mito-ai 0.1.36__tar.gz → 0.1.38__tar.gz

mito_ai-0.1.38/PKG-INFO

@@ -0,0 +1,218 @@
+Metadata-Version: 2.4
+Name: mito_ai
+Version: 0.1.38
+Summary: AI chat for JupyterLab
+Project-URL: Homepage, https://trymito.io
+Project-URL: Bug Tracker, https://github.com/mito-ds/monorepo/issues
+Project-URL: Repository, https://github.com/mito-ds/monorepo
+Author-email: Aaron Diamond-Reivich <aaron@sagacollab.com>
+License: Copyright (c) 2020-2024 Saga Inc.
+        
+        See the LICENSE.txt file at the root of this monorepo for licensing information.
+License-File: LICENSE
+Keywords: jupyter,jupyterlab,jupyterlab-extension
+Classifier: Framework :: Jupyter
+Classifier: Framework :: Jupyter :: JupyterLab
+Classifier: Framework :: Jupyter :: JupyterLab :: 4
+Classifier: Framework :: Jupyter :: JupyterLab :: Extensions
+Classifier: Framework :: Jupyter :: JupyterLab :: Extensions :: Prebuilt
+Classifier: License :: OSI Approved :: BSD License
+Classifier: Programming Language :: Python
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.9
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Requires-Python: >=3.9
+Requires-Dist: analytics-python
+Requires-Dist: anthropic
+Requires-Dist: google-genai
+Requires-Dist: jinja2>=3.0.3
+Requires-Dist: jupyterlab<5,>=4.1.0
+Requires-Dist: openai>=1.0.0
+Requires-Dist: pipreqs
+Requires-Dist: pydantic
+Requires-Dist: requests>=2.25.0
+Requires-Dist: sqlalchemy
+Requires-Dist: tornado>=6.2.0
+Requires-Dist: traitlets
+Provides-Extra: deploy
+Requires-Dist: hatch-jupyter-builder>=0.5; extra == 'deploy'
+Requires-Dist: hatch-nodejs-version>=0.3.2; extra == 'deploy'
+Requires-Dist: hatchling>=1.27.0; extra == 'deploy'
+Requires-Dist: twine>=4.0.0; extra == 'deploy'
+Provides-Extra: test
+Requires-Dist: mypy>=1.8.0; extra == 'test'
+Requires-Dist: pytest-asyncio==0.25.3; extra == 'test'
+Requires-Dist: pytest==8.3.4; extra == 'test'
+Requires-Dist: types-requests>=2.25.0; extra == 'test'
+Requires-Dist: types-setuptools; extra == 'test'
+Requires-Dist: types-tornado>=5.1.1; extra == 'test'
+Description-Content-Type: text/markdown
+
+# mito_ai
+
+[![Github Actions Status](/workflows/Build/badge.svg)](/actions/workflows/build.yml)
+
+AI chat for JupyterLab. This codebase contains two main components:
+1. A Jupyter server extension that handles the backend logic for the chat.
+2. Several JupyterLab extensions that handle the frontend logic for interacting with the AI, including the chat sidebar and the error message rendermime.
+
+## Requirements
+
+- JupyterLab >= 4.0.0
+
+## Install
+
+To install the extension, execute:
+
+```bash
+pip install mito-ai
+```
+
+## Configuration
+
+This extension has two AI providers; OpenAI and Mito (calling OpenAI).
+Mito is the fallback but the number of request is limited for free tier.
+To use OpenAI directly, you will to create an API key on https://platform.openai.com/docs/overview.
+Then set the environment variable `OPENAI_API_KEY` with that key.
+
+The OpenAI model can be configured with 1 parameters:
+- `OpenAIProvider.model`: Name of the AI model; default _gpt-4o-mini_.
+
+You can set those parameters through command line when starting JupyterLab; e.g.
+
+```sh
+jupyter lab --OpenAIProvider.max_completion_tokens 20 --OpenAIProvider.temperature 1.5
+```
+
+> If a value is incorrect, an error message will be displayed in the terminal logs.
+
+## Uninstall
+
+To remove the extension, execute:
+
+```bash
+pip uninstall mito-ai
+```
+
+## Contributing
+
+### Development install
+
+To ensure consistent package management, please use `jlpm` instead of `npm` for this project.
