Build the frontend with the docker container, make runnable after ins…

…tall (#37)

.dockerignore

@@ -3,11 +3,12 @@
 
 # Include (don't ignore) the application code
 !dewy
+!frontend
 !./pyproject.toml
 !./poetry.lock
 
 # Re-ignore pycache within `dewy`.
 **/__pycache__
 
 # Include (don't ignore) the migrations.
-!migrations/*.sql
+!migrations/*.sql

Dockerfile

@@ -9,7 +9,17 @@ COPY ./pyproject.toml ./poetry.lock* /tmp/
 RUN poetry export -f requirements.txt --output requirements.txt --without-hashes
 
 ######
-# 2. Create the actual image.
+# 2. Compile the frontend 
+FROM node:20.9.0-alpine as frontend-stage
+WORKDIR /app
+COPY ./frontend/package.json ./package.json
+RUN npm install --silent
+
+COPY ./frontend/ ./
+RUN npm run build
+
+######
+# 3. Create the actual image.
 FROM python:3.11
 WORKDIR /code
 
@@ -19,6 +29,7 @@ RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
 
 # Finally, copy in the application code.
 COPY ./dewy /code/dewy
+COPY --from=frontend-stage /app/dist /code/dewy/frontend/dist
 
 COPY ./migrations/0001_schema.sql /code/migrations/0001_schema.sql
 

README.md

@@ -47,21 +47,33 @@ Dewy handles all of the parts shown in brown, letting you focus on your applicat
 
 To get a local copy up and running follow these steps.
 
+1. (Optional) Start a `pgvector` instance to persist your data
+    ```sh
+    docker run -d \
+      -p 5432:5432 \
+      -e POSTGRES_DB=dewydb \
+      -e POSTGRES_USER=dewydbuser \
+      -e POSTGRES_PASSWORD=dewydbpwd \
+      -e POSTGRES_HOST_AUTH_METHOD=trust \
+      ankane/pgvector
+    ```
+1. Install Dewy
+    ```
+    pip install dewy
+    ```
 1. Fire up Dewy
     ```sh
     // Configure your OpenAI key (optional - local models will be used if not provided)
     export OPENAI_API_KEY=...
 
+    // Configure your pgvector endpoint (this assumes you installed with Docker as shown previously)
+    export DB=postgresql://dewydbuser:dewydbpwd@localhost/dewydb
+
     // Run the docker container
-<<<<<<< Updated upstream
-    docker run -d dewy-kb
+    dewy
 
-=======
-    docker run -d dewykb/dewy
-
->>>>>>> Stashed changes
     // Go to the management console to start creating resources!
-    open localhost:3001
+    open http://localhost/admin
     ```
 1. Install the API client library
     ```sh
@@ -136,15 +148,13 @@ Don't forget to give the project a star! Thanks again!
     OPENAI_API_KEY=...
     EOF
     ```
-1. Run the Dewy service
+1. Build the frontend
     ```sh
-    poetry run uvicorn dewy.main:app --host 0.0.0.0 --port 8000
+    cd frontend && npm install && npm run build
     ```
-1. Run the admin frontend (optional)
+1. Run the Dewy service
     ```sh
-    cd frontend
-    npm install
-    npm run dev
+    poetry run dewy
     ```
 
 ### Practices
@@ -167,4 +177,4 @@ If you're in a `poetry shell`, you can omit the `poetry run`:
 
 Distributed under the Apache 2 License. See `LICENSE.txt` for more information.
 
-<p align="right">(<a href="#readme-top">back to top</a>)</p>
+<p align="right">(<a href="#readme-top">back to top</a>)</p>

dewy/collections/router.py → dewy/collection/router.py

@@ -3,7 +3,7 @@
 from fastapi import APIRouter, Path
 from loguru import logger
 
-from dewy.collections.models import Collection, CollectionCreate
+from dewy.collection.models import Collection, CollectionCreate
 from dewy.common.collection_embeddings import get_dimensions
 from dewy.common.db import PgConnectionDep
 

dewy/common/collection_embeddings.py

@@ -6,8 +6,8 @@
 from llama_index.schema import TextNode
 from loguru import logger
 
-from dewy.chunks.models import TextResult
-from dewy.collections.models import DistanceMetric
+from dewy.chunk.models import TextResult
+from dewy.collection.models import DistanceMetric
 from dewy.config import settings
 
 from .extract import extract

dewy/config.py

@@ -18,6 +18,9 @@ class Config(BaseSettings):
         env_file_encoding="utf-8",
     )
 
