make visualwebarena package

README.md

@@ -27,17 +27,34 @@
 - [01/25/2024]: GitHub repo released with tasks and scripts for setting up the VWA environments.
 
 ## Install
+
+**Option 1: pip install from GitHub (Recommended)**
 ```bash
 # Python 3.10 (or 3.11, but not 3.12 cause 3.12 deprecated distutils needed here)
-python -m venv venv
-source venv/bin/activate
-pip install -r requirements.txt
+pip install git+https://github.com/web-arena-x/visualwebarena.git
 playwright install
-pip install -e .
 ```
 
-You can also run the unit tests to ensure that VisualWebArena is installed correctly:
+**Option 2: Install from Source**
+```bash
+# Python 3.10 (or 3.11, but not 3.12 cause 3.12 deprecated distutils needed here)
+git clone https://github.com/web-arena-x/visualwebarena.git
+cd visualwebarena
+pip install .
+playwright install
 ```
+
+**Option 3: Development Installation**
+```bash
+# Python 3.10 (or 3.11, but not 3.12 cause 3.12 deprecated distutils needed here)
+git clone https://github.com/web-arena-x/visualwebarena.git
+cd visualwebarena
+pip install -e .
+playwright install
+```
+
+You can run the unit tests to ensure that VisualWebArena is installed correctly:
+```bash
 pytest -x
 ```
 

pyproject.toml

@@ -0,0 +1,72 @@
+[build-system]
+requires = ["setuptools>=64", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "visualwebarena"
+version = "0.1.0"
+description = "Evaluating Multimodal Agents on Realistic Visual Web Tasks"
+readme = "README.md"
+license = {text = "MIT"}
+authors = [
+    {name = "VisualWebArena Team"},
+]
+requires-python = ">=3.7"
+classifiers = [
+    "Development Status :: 3 - Alpha",
+    "Intended Audience :: Researchers",
+    "License :: OSI Approved :: MIT License",
+    "Operating System :: OS Independent",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.7",
+    "Programming Language :: Python :: 3.8",
+    "Programming Language :: Python :: 3.9",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Topic :: Scientific/Engineering :: Artificial Intelligence",
+]
+dependencies = [
+    "beartype>=0.12.0",
+    "beautifulsoup4>=4.12.2",
+    "gymnasium>=0.29.1",
+    "numpy>=1.25.2",
+    "pillow>=10.0.1",
+    "playwright>=1.37.0",
+    "pydantic>=2.4.2",
+    "requests>=2.31.0",
+    "openai>=1.3.5",
+    "torch>=2.0.1",
+    "transformers>=4.34.0",
+    "nltk>=3.8.1",
+    "scikit-image>=0.22.0",
+    "tiktoken>=0.4.0",
+    "matplotlib>=3.8.0",
+    "text-generation>=0.6.1",
+    "aiolimiter>=1.1.0",
+    "evaluate>=0.4.0",
+]
+
+[project.optional-dependencies]
+dev = [
+    "pre-commit==3.0.1",
+    "pytest==7.1.2",
+    "mypy==0.991",
+    "nbmake",
+    "pytest-asyncio",
+    "types-requests",
+]
+
+[project.urls]
+Homepage = "https://jykoh.com/vwa"
+Repository = "https://github.com/web-arena-x/visualwebarena"
+Issues = "https://github.com/web-arena-x/visualwebarena/issues"
+
+[tool.setuptools.packages.find]
+include = ["visualwebarena*"]
+
+[tool.pytest.ini_options]
+testpaths = ["tests"]
+python_files = "test_*.py"
+
+[tool.mypy]
+strict = true

run.py

@@ -19,26 +19,26 @@
 import torch
 from PIL import Image
 
-from agent import (
+from visualwebarena.agent import (
     PromptAgent,
     construct_agent,
 )
-from agent.prompts import *
-from browser_env import (
+from visualwebarena.agent.prompts import *
+from visualwebarena.browser_env import (
     Action,
     ActionTypes,
     ScriptBrowserEnv,
     StateInfo,
     Trajectory,
     create_stop_action,
 )
-from browser_env.actions import is_equivalent
-from browser_env.auto_login import get_site_comb_from_filepath
-from browser_env.helper_functions import (
+from visualwebarena.browser_env.actions import is_equivalent
+from visualwebarena.browser_env.auto_login import get_site_comb_from_filepath
+from visualwebarena.browser_env.helper_functions import (
     RenderHelper,
     get_action_description,
 )
-from evaluation_harness import evaluator_router, image_utils
+from visualwebarena.evaluation_harness import evaluator_router, image_utils
 
 DATASET = os.environ["DATASET"]
 
@@ -469,7 +469,7 @@ def test(
 
 def prepare(args: argparse.Namespace) -> None:
     # convert prompt python files to json
-    from agent.prompts import to_json
+    from visualwebarena.agent.prompts import to_json
 
     to_json.run()
 

run_demo.py

@@ -17,25 +17,25 @@
 from beartype import beartype
 from PIL import Image
 
-from agent import (
+from visualwebarena.agent import (
     PromptAgent,
     construct_agent,
 )
-from agent.prompts import *
-from browser_env import (
+from visualwebarena.agent.prompts import *
+from visualwebarena.browser_env import (
     Action,
     ActionTypes,
     ScriptBrowserEnv,
     StateInfo,
     Trajectory,
     create_stop_action,
 )
-from browser_env.actions import is_equivalent
-from browser_env.helper_functions import (
+from visualwebarena.browser_env.actions import is_equivalent
+from visualwebarena.browser_env.helper_functions import (
     RenderHelper,
     get_action_description,
 )
-from evaluation_harness import image_utils
+from visualwebarena.evaluation_harness import image_utils
 
