Feature/better usability (#6)

* refactor: api, core

* feat: static uploader

* doc: update readme

* fix: mkdir static file

* doc: typo

.env.example

@@ -1,10 +1,4 @@
 BOT_NAME=<your-bot-name>
-AWS_ACCESS_KEY_ID=***
-AWS_SECRET_ACCESS_KEY=***
-AWS_REGION=***
-AWS_S3_BUCKET=***
-WINEDB_HOST=***
-WINEDB_PASSWORD=***
 OPENAI_API_KEY=***
 BING_SEARCH_URL=***
 BING_SUBSCRIPTION_KEY=***

.gitignore

@@ -6,4 +6,6 @@ __pycache__/
 image/
 audio/
 video/
-dataframe/
+dataframe/
+
+static/*

Dockerfile

@@ -20,4 +20,6 @@ RUN poetry install --with tools
 
 COPY . .
 
-ENTRYPOINT ["poetry", "run", "python3", "-m", "uvicorn", "main:app", "--host=0.0.0.0", "--port=8000"]
+ENV PORT 8000
+
+ENTRYPOINT ["poetry", "run", "serve"]

README.md

@@ -43,52 +43,37 @@ We also don't know what tools EVAL will create. Every day, It will create the ri
 
 ## Usage
 
-1. S3 Settings
-2. environments settings
-3. `docker-compose up -d`
-
-### S3
-
-1. Create a bucket.
-2. Turn off the "Block all public access" setting for the bucket. ![image](assets/block_public_access.png)
-3. Add the following text to Bucket Policy.
-   ```json
-   {
-     "Version": "2012-10-17",
-     "Statement": [
-       {
-         "Sid": "AllowPublicRead",
-         "Effect": "Allow",
-         "Principal": {
-           "AWS": "*"
-         },
-         "Action": "s3:GetObject",
-         "Resource": "arn:aws:s3:::{your-bucket-name}/*"
-       }
-     ]
-   }
-   ```
+1. environments settings
+2. `docker-compose up -d`
 
 ### Environment
 
-These environmental variables are essential, so please set them.
+You need to write some environment variables in the `.env` file. Refer [.env.example](.env.example) if you don't know how to format it.
 
-```
-BOT_NAME: your custom bot name
-OPENAI_API_KEY: openai api key
-AWS_ACCESS_KEY_ID
-AWS_SECRET_ACCESS_KEY
-AWS_REGION
-AWS_S3_BUCKET
-```
+**Mandatory**
 
-These environment variables are necessary to use the following tools:  
-If you want to use it, set it up, and if you don't need it, you don't have to set it up.
+Manatory envs are required in order to serve EVAL.
 
-```
-SERPAPI_API_KEY: need to append google search tool
-BING_SEARCH_URL, BING_SUBSCRIPTION_KEY: need to append bing search tool
-```
+- `OPENAI_API_KEY` - OpenAI api key
+
+**Optional**
+
+Each optional env has default value, so you don't need to set unless you want to change it.
+
+- `PORT` - port (default: 8000)
+- `SERVER` - server address (default: http://localhost:8000)
+- `LOG_LEVEL` - INFO | DEBUG (default: INFO)
+- `BOT_NAME` - give it a name! (default: Orca)
+
+**For More Tools**
+
+Some tools requires environment variables. Set envs depend on which tools you want to use.
+
+- Google search tool
+  - `SERPAPI_API_KEY`
+- Bing search tool
+  - `BING_SEARCH_URL`
+  - `BING_SUBSCRIPTION_KEY`
 
 ## TODO
 

api/main.py

@@ -1,35 +1,41 @@
 from typing import Dict, List, TypedDict
 import re
+import uvicorn
 
 from fastapi import FastAPI
+from fastapi.staticfiles import StaticFiles
+
 from pydantic import BaseModel
 