+
+Note: You will need NodeJS to build the extension package.
+
+The `jlpm` command is JupyterLab's pinned version of
+[yarn](https://yarnpkg.com/) that is installed with JupyterLab. 
+
+```bash
+# Clone the repo to your local environment
+# Change directory to the mito-ai directory
+
+# Required to deal with Yarn 3 workspace rules
+touch yarn.lock
+
+# Install package in development mode
+pip install -e ".[test]"
+
+# Install the node modules
+jlpm install
+
+# Build the extension
+jlpm build
+
+# Link your development version of the extension with JupyterLab
+jupyter labextension develop . --overwrite
+
+# Start the jupyter server extension for development
+jupyter server extension enable --py mito_ai
+
+# Watch the source directory in one terminal, automatically rebuilding when needed
+# In case of Error: If this command fails because the lib directory was not created (the error will say something like
+# unable to find main entry point) then run `jlpm run clean:lib` first to get rid of the old buildcache 
+# that might be preventing a new lib directory from getting created. 
+jlpm watch
+```
+
+Then, in a new terminal, run:
+
+```bash
+# Run JupyterLab in another terminal
+jupyter lab --autoreload
+```
+
+With the watch command running, every saved change will immediately be built locally and available in your running JupyterLab. With the `--autoreload` flag, you don't need to refresh JupyterLab to load the change in your browser. It will launch a new window each time you save a change to the backend.
+
+By default, the `jlpm build` command generates the source maps for this extension to make it easier to debug using the browser dev tools. To also generate source maps for the JupyterLab core extensions, you can run the following command:
+
+```bash
+jupyter lab build --minimize=False
+```
+
+### Development uninstall
+
+```bash
+pip uninstall mito-ai
+```
+
+In development mode, you will also need to remove the symlink created by `jupyter labextension develop`
+command. To find its location, you can run `jupyter labextension list` to figure out where the `labextensions`
+folder is located. Then you can remove the symlink named `mito-ai` within that folder.
+
+### Testing the extension
+
+#### Integration tests
+
+Integration tests for mito-ai are written using Playwright and Gelata in the mito/tests directory.
+
+To run these tests, follow the directions in the tests/README.md file.
+
+#### Backend Unit tests
+
+Backend tests for mito-ai are written using pytest in the mito/mito-ai/mito_ai/tests directory.
+
+To run the pytests, just run `pytest` in the mito-ai directory.
+
+#### Backend Mypy tests
+
+To run the mypy tests, just run `mypy mito_ai/ --ignore-missing-imports` in the mito-ai directory.
+
+#### Frontend Unit tests 
+
+Frontend unit tests for mito-ai are written using Jest in the mito/mito-ai/src/tests directory.
+
+To run the Jest tests, just run `npm test` in the mito-ai directory.
+
+#### Frontend Tests
+
+Frontend tests for mito-ai are written using Playwright and Gelata in the mito/tests directory. See the [tests/README.md](tests/README.md) file for more information.
+
+#### Frontend Linting
+
+Frontend linting for mito-ai is done using ESLint in the mito-ai directory.
+
+To run the ESLint tests, just run `jlpm eslint` in the mito-ai directory.
+
+#### Performance Tests
+
+Performance tests for mito-ai are written using pytest in the mito-ai/tests directory.
+
+To run the performance tests, just run `python -m pytest mito_ai/tests/performance_test.py -v -s` in the mito-ai directory.
+
+Note that you'll have to edit `open_ai_utils.py`, specifically the `is_running_test` condition.
+
+#### Running Databases
+
+To ensure reproducibility, databases, like Postgres, are created using Docker. To run:
+
+```bash
+docker-compose -f mito_ai/tests/docker/postgres.yml up
+```
+
+When you're done, stop and remove the container and its volumes with:
+
+```bash
+docker-compose -f mito_ai/tests/docker/postgres.yml down -v
+```

mito_ai-0.1.38/mito_ai/__init__.py

@@ -0,0 +1,84 @@
+# Copyright (c) Saga Inc.
+# Distributed under the terms of the GNU Affero General Public License v3.0 License.