 LOG_FOLDER = "log_files"
 Path(LOG_FOLDER).mkdir(parents=True, exist_ok=True)
@@ -395,7 +395,7 @@ def test(
 
 def prepare(args: argparse.Namespace) -> None:
     # convert prompt python files to json
-    from agent.prompts import to_json
+    from visualwebarena.agent.prompts import to_json
 
     to_json.run()
 

scripts/collect_obs.py

@@ -8,14 +8,14 @@
 import pytest
 from playwright.sync_api import Page, expect
 import browser_env
-from browser_env import (
+from visualwebarena.browser_env import (
     ScriptBrowserEnv,
     create_id_based_action,
     create_key_press_action,
     create_playwright_action,
     create_scroll_action,
 )
-from browser_env.env_config import *
+from visualwebarena.browser_env.env_config import *
 
 HEADLESS = False
 

scripts/generate_test_data.py

@@ -3,7 +3,7 @@
 import json
 import os
 
-from browser_env.env_config import *
+from visualwebarena.browser_env.env_config import *
 
 
 def main() -> None:

tests/conftest.py

@@ -3,7 +3,7 @@
 import pytest
 import pytest_asyncio
 
-from browser_env import AsyncScriptBrowserEnv, ScriptBrowserEnv
+from visualwebarena.browser_env import AsyncScriptBrowserEnv, ScriptBrowserEnv
 
 HEADLESS = True
 SLOW_MO = 0

tests/test_browser_env/test_action_functionalities.py

@@ -4,7 +4,7 @@
 import pytest
 from playwright.sync_api import Page, expect
 
-from browser_env import (
+from visualwebarena.browser_env import (
     ScriptBrowserEnv,
     create_id_based_action,
     create_key_press_action,

tests/test_browser_env/test_actions.py

@@ -1,6 +1,6 @@
 import numpy as np
 
-from browser_env import *
+from visualwebarena.browser_env import *
 
 
 def test_is_equivalent() -> None:

tests/test_browser_env/test_auth_cookie.py

@@ -1,7 +1,7 @@
 import asyncio
 import json
 
-from browser_env import *
+from visualwebarena.browser_env import *
 
 auth_json = {
     "cookies": [

tests/test_browser_env/test_playwright_actions.py

@@ -3,7 +3,7 @@
 import pytest
 from playwright.sync_api import Page
 
-from browser_env import ScriptBrowserEnv, create_playwright_action
+from visualwebarena.browser_env import ScriptBrowserEnv, create_playwright_action
 
 HEADLESS = True
 SLOW_MO = 0

tests/test_browser_env/test_script_browser_env.py

@@ -9,7 +9,7 @@
 from gymnasium.vector import AsyncVectorEnv
 from playwright.sync_api import Page
 
-from browser_env import (
+from visualwebarena.browser_env import (
     Action,
     AsyncScriptBrowserEnv,
     DetachedPage,
@@ -20,8 +20,8 @@
     create_playwright_action,
     create_scroll_action,
 )
-from browser_env.actions import create_id_based_action
-from browser_env.env_config import ACCOUNTS, REDDIT, SHOPPING
+from visualwebarena.browser_env.actions import create_id_based_action
+from visualwebarena.browser_env.env_config import ACCOUNTS, REDDIT, SHOPPING
 
 @pytest.mark.skip(reason="The actions are deprecated")
 def test_script_browser_env(script_browser_env: ScriptBrowserEnv) -> None:

tests/test_evaluation_harness/test_exact_evaluators.py

@@ -10,17 +10,17 @@
 from PIL import Image
 from py import test
 
-from agent import Agent, TeacherForcingAgent
-from browser_env import ActionTypes, ScriptBrowserEnv
-from browser_env.env_config import *
-from evaluation_harness import (
+from visualwebarena.agent import Agent, TeacherForcingAgent
+from visualwebarena.browser_env import ActionTypes, ScriptBrowserEnv
+from visualwebarena.browser_env.env_config import *
+from visualwebarena.evaluation_harness import (
     HTMLContentExactEvaluator,
     PageImageEvaluator,
     StringEvaluator,
     URLExactEvaluator,
     image_utils,
 )
-from evaluation_harness.evaluators import EvaluatorComb
+from visualwebarena.evaluation_harness.evaluators import EvaluatorComb
 
 IN_GITHUB_ACTIONS = os.getenv("GITHUB_ACTIONS") == "true"
 HEADLESS = True

tests/test_evaluation_harness/test_helper_functions.py

@@ -1,9 +1,9 @@
 import json
 import os
 
-from browser_env import ScriptBrowserEnv
-from browser_env.env_config import *
-from evaluation_harness.helper_functions import (
+from visualwebarena.browser_env import ScriptBrowserEnv
+from visualwebarena.browser_env.env_config import *
+from visualwebarena.evaluation_harness.helper_functions import (
     get_query_text,
     get_query_text_lowercase,
     reddit_get_latest_comment_content_by_username,

visualwebarena/__init__.py

@@ -0,0 +1,13 @@
+"""
+VisualWebArena: Evaluating Multimodal Agents on Realistic Visual Web Tasks
+
+This package provides components for evaluating multimodal autonomous language agents
+on web-based visual tasks.
+"""
+
+__version__ = "0.1.0"
+
+from .browser_env import ScriptBrowserEnv
+from .agent import PromptAgent
+
+__all__ = ["ScriptBrowserEnv", "PromptAgent"]