-from s3 import upload
 from env import settings
 
-from prompts.error import ERROR_PROMPT
-from agents.manager import AgentManager
-from tools.base import BaseToolSet
-from tools.cpu import (
+from core.prompts.error import ERROR_PROMPT
+from core.agents.manager import AgentManager
+from core.tools.base import BaseToolSet
+from core.tools.cpu import (
     Terminal,
     CodeEditor,
     RequestsGet,
     WineDB,
     ExitConversation,
 )
-from tools.gpu import (
+from core.tools.gpu import (
     ImageEditing,
     InstructPix2Pix,
     Text2Image,
     VisualQuestionAnswering,
 )
-from handlers.base import BaseHandler, FileHandler, FileType
-from handlers.image import ImageCaptioning
-from handlers.dataframe import CsvToDataframe
+from core.handlers.base import BaseHandler, FileHandler, FileType
+from core.handlers.image import ImageCaptioning
+from core.handlers.dataframe import CsvToDataframe
+from core.upload import StaticUploader
+
 from logger import logger
 
 app = FastAPI()
 
+app.mount("/static", StaticFiles(directory=StaticUploader.STATIC_DIR), name="static")
+uploader = StaticUploader.from_settings(settings)
 
 toolsets: List[BaseToolSet] = [
     Terminal(),
@@ -104,6 +110,10 @@ async def command(request: Request) -> Response:
 
     return {
         "response": res["output"],
-        "files": [upload(image) for image in images]
-        + [upload(dataframe) for dataframe in dataframes],
+        "files": [uploader.upload(image) for image in images]
+        + [uploader.upload(dataframe) for dataframe in dataframes],
     }
+
+
+def serve():
+    uvicorn.run("api.main:app", host="0.0.0.0", port=settings["PORT"])

core/agents/builder.py

@@ -1,15 +1,15 @@
 from langchain.chat_models.base import BaseChatModel
 from langchain.output_parsers.base import BaseOutputParser
 
-from prompts.input import EVAL_PREFIX, EVAL_SUFFIX
 from env import settings
 
-from tools.base import BaseToolSet
-from tools.factory import ToolsFactory
+from core.prompts.input import EVAL_PREFIX, EVAL_SUFFIX
+from core.tools.base import BaseToolSet
+from core.tools.factory import ToolsFactory
 
-from agents.llm import ChatOpenAI
-from agents.chat_agent import ConversationalChatAgent
-from agents.parser import EvalOutputParser
+from .llm import ChatOpenAI
+from .chat_agent import ConversationalChatAgent
+from .parser import EvalOutputParser
 
 
 class AgentBuilder:

core/agents/chat_agent.py

@@ -20,7 +20,7 @@ from langchain.schema import (
 )
 from langchain.tools.base import BaseTool
 
-from prompts.input import EVAL_TOOL_RESPONSE
+from core.prompts.input import EVAL_TOOL_RESPONSE
 
 
 class ConversationalChatAgent(Agent):

core/agents/manager.py

@@ -1,14 +1,14 @@
-from typing import Dict, Any
+from typing import Dict
 
 from langchain.agents.tools import BaseTool
 from langchain.agents.agent import Agent, AgentExecutor
 from langchain.chains.conversation.memory import ConversationBufferMemory
 from langchain.memory.chat_memory import BaseChatMemory
 
-from tools.base import BaseToolSet
-from tools.factory import ToolsFactory
+from core.tools.base import BaseToolSet
+from core.tools.factory import ToolsFactory
 
-from agents.builder import AgentBuilder
+from .builder import AgentBuilder
 
 
 class AgentManager:

core/agents/parser.py

@@ -3,7 +3,7 @@ from typing import Dict
 
 from langchain.output_parsers.base import BaseOutputParser
 
-from prompts.input import EVAL_FORMAT_INSTRUCTIONS
+from core.prompts.input import EVAL_FORMAT_INSTRUCTIONS
 
 
 class EvalOutputParser(BaseOutputParser):

core/handlers/dataframe.py

@@ -1,8 +1,8 @@
 import pandas as pd
 
-from prompts.file import DATAFRAME_PROMPT
+from core.prompts.file import DATAFRAME_PROMPT
 
-from handlers.base import BaseHandler
+from .base import BaseHandler
 
 
 class CsvToDataframe(BaseHandler):

core/handlers/image.py

@@ -5,9 +5,9 @@ from transformers import (
     BlipForConditionalGeneration,
 )
 
