add create_web_app (#12)

* add create_web_app * update readme * update fastapi docs * fix styling * some fixes, add tests for web app, update readme * update requirements * update httpx versioning * add examples * minor fixes, llama_index version update * major refactor * minor fix * another minor fix * bugfix * fix tests --------- Co-authored-by: fatih <34196005+fcakyon@users.noreply.github.com> Co-authored-by: fcakyon <fcakyon@gmail.com>
safevideo · Oct 19, 2023 · 7ea85ba · 7ea85ba
1 parent f2a3d7f
commit 7ea85ba
Show file tree

Hide file tree

Showing 22 changed files with 254 additions and 147 deletions.
diff --git a/README.md b/README.md
@@ -79,11 +79,14 @@ from autollm import AutoVectorStoreIndex
 
 # Dynamically initialize a VectorStoreIndex instance with the same AutoVectorStoreIndex interface:
 
-vector_store_index = AutoVectorStoreIndex.from_defaults(vector_store_type="PineconeVectorStore", pinecone_index=pinecone.Index("quickstart"))
+vector_store_index = AutoVectorStoreIndex.from_defaults(
+ vector_store_type="PineconeVectorStore", pinecone_index=pinecone.Index("quickstart")
+)
 
 
 vector_store_index = AutoVectorStoreIndex.from_defaults(
- vector_store_type="VectorStoreIndex", documents=documents)
+ vector_store_type="VectorStoreIndex", documents=documents
+)
 ```
 
 ### AutoQueryEngine (Creates a query engine pipeline in a single line of code)
@@ -200,6 +203,30 @@ input_dir = "/local/documents/path"
 documents = local_document_provider(input_dir=input_dir)
 ```
 
+### FastAPI Integration in 1-Line 🚀
+
+Creating a FastAPI application integrated with AutoLLM has never been easier. Follow the quick guide below to get started.
+
+#### Create Your FastAPI Application
+
+In your `main.py`, include the following line of code:
+
+```python
+from autollm import create_web_app
+
+app = create_web_app(config_path, env_path)
+```
+
+Here, `config` and `env` should be replaced by your configuration and environment file paths.
+
+#### Run Your Application
+
+After creating your FastAPI app, run the following command in your terminal to get it up and running:
+
+```bash
+uvicorn main:app
+```
+
 ______________________________________________________________________
 
 ## FAQ
@@ -250,7 +277,7 @@ ______________________________________________________________________
 
 ## License
 
-AutoLLM is available under the [GNU Affero General Public License (AGPL 3.0)](LICENSE.txt).
+AutoLLM is available under the [GNU Affero General Public License (AGPL 3.0)](LICENSE).
 
 ______________________________________________________________________
 

diff --git a/autollm/__init__.py b/autollm/__init__.py
@@ -12,5 +12,6 @@
 from autollm.auto.query_engine import AutoQueryEngine
 from autollm.auto.service_context import AutoServiceContext
 from autollm.auto.vector_store_index import AutoVectorStoreIndex
+from autollm.serve.utils import create_web_app
 
-__all__ = ['AutoLLM', 'AutoServiceContext', 'AutoVectorStoreIndex', 'AutoQueryEngine']
+__all__ = ['AutoLLM', 'AutoServiceContext', 'AutoVectorStoreIndex', 'AutoQueryEngine', 'create_web_app']
diff --git a/autollm/app/api/__init__.py b/autollm/app/api/__init__.py
diff --git a/autollm/app/api/ask_question.py b/autollm/app/api/ask_question.py
diff --git a/autollm/app/api/health_check.py b/autollm/app/api/health_check.py
diff --git a/autollm/app/docs.py b/autollm/app/docs.py
diff --git a/autollm/app/main.py b/autollm/app/main.py
diff --git a/autollm/auto/vector_store_index.py b/autollm/auto/vector_store_index.py
@@ -42,7 +42,7 @@ def from_defaults(
  if documents is None:
  documents = [Document.example()]
  if vector_store_type == "VectorStoreIndex":
- index = VectorStoreIndex.from_documents(documents=[documents], *args, **kwargs)
+ index = VectorStoreIndex.from_documents(documents=documents, *args, **kwargs)
  else:
  VectorStoreClass = import_vector_store_class(vector_store_type)
  vector_store = VectorStoreClass(*args, **kwargs)

