Localhost plugin support

.dockerignore

@@ -4,6 +4,7 @@
 scripts/
 tests/
 examples/
+local-server/
 *.md
 *.pyc
 .dockerignore

README.md

@@ -42,9 +42,11 @@ This README provides detailed information on how to set up, develop, and deploy
     - [Qdrant](#qdrant)
     - [Redis](#redis)
   - [Running the API Locally](#running-the-api-locally)
+  - [Testing a Localhost Plugin in ChatGPT](#testing-a-localhost-plugin-in-chatgpt)
   - [Personalization](#personalization)
   - [Authentication Methods](#authentication-methods)
 - [Deployment](#deployment)
+- [Installing a Developer Plugin](#installing-a-developer-plugin)
 - [Webhooks](#webhooks)
 - [Scripts](#scripts)
 - [Limitations](#limitations)
@@ -122,6 +124,13 @@ Follow these steps to quickly set up and run the ChatGPT Retrieval Plugin:
 9. Run the API locally: `poetry run start`
 10. Access the API documentation at `http://0.0.0.0:8000/docs` and test the API endpoints (make sure to add your bearer token).
 
+### Testing in ChatGPT
+
+To test a locally hosted plugin in ChatGPT, follow these steps:
+
+1. Run the API on localhost: `poetry run dev`
+2. Follow the instructions in the [Testing a Localhost Plugin in ChatGPT](#testing-a-localhost-plugin-in-chatgpt) section of the README.
+
 For more detailed information on setting up, developing, and deploying the ChatGPT Retrieval Plugin, refer to the full Development section below.
 
 ## About
@@ -257,7 +266,6 @@ For more detailed instructions on setting up and using each vector database prov
 
 [Redis](https://redis.com/solutions/use-cases/vector-database/) is a real-time data platform suitable for a variety of use cases, including everyday applications and AI/ML workloads. It can be used as a low-latency vector engine by creating a Redis database with the [Redis Stack docker container](/examples/docker/redis/docker-compose.yml). For a hosted/managed solution, [Redis Cloud](https://app.redislabs.com/#/) is available. For detailed setup instructions, refer to [`/docs/providers/redis/setup.md`](/docs/providers/redis/setup.md).
 
-
 #### LlamaIndex
 
 [LlamaIndex](https://github.com/jerryjliu/llama_index) is a central interface to connect your LLM's with external data.
@@ -289,6 +297,22 @@ Append `docs` to the URL shown in the terminal and open it in a browser to acces
 
 **Note:** If you add new dependencies to the pyproject.toml file, you need to run `poetry lock` and `poetry install` to update the lock file and install the new dependencies.
 
+### Testing a Localhost Plugin in ChatGPT
+
+To test a localhost plugin in ChatGPT, use the provided [`local-server/main.py`](/local-server/main.py) file, which is specifically configured for localhost testing with CORS settings, no authentication and routes for the manifest, OpenAPI schema and logo.
+
+Follow these steps to test your localhost plugin:
+
+1. Run the localhost server using the `poetry run dev` command. This starts the server at the default address (e.g. `localhost:3333`).
+
+2. Visit [ChatGPT](https://chat.openai.com/), select "Plugins" from the model picker, click on the plugins picker, and click on "Plugin store" at the bottom of the list.
+
+3. Choose "Develop your own plugin" and enter your localhost URL (e.g. `localhost:3333`) when prompted.
+
+4. Your localhost plugin is now enabled for your ChatGPT session.
+
+For more information, refer to the [OpenAI documentation](https://platform.openai.com/docs/plugins/getting-started/openapi-definition).
+
 ### Personalization
 