-from prompts.file import IMAGE_PROMPT
+from core.prompts.file import IMAGE_PROMPT
 
-from handlers.base import BaseHandler
+from .base import BaseHandler
 
 
 class ImageCaptioning(BaseHandler):

core/tools/base.py

@@ -1,4 +1,4 @@
-from typing import Optional, Callable, Tuple
+from typing import Callable, Tuple
 from enum import Enum
 
 from langchain.agents.tools import Tool, BaseTool

core/tools/cpu.py

@@ -9,7 +9,7 @@ from bs4 import BeautifulSoup
 
 import subprocess
 
-from tools.base import tool, BaseToolSet, ToolScope, SessionGetter
+from .base import tool, BaseToolSet, ToolScope, SessionGetter
 from logger import logger
 
 

core/tools/factory.py

@@ -3,7 +3,7 @@ from langchain.agents import load_tools
 from langchain.agents.tools import BaseTool
 from langchain.llms.base import BaseLLM
 
-from tools.base import BaseToolSet, SessionGetter
+from .base import BaseToolSet, SessionGetter
 
 
 class ToolsFactory:

core/tools/gpu.py

@@ -23,7 +23,7 @@ from diffusers import EulerAncestralDiscreteScheduler
 from utils import get_new_image_name
 from logger import logger
 
-from tools.base import tool, BaseToolSet
+from .base import tool, BaseToolSet
 
 
 class MaskFormer(BaseToolSet):

core/upload/__init__.py

@@ -0,0 +1,2 @@
+from .s3 import S3Uploader
+from .static import StaticUploader

core/upload/base.py

@@ -0,0 +1,15 @@
+from abc import ABC, abstractmethod, abstractstaticmethod
+
+from env import DotEnv
+
+STATIC_DIR = "static"
+
+
+class AbstractUploader(ABC):
+    @abstractmethod
+    def upload(self, filepath: str) -> str:
+        pass
+
+    @abstractstaticmethod
+    def from_settings(settings: DotEnv) -> "AbstractUploader":
+        pass

core/upload/s3.py

@@ -0,0 +1,35 @@
+import os
+import boto3
+
+from env import DotEnv
+from .base import AbstractUploader
+
+
+class S3Uploader(AbstractUploader):
+    def __init__(self, accessKey: str, secretKey: str, region: str, bucket: str):
+        self.accessKey = accessKey
+        self.secretKey = secretKey
+        self.region = region
+        self.bucket = bucket
+        self.client = boto3.client(
+            "s3",
+            aws_access_key_id=self.accessKey,
+            aws_secret_access_key=self.secretKey,
+        )
+
+    @staticmethod
+    def from_settings(settings: DotEnv) -> "S3Uploader":
+        return S3Uploader(
+            settings["AWS_ACCESS_KEY_ID"],
+            settings["AWS_SECRET_ACCESS_KEY"],
+            settings["AWS_REGION"],
+            settings["AWS_S3_BUCKET"],
+        )
+
+    def get_url(self, object_name: str) -> str:
+        return f"https://{self.bucket}.s3.{self.region}.amazonaws.com/{object_name}"
+
+    def upload(self, filepath: str) -> str:
+        object_name = os.path.basename(filepath)
+        self.client.upload_file(filepath, self.bucket, object_name)
+        return self.get_url(object_name)

core/upload/static.py

@@ -0,0 +1,25 @@
+import os
+import shutil
+
+from env import DotEnv
+from .base import AbstractUploader
+
+
+class StaticUploader(AbstractUploader):
+    STATIC_DIR = "static"
+
+    def __init__(self, server: str):
+        self.server = server
+
+    @staticmethod
+    def from_settings(settings: DotEnv) -> "StaticUploader":
+        return StaticUploader(settings["SERVER"])
+
+    def get_url(self, uploaded_path: str) -> str:
+        return f"{self.server}/{uploaded_path}"
+
+    def upload(self, filepath: str):
+        upload_path = os.path.join(StaticUploader.STATIC_DIR, filepath)
+        os.makedirs(os.path.dirname(upload_path), exist_ok=True)
+        shutil.copy(filepath, upload_path)
+        return f"{self.server}/{upload_path}"