+
+from typing import List, Dict
+from jupyter_server.utils import url_path_join
+from mito_ai.completions.handlers import CompletionHandler
+from mito_ai.completions.providers import OpenAIProvider
+from mito_ai.app_builder.handlers import AppBuilderHandler
+from mito_ai.streamlit_preview.handlers import StreamlitPreviewHandler
+from mito_ai.log.urls import get_log_urls
+from mito_ai.version_check import VersionCheckHandler
+from mito_ai.db.urls import get_db_urls
+from mito_ai.settings.urls import get_settings_urls
+from mito_ai.rules.urls import get_rules_urls
+from mito_ai.auth.urls import get_auth_urls
+from mito_ai.streamlit_preview.urls import get_streamlit_preview_urls
+try:
+    from _version import __version__
+except ImportError:
+    # Fallback when using the package in dev mode without installing in editable mode with pip. It is highly recommended to install
+    # the package from a stable release or in editable mode: https://pip.pypa.io/en/stable/topics/local-project-installs/#editable-installs
+    import warnings
+    
+    warnings.warn("Importing 'mito_ai' outside a proper installation.")
+    __version__ = "dev"
+
+def _jupyter_labextension_paths() -> List[Dict[str, str]]:
+    return [{"src": "labextension", "dest": "mito_ai"}]
+
+
+def _jupyter_server_extension_points() -> List[Dict[str, str]]:
+    """
+    Returns a list of dictionaries with metadata describing
+    where to find the `_load_jupyter_server_extension` function.
+    """
+    return [{"module": "mito_ai"}]
+
+
+# Jupyter Server is the backend used by JupyterLab. A sever extension lets
+# us add new API's to the backend, so we can do some processing that we don't
+# want to exist in the users's javascript.
+# For a further explanation of the Jupyter architecture watch the first 35 minutes
+# of this video: https://www.youtube.com/watch?v=9_-siU-_XoI
+def _load_jupyter_server_extension(server_app) -> None: # type: ignore
+    host_pattern = ".*$"
+    web_app = server_app.web_app
+    base_url = web_app.settings["base_url"]
+
+    open_ai_provider = OpenAIProvider(config=server_app.config)
+
+    # WebSocket handlers
+    handlers = [
+        (
+            url_path_join(base_url, "mito-ai", "completions"),
+            CompletionHandler,
+            {"llm": open_ai_provider},
+        ),
+        (
+            url_path_join(base_url, "mito-ai", "app-builder"),
+            AppBuilderHandler,
+            {}
+        ),
+        (
+            url_path_join(base_url, "mito-ai", "streamlit-preview"),
+            StreamlitPreviewHandler,
+            {}
+        ),
+        (
+            url_path_join(base_url, "mito-ai", "version-check"),
+            VersionCheckHandler,
+            {},
+        )
+    ]
+    
+    # REST API endpoints
+    handlers.extend(get_db_urls(base_url))  # type: ignore
+    handlers.extend(get_settings_urls(base_url))  # type: ignore
+    handlers.extend(get_rules_urls(base_url))  # type: ignore
+    handlers.extend(get_log_urls(base_url, open_ai_provider.key_type))  # type: ignore
+    handlers.extend(get_auth_urls(base_url))  # type: ignore
+    handlers.extend(get_streamlit_preview_urls(base_url))  # type: ignore
+    
+    web_app.add_handlers(host_pattern, handlers)
+    server_app.log.info("Loaded the mito_ai server extension")

mito_ai-0.1.38/mito_ai/_version.py

@@ -0,0 +1,4 @@
+# This file is auto-generated by Hatchling. As such, do not:
+#   - modify
+#   - track in version control e.g. be sure to add to .gitignore
+__version__ = VERSION = '0.1.38'

mito_ai-0.1.38/mito_ai/anthropic_client.py

@@ -0,0 +1,250 @@
+# Copyright (c) Saga Inc.
+# Distributed under the terms of the GNU Affero General Public License v3.0 License.
