
-# Language Model for Azure OpenAI.
-# LLM_MODEL=
+# Language Model for Azure OpenAI. Specify the model you wish to use.
+# LLM_MODEL=gpt-4
 
-# Language Model Deployment Name for Azure OpenAI
-# LLM_DEPLOYMENT_NAME=
+# Language Model Deployment Name for Azure OpenAI If you have a custom
+# deployment, specify its name here. Otherwise, leave commented
+# LLM_DEPLOYMENT_NAME=ailab-llm
 
-# Embedding Model for Azure OpenAI
-# EMBED_MODEL=
+# Embedding Model for Azure OpenAI Specify the embedding model you wish to use.
+# Check Azure OpenAI documentation for available models 
+# EMBED_MODEL=ada
 
-# Embedding Model Deployment Name for Azure OpenAI
-# EMBED_DEPLOYMENT_NAME=
+# Embedding Model Deployment Name for Azure OpenAI.
+# EMBED_DEPLOYMENT_NAME=text-embedding-ada-002
 
-# Database Connection Configuration
+# Database Connection Configuration These details will be specific to your
+# database. Obtain them from your database administrator or setup
 DB_SCHEME=
 DB_HOST=
 DB_PORT=
 DB_USER=
 DB_PASSWORD=
 DB_NAME=
 
-# Table and Column for Querying the Database
+# Table and Column for Querying the Database Specify the table and column you
+# wish to query. This will depend on your database schema 
 # TABLE_NAME=
 # COLUMN=

.gitignore

@@ -56,3 +56,12 @@ venv/
 ENV/
 env.bak/
 venv.bak/
+
+# Index files
+index/
+
+# Pickle files
+*.pkl
+
+# OS specific
+.DS_Store

docs/puml/components.plantuml

@@ -0,0 +1,24 @@
+@startuml components
+left to right direction
+interface REST
+
+component "LlamaIndex API" as api {
+    package app as "flask app"
+    package LlamaIndex as "llamaIndex_db" {
+        package llama_index
+    }
+}
+app --> LlamaIndex: uses
+REST -- api
+
+folder Container as "Index folder" {
+    file doc as "default__vector_store.json"
+    file index as "docstore.json"
+    file vector as "graph_store.json"
+    file graph as "image__vector_store.json"
+    file image as "index_store.json"
+}
+
+llama_index --> Container : reads
+
+@enduml

docs/puml/deployment.plantuml

@@ -0,0 +1,16 @@
+@startuml deployment
+left to right direction
+interface REST
+
+cloud Azure {
+    node docker as "Docker container" {
+        component "LlamaIndex API" as api
+    }
+    database volume as "Docker Volume" {
+        folder Container as "Index folder"
+    }
+}
+
+REST -- api
+api --> Container : reads
+@enduml

docs/puml/search_sequence.plantuml

@@ -0,0 +1,19 @@
+@startuml search sequence
+
+actor user
+participant "app:Flask" as app
+participant "config:Config" as config
+participant ":VectorIndexRetriever" as retriever
+
+user -> app: POST /search query_string
+activate app
+app -> config: get retriever
+app -> retriever: retrieve with query_string
+activate retriever
+retriever --> app: nodes
+deactivate retriever
+app -> app: generate results from nodes
+app --> user: results
+deactivate app
+
+@enduml

docs/puml/startup_sequence.plantuml

@@ -0,0 +1,20 @@
+@startuml startup sequence
+
+participant "App Creator" as creator
+
+create ":StorageContext" as sc
+creator -> sc: create with index_folder
+
+create ":BaseIndex" as index
+creator -> index: create with storage_context
+
+create ":VectorIndexRetriever" as retriever
+creator -> retriever: create with index
+
+create ":Config" as conf
+creator -> conf: create with retriever
+
+create ":Flask" as app
+creator -> app: create with config
+
+@enduml