 You can personalize the Retrieval Plugin for your own use case by doing the following:

local-server/ai-plugin.json

@@ -0,0 +1,18 @@
+{
+  "schema_version": "v1",
+  "name_for_model": "retrieval",
+  "name_for_human": "Retrieval Plugin",
+  "description_for_model": "Plugin for searching through the user's documents (such as files, emails, and more) to find answers to questions and retrieve relevant information. Use it whenever a user asks something that might be found in their personal information.",
+  "description_for_human": "Search through your documents.",
+  "auth": {
+    "type": "none"
+  },
+  "api": {
+    "type": "openapi",
+    "url": "http://localhost:3333/.well-known/openapi.yaml"
+  },
+  "logo_url": "http://localhost:3333/.well-known/logo.png",
+  "contact_email": "[email protected]", 
+  "legal_info_url": "[email protected]"
+}
+

local-server/main.py

@@ -0,0 +1,145 @@
+# This is a version of the main.py file found in ../../../server/main.py for testing the plugin locally.
+# Use the command `poetry run dev` to run this.
+from typing import Optional
+import uvicorn
+from fastapi import FastAPI, File, Form, HTTPException, Body, UploadFile
+
+from models.api import (
+    DeleteRequest,
+    DeleteResponse,
+    QueryRequest,
+    QueryResponse,
+    UpsertRequest,
+    UpsertResponse,
+)
+from datastore.factory import get_datastore
+from services.file import get_document_from_file
+
+from starlette.responses import FileResponse
+
+from models.models import DocumentMetadata, Source
+from fastapi.middleware.cors import CORSMiddleware
+
+
+app = FastAPI()
+
+PORT = 3333
+
+origins = [
+    f"http://localhost:{PORT}",
+    "https://chat.openai.com",
+]
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=origins,
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+
+@app.route("/.well-known/ai-plugin.json")
+async def get_manifest(request):
+    file_path = "./local-server/ai-plugin.json"
+    return FileResponse(file_path, media_type="text/json")
+
+
+@app.route("/.well-known/logo.png")
+async def get_logo(request):
+    file_path = "./local-server/logo.png"
+    return FileResponse(file_path, media_type="text/json")
+
+
+@app.route("/.well-known/openapi.yaml")
+async def get_openapi(request):
+    file_path = "./local-server/openapi.yaml"
+    return FileResponse(file_path, media_type="text/json")
+
+
+@app.post(
+    "/upsert-file",
+    response_model=UpsertResponse,
+)
+async def upsert_file(
+    file: UploadFile = File(...),
+    metadata: Optional[str] = Form(None),
+):
+    try:
+        metadata_obj = (
+            DocumentMetadata.parse_raw(metadata)
+            if metadata
+            else DocumentMetadata(source=Source.file)
+        )
+    except:
+        metadata_obj = DocumentMetadata(source=Source.file)
+
+    document = await get_document_from_file(file, metadata_obj)
+
+    try:
+        ids = await datastore.upsert([document])
+        return UpsertResponse(ids=ids)
+    except Exception as e:
+        print("Error:", e)
+        raise HTTPException(status_code=500, detail=f"str({e})")
+
+
+@app.post(
+    "/upsert",
+    response_model=UpsertResponse,
+)
+async def upsert(
+    request: UpsertRequest = Body(...),
+):
+    try:
+        ids = await datastore.upsert(request.documents)
+        return UpsertResponse(ids=ids)
+    except Exception as e:
+        print("Error:", e)
+        raise HTTPException(status_code=500, detail="Internal Service Error")
+
+
+@app.post("/query", response_model=QueryResponse)
+async def query_main(request: QueryRequest = Body(...)):
+    try:
+        results = await datastore.query(
+            request.queries,
+        )
+        return QueryResponse(results=results)
+    except Exception as e:
+        print("Error:", e)
+        raise HTTPException(status_code=500, detail="Internal Service Error")
+
+
+@app.delete(
+    "/delete",
+    response_model=DeleteResponse,
+)
+async def delete(
+    request: DeleteRequest = Body(...),
+):
+    if not (request.ids or request.filter or request.delete_all):
+        raise HTTPException(
+            status_code=400,
+            detail="One of ids, filter, or delete_all is required",
+        )
+    try:
+        success = await datastore.delete(
+            ids=request.ids,
+            filter=request.filter,
+            delete_all=request.delete_all,
+        )
+        return DeleteResponse(success=success)
+    except Exception as e:
+        print("Error:", e)
+        raise HTTPException(status_code=500, detail="Internal Service Error")
+
+
+@app.on_event("startup")
+async def startup():
+    global datastore
+    datastore = await get_datastore()
+
+
+def start():
+    uvicorn.run("local-server.main:app", host="localhost", port=PORT, reload=True)