docker-compose.yml

@@ -9,6 +9,7 @@ services:
       context: .
     volumes: # if you want to decrease your model download time, use this.
       - ../.cache/huggingface/:/root/.cache/huggingface/
+      - ./static/:/app/static/
     ports:
       - "8000:8000"
     env_file:

env.py

@@ -7,30 +7,33 @@ load_dotenv()
 
 
 class DotEnv(TypedDict):
+    OPENAI_API_KEY: str
+
+    PORT: int
+    SERVER: str
+
     LOG_LEVEL: str  # optional
-    BOT_NAME: str
-    AWS_ACCESS_KEY_ID: str
-    AWS_SECRET_ACCESS_KEY: str
-    AWS_REGION: str
-    AWS_S3_BUCKET: str
+    BOT_NAME: str  # optional
+    AWS_ACCESS_KEY_ID: str  # optional
+    AWS_SECRET_ACCESS_KEY: str  # optional
+    AWS_REGION: str  # optional
+    AWS_S3_BUCKET: str  # optional
     WINEDB_HOST: str  # optional
     WINEDB_PASSWORD: str  # optional
-    OPENAI_API_KEY: str
     BING_SEARCH_URL: str  # optional
     BING_SUBSCRIPTION_KEY: str  # optional
     SERPAPI_API_KEY: str  # optional
 
 
+PORT = int(os.getenv("PORT", 8000))
 settings: DotEnv = {
+    "PORT": PORT,
+    "SERVER": os.getenv("SERVER", f"http://localhost:{PORT}"),
+    "OPENAI_API_KEY": os.getenv("OPENAI_API_KEY"),
     "LOG_LEVEL": os.getenv("LOG_LEVEL", "INFO"),
     "BOT_NAME": os.getenv("BOT_NAME", "Orca"),
-    "AWS_ACCESS_KEY_ID": os.getenv("AWS_ACCESS_KEY_ID"),
-    "AWS_SECRET_ACCESS_KEY": os.getenv("AWS_SECRET_ACCESS_KEY"),
-    "AWS_REGION": os.getenv("AWS_REGION"),
-    "AWS_S3_BUCKET": os.getenv("AWS_S3_BUCKET"),
     "WINEDB_HOST": os.getenv("WINEDB_HOST"),
     "WINEDB_PASSWORD": os.getenv("WINEDB_PASSWORD"),
-    "OPENAI_API_KEY": os.getenv("OPENAI_API_KEY"),
     "BING_SEARCH_URL": os.getenv("BING_SEARCH_URL"),
     "BING_SUBSCRIPTION_KEY": os.getenv("BING_SUBSCRIPTION_KEY"),
     "SERPAPI_API_KEY": os.getenv("SERPAPI_API_KEY"),

pyproject.toml

@@ -4,7 +4,10 @@ version = "0.1.0"
 description = ""
 authors = ["Taeho Lee <taeho@corca.ai>", "Chung Hwan Han <hanch@corca.ai>"]
 readme = "README.md"
+packages = [{include = "api"},{include = "core"}]
 
+[tool.poetry.scripts]
+serve = "api.main:serve"
 
 [tool.poetry.dependencies]
 python = "^3.10"

s3.py

@@ -1,22 +0,0 @@
-import os
-import boto3
-
-from env import settings
-
-
-def upload(file_name: str):
-    return upload_file(file_name, settings["AWS_S3_BUCKET"])
-
-
-def upload_file(file_name, bucket, object_name=None):
-    if object_name is None:
-        object_name = os.path.basename(file_name)
-
-    s3_client = boto3.client(
-        "s3",
-        aws_access_key_id=settings["AWS_ACCESS_KEY_ID"],
-        aws_secret_access_key=settings["AWS_SECRET_ACCESS_KEY"],
-    )
-    s3_client.upload_file(file_name, bucket, object_name)
-
-    return f"https://{bucket}.s3.{settings['AWS_REGION']}.amazonaws.com/{object_name}"