diff --git a/autollm/app/__init__.py → autollm/serve/__init__.py b/autollm/app/__init__.py → autollm/serve/__init__.py
diff --git a/autollm/serve/docs.py b/autollm/serve/docs.py
@@ -0,0 +1,16 @@
+# Purpose: FastAPI documentation configuration.
+# Metadata
+title = "AutoLLM Query Engine"
+description = """
+This is a FastAPI service for AutoLLM's natural language query engine.
+It's designed to query multiple and big documents and get the most relevant results.
+"""
+version = "0.0.1"
+openapi_url = "/api/v1/openapi.json"
+terms_of_service = "Local Deployment, All Rights Reserved."
+tags_metadata = [
+ {
+ "name": "query",
+ "description": "Operations related to querying the header-documents."
+ },
+]
diff --git a/autollm/serve/utils.py b/autollm/serve/utils.py
@@ -0,0 +1,72 @@
+import logging
+from typing import Dict
+
+import yaml
+from dotenv import load_dotenv
+from fastapi import FastAPI, HTTPException
+from pydantic import BaseModel, Field
+
+from autollm.auto.query_engine import AutoQueryEngine
+from autollm.serve.docs import description, openapi_url, tags_metadata, terms_of_service, title, version
+
+logging.basicConfig(level=logging.INFO)
+
+
+# Function to load the configuration for tasks and initialize query engines
+def load_config_and_initialize_engines(config_file_path: str,
+ env_file_path: str = None) -> Dict[str, AutoQueryEngine]:
+ # Optionally load environment variables from a .env file
+ if env_file_path:
+ load_dotenv(dotenv_path=env_file_path)
+
+ # Load the YAML configuration file
+ with open(config_file_path) as f:
+ config = yaml.safe_load(f)
+
+ # Initialize query engines based on the config
+ query_engines = {}
+ for task_params in config['tasks']:
+ task_name = task_params.pop('name')
+ query_engines[task_name] = AutoQueryEngine.from_parameters(**task_params)
+
+ return query_engines
+
+
+class QueryPayload(BaseModel):
+ task: str = Field(..., description="Task to execute")
+ user_query: str = Field(..., description="User's query")
+
+
+# Function to create the FastAPI web app
+def create_web_app(config_file_path: str, env_file_path: str = None):
+ app = FastAPI(
+ title=title,
+ description=description,
+ version=version,
+ openapi_url=openapi_url,
+ terms_of_service=terms_of_service,
+ openapi_tags=tags_metadata,
+ )
+
+ query_engines = load_config_and_initialize_engines(config_file_path, env_file_path)
+
+ @app.post("/query")
+ async def query(payload: QueryPayload):
+ task = payload.task
+ user_query = payload.user_query
+
+ if task not in query_engines:
+ raise HTTPException(status_code=400, detail="Invalid task name")
+
+ # Use the appropriate query engine for the task
+ query_engine = query_engines[task]
+ response = query_engine.query(user_query)
+
+ return response
+
+ return app
+
+
+# For demonstration, let's assume we have a config.yaml with task configurations and an optional .env file
+# This function call would typically be in your main application file
+# app = create_web_app("path/to/config.yaml", "path/to/.env")
diff --git a/examples/configs/.env.example b/examples/configs/.env.example
@@ -0,0 +1,2 @@
+# .env.example
+OPENAI_API_KEY=your_openai_api_key_here
diff --git a/examples/configs/config.example.yaml b/examples/configs/config.example.yaml
@@ -0,0 +1,25 @@
+# config.example.yaml
+version: '1.0' # Version of this configuration file
+tasks:
+ - name: "summarize"
+ system_prompt: "Please summarize the following text:" # System prompt for this task
+ vector_store_params:
+ vector_store_type: "VectorStoreIndex"
+ llm_params:
+ model: "gpt-3.5-turbo"
+ service_context_params:
+ chunk_size: 1024
+ query_engine_params:
+ similarity_top_k: 5
+ enable_cost_calculator: true
+ - name: "qa"
+ system_prompt: "Question:" # System prompt for this task
+ vector_store_params:
+ vector_store_type: "VectorStoreIndex"
+ llm_params:
+ model: "gpt-3.5-turbo"
+ service_context_params:
+ chunk_size: 1024
+ query_engine_params:
+ similarity_top_k: 3
+ enable_cost_calculator: false
diff --git a/requirements.txt b/requirements.txt
@@ -1,6 +1,7 @@
-llama-index==0.8.41
+llama-index==0.8.46
 litellm==0.8.4
 gitpython==3.1.37
 uvicorn==0.23.2
 fastapi==0.103.2
 python-dotenv
+httpx
diff --git a/setup.py b/setup.py
@@ -28,12 +28,26 @@ def get_version():
  return re.search(r'^__version__ = [\'"]([^\'"]*)[\'"]', f.read(), re.M).group(1)
 
 
+def get_author():
+ current_dir = os.path.abspath(os.path.dirname(__file__))
+ init_file = os.path.join(current_dir, 'autollm', '__init__.py')
+ with open(init_file, encoding='utf-8') as f:
+ return re.search(r'^__author__ = [\'"]([^\'"]*)[\'"]', f.read(), re.M).group(1)
+
+
+def get_license():
+ current_dir = os.path.abspath(os.path.dirname(__file__))
+ init_file = os.path.join(current_dir, 'autollm', '__init__.py')
+ with open(init_file, encoding='utf-8') as f:
+ return re.search(r'^__license__ = [\'"]([^\'"]*)[\'"]', f.read(), re.M).group(1)
+
+
 setuptools.setup(
  name='autollm',
  version=get_version(),
- author='safevideo',
+ author=get_author(),
  author_email='support@safevideo.ai',
- license='AGPL-3.0',
+ license=get_license(),
  description="Ship RAG based LLM Web API's, in seconds.",
  long_description=get_long_description(),
  long_description_content_type='text/markdown',

diff --git a/tests/config.yaml b/tests/config.yaml
@@ -0,0 +1,24 @@
+# config.example.yaml
+tasks:
+ - name: "summarize"
+ system_prompt: "Please summarize the following text:" # System prompt for this task
+ vector_store_params:
+ vector_store_type: "VectorStoreIndex"
+ llm_params:
+ model: "gpt-3.5-turbo"
+ service_context_params:
+ chunk_size: 1024
+ query_engine_params:
+ similarity_top_k: 5
+ enable_cost_calculator: true
+ - name: "qa"
+ system_prompt: "Question:" # System prompt for this task
+ vector_store_params:
+ vector_store_type: "VectorStoreIndex"
+ llm_params:
+ model: "gpt-3.5-turbo"
+ service_context_params:
+ chunk_size: 1024
+ query_engine_params:
+ similarity_top_k: 3
+ enable_cost_calculator: false