+
+import json
+import anthropic
+from typing import Dict, Any, Optional, Tuple, Union, Callable, List, cast
+
+from anthropic.types import Message, MessageParam
+from mito_ai.completions.models import CompletionError, ResponseFormatInfo, CompletionReply, CompletionStreamChunk, CompletionItem, MessageType
+from mito_ai.utils.mito_server_utils import ProviderCompletionException
+from openai.types.chat import ChatCompletionMessageParam
+from mito_ai.utils.anthropic_utils import get_anthropic_completion_from_mito_server, stream_anthropic_completion_from_mito_server, get_anthropic_completion_function_params
+
+# Max tokens is a required parameter for the Anthropic API. 
+# We set it to a high number so that we can edit large code cells
+# 8192 is the maximum allowed number of output tokens for claude-3-5-haiku-20241022
+MAX_TOKENS = 8_000
+
+def extract_and_parse_anthropic_json_response(response: Message) -> Union[object, Any]:
+    """
+    Extracts and parses the JSON response from the Claude API.
+    """
+    try:
+        # Check for tool use in the response
+        for content_block in response.content:
+            if content_block.type == "tool_use" and content_block.name == "agent_response":
+                result = content_block.input
+                return result
+
+        # If no tool use was found, try to parse the text response
+        text_response = None
+        for content_block in response.content:
+            if content_block.type == "text":
+                text_response = content_block.text
+                break
+
+        if text_response:
+            # Try to extract JSON from the text response
+            import re
+            json_pattern = r'(\{.*\})'
+            match = re.search(json_pattern, text_response, re.DOTALL)
+            if match:
+                try:
+                    json_response = json.loads(match.group(0))
+                    return json_response
+                except json.JSONDecodeError:
+                    pass
+
+        raise Exception("No valid AgentResponse format found in the response")
+    except Exception as e:
+        raise Exception(f"Failed to parse response: {e}")
+
+
+def get_anthropic_system_prompt_and_messages(messages: List[ChatCompletionMessageParam]) -> Tuple[
+    Union[str, anthropic.NotGiven], List[MessageParam]]:
+    """
+    Convert a list of OpenAI messages to a list of Anthropic messages.
+    """
+    
+    system_prompt: Union[str, anthropic.NotGiven] = anthropic.NotGiven()
+    anthropic_messages: List[MessageParam] = []
+
+    for message in messages:
+        if 'content' not in message:
+            continue
+
+        # We assume that the conversation only has one system message.
+        # Or if there are multiple, we take the last one.
+        if message['role'] == 'system':
+            system_prompt = str(message['content'])
+
+        # Construct the messages for the user and assistant in Anthropic format.
+        if message['role'] == 'user':
+            content = message['content']
+
+            # Handle mixed content (text + images)
+            if isinstance(content, list):
+                anthropic_content = []
+
+                for item in content:
+                    if isinstance(item, dict):
+                        item_dict = cast(Dict[str, Any], item)
+                        if item_dict.get('type') == 'text':
+                            # Add text content
+                            text_content = item_dict.get('text', '')
+                            anthropic_content.append({
+                                "type": "text",
+                                "text": text_content
+                            })
+                        elif item_dict.get('type') == 'image_url':
+                            # Convert OpenAI image format to Anthropic format
+                            image_url_obj = item_dict.get('image_url', {})
+                            if isinstance(image_url_obj, dict):
+                                image_url = image_url_obj.get('url', '')
+                            else:
+                                image_url = str(image_url_obj)
+
+                            # Extract media type and base64 data
+                            if image_url.startswith('data:'):
+                                # Format: data:image/png;base64,<base64_data>
+                                header, base64_data = image_url.split(',', 1)
+                                media_type = header.split(';')[0].split(':')[1]  # Extract image/png or image/jpeg
+                            else:
+                                # If it's not a data URL, assume it's direct base64 and default to image/png
+                                media_type = "image/png"
+                                base64_data = image_url
+
+                            anthropic_content.append({
+                                "type": "image",
+                                "source": {
+                                    "type": "base64",
+                                    "media_type": media_type,
+                                    "data": base64_data
+                                }
+                            })
+
+                anthropic_messages.append(MessageParam(role='user', content=cast(Any, anthropic_content)))
+            else:
+                # Handle simple text content
+                anthropic_messages.append(MessageParam(role='user', content=str(content)))
+
+        elif message['role'] == 'assistant':
+            anthropic_messages.append(MessageParam(role='assistant', content=str(message['content'])))
+
+    return system_prompt, anthropic_messages
+
+
+class AnthropicClient:
+    """
+    A client for interacting with the Anthropic API or the Mito server fallback.