+    SERVE_ADMIN_UI: bool = True
+    """If true, serve the admin UI."""
+
     DB: PostgresDsn
     """The Postgres database to connect to."""
 

dewy/documents/router.py → dewy/document/router.py

@@ -5,7 +5,7 @@
 
 from dewy.common.collection_embeddings import CollectionEmbeddings
 from dewy.common.db import PgConnectionDep, PgPoolDep
-from dewy.documents.models import Document
+from dewy.document.models import Document
 
 from .models import AddDocumentRequest
 

dewy/main.py

@@ -1,9 +1,13 @@
 import contextlib
 from typing import AsyncIterator, TypedDict
+from pathlib import Path
+import os
 
+import uvicorn
 import asyncpg
 from fastapi import FastAPI
 from fastapi.middleware.cors import CORSMiddleware
+from fastapi.staticfiles import StaticFiles
 from loguru import logger
 
 from dewy.common import db
@@ -15,6 +19,10 @@
 class State(TypedDict):
     pg_pool: asyncpg.Pool
 
+# Resolve paths, independent of PWD
+current_file_path = Path(__file__).resolve()
+react_build_path = current_file_path.parent.parent / 'frontend' / 'dist'
+migrations_path = current_file_path.parent.parent / 'migrations'
 
 @contextlib.asynccontextmanager
 async def lifespan(_app: FastAPI) -> AsyncIterator[State]:
@@ -25,7 +33,7 @@ async def lifespan(_app: FastAPI) -> AsyncIterator[State]:
     async with db.create_pool(settings.DB.unicode_string()) as pg_pool:
         if settings.APPLY_MIGRATIONS:
             async with pg_pool.acquire() as conn:
-                await apply_migrations(conn, migration_dir="migrations")
+                await apply_migrations(conn, migration_dir=migrations_path)
 
         logger.info("Created database connection")
         state = {
@@ -54,3 +62,12 @@ async def healthcheck() -> dict[str, str]:
 
 
 app.include_router(api_router)
+
+if settings.SERVE_ADMIN_UI and os.path.isdir(react_build_path):
+    logger.info("Running admin UI at http://localhost/admin")
+    # Serve static files from the React app build directory
+    app.mount("/admin", StaticFiles(directory=str(react_build_path), html=True), name="static")
+
+# Function for running Dewy as a script
+def run(*args):
+   uvicorn.run("dewy.main:app", host="0.0.0.0", port=80)

dewy/routes.py

@@ -1,8 +1,8 @@
 from fastapi import APIRouter
 
-from dewy.chunks.router import router as chunks_router
-from dewy.collections.router import router as collections_router
-from dewy.documents.router import router as documents_router
+from dewy.chunk.router import router as chunks_router
+from dewy.collection.router import router as collections_router
+from dewy.document.router import router as documents_router
 
 api_router = APIRouter(prefix="/api")
 

docker-compose.yml

@@ -46,4 +46,4 @@ volumes:
   llama-cache:
 
 networks:
-  kb-network:
+  kb-network:

frontend/src/Chunk.tsx

@@ -66,4 +66,4 @@ export const ChunkList = () => (
         <ChunkListView />
         <Pagination />
     </ListBase>
-);
+);

frontend/src/Document.tsx

@@ -60,4 +60,4 @@ export const DocumentEdit = () => (
         <TextInput source="url"/>
     </SimpleForm>
     </Edit>
-);
+);

frontend/src/dataProvider.ts

@@ -5,7 +5,7 @@ import data from "./data.json";
 
 export const fakeDataProvider = fakeRestDataProvider(data, true);
 
-const apiUrl = 'http://localhost:8000';
+const apiUrl = 'http://localhost';
 const httpClient = fetchUtils.fetchJson;
 
 

pyproject.toml

@@ -5,6 +5,13 @@ description = "Knowledge base service."
 authors = ["Ben Chambers <ben.chambers@datastax.com>"]
 readme = "README.md"
 license = "Apache-2.0"
+include = [
+    { path = "migrations" },
+    { path = "frontend/dist" },
+]
+
+[tool.poetry.scripts]
+dewy = "dewy.main:run"
 
 [tool.poetry.dependencies]
 python = "^3.11"