+    """
+
+    def __init__(self, api_key: Optional[str], timeout: int = 30, max_retries: int = 1):
+        self.api_key = api_key
+        self.timeout = timeout
+        self.max_retries = max_retries
+        self.client: Optional[anthropic.Anthropic]
+        if api_key:
+            self.client = anthropic.Anthropic(api_key=api_key)
+        else:
+            self.client = None
+
+    async def request_completions(
+        self, messages: List[ChatCompletionMessageParam],
+        model: str,
+        response_format_info: Optional[ResponseFormatInfo] = None,
+        message_type: MessageType = MessageType.CHAT
+    ) -> Any:
+        """
+        Get a response from Claude or the Mito server that adheres to the AgentResponse format.
+        """
+        anthropic_system_prompt, anthropic_messages = get_anthropic_system_prompt_and_messages(messages)
+        
+        provider_data = get_anthropic_completion_function_params(
+            message_type=message_type,
+            model=model,
+            messages=anthropic_messages,
+            max_tokens=MAX_TOKENS,
+            temperature=0,
+            system=anthropic_system_prompt,
+            stream=None,
+            response_format_info=response_format_info
+        )
+        
+        if self.api_key:
+            # Unpack provider_data for direct API call
+            assert self.client is not None
+            response = self.client.messages.create(**provider_data)
+            if provider_data.get("tool_choice") is not None:
+                result = extract_and_parse_anthropic_json_response(response)
+                return json.dumps(result) if not isinstance(result, str) else result
+            else:
+                content = response.content
+                if content[0].type == "text":
+                    return content[0].text
+                else:
+                    return ""
+        else:
+            # Only pass provider_data to the server
+            response = await get_anthropic_completion_from_mito_server(
+                model=provider_data["model"],
+                max_tokens=provider_data["max_tokens"],
+                temperature=provider_data["temperature"],
+                system=provider_data["system"],
+                messages=provider_data["messages"],
+                tools=provider_data.get("tools"),
+                tool_choice=provider_data.get("tool_choice"),
+                message_type=message_type
+            )
+            return response
+
+    async def stream_completions(self, messages: List[ChatCompletionMessageParam], model: str, message_id: str, message_type: MessageType,
+                              reply_fn: Callable[[Union[CompletionReply, CompletionStreamChunk]], None]) -> str:
+        try:
+            anthropic_system_prompt, anthropic_messages = get_anthropic_system_prompt_and_messages(messages)
+            accumulated_response = ""
+
+            if self.api_key:
+                assert self.client is not None
+                stream = self.client.messages.create(
+                    model=model,
+                    max_tokens=MAX_TOKENS,
+                    temperature=0,
+                    system=anthropic_system_prompt,
+                    messages=anthropic_messages,
+                    stream=True
+                )
+
+
+                for chunk in stream:
+                    if chunk.type == "content_block_delta" and chunk.delta.type == "text_delta":
+                        content = chunk.delta.text
+                        accumulated_response += content
+
+                        is_finished = chunk.type == "message_stop"
+
+                        reply_fn(CompletionStreamChunk(
+                            parent_id=message_id,
+                            chunk=CompletionItem(
+                                content=content,
+                                isIncomplete=not is_finished,
+                                token=message_id,
+                            ),
+                            done=is_finished,
+                        ))
+
+            else:
+                async for stram_chunk in stream_anthropic_completion_from_mito_server(
+                    model=model,
+                    max_tokens=MAX_TOKENS,
+                    temperature=0,
+                    system=anthropic_system_prompt,
+                    messages=anthropic_messages,
+                    stream=True,
+                    message_type=message_type,
+                    reply_fn=reply_fn,
+                    message_id=message_id
+                ):
+                    accumulated_response += stram_chunk
+
+            return accumulated_response
+
+        except anthropic.RateLimitError:
+            raise Exception("Rate limit exceeded. Please try again later or reduce your request frequency.")
+
+        except Exception as e:
+            print(f"Error streaming content: {str(e)}")
+            raise e
+
+