The constructor of VectorStoreRouterAgent expects a non-list object for vectorstoreroutertoolkit. However, in some cases, vectorstoreroutertoolkit is already a list. This commit fixes the issue by checking if vectorstoreroutertoolkit is a list and using it directly if it is.
The cachetools package is added as a dependency to improve caching. In chat.py, a ValueError is raised if no ID is provided in init_build to prevent errors. The flow_data_store dictionary is replaced with an LRUCache from cachetools to limit the size of the cache to 10 items. This improves performance by reducing the memory usage of the application.
📦 chore(pyproject.toml): add cachetools dependency to improve caching
🐛 fix(chat.py): raise ValueError if no ID is provided in init_build
🐛 fix(endpoints.py): add check for flow data before processing
🔨 refactor(process.py): add type hints to process_tweaks function
The get_flow_from_token function and HTTPBearer import are removed as they are not used in the code. A check for flow data is added before processing to avoid errors when the flow data is None. The process_tweaks function is updated to include type hints for the graph_data and tweaks parameters.
🔥 refactor(utils/validate.py): remove try-except block in create_function method
The pass statement in the BaseCache abstract method is redundant and can be removed. Similarly, the try-except block in the create_function method is not necessary as the exception is being suppressed.
The name and display_name of the InitializeAgentNode class have been changed to "AgentInitializer" and "AgentInitializer" respectively. This improves the readability of the code and makes it easier to understand the purpose of the class.
The "initialize_agent" key in the CUSTOM_AGENTS dictionary has been renamed to "AgentInitializer" to improve naming consistency with the other agents in the dictionary.
🔨 refactor(chains.py): comment out unused code block
The `from_method_nodes` dictionary in `base.py` has been updated to include the `LLMCheckerChain` class. This allows the `from_llm` method to be called on the `LLMCheckerChain` class.
The code block in `chains.py` that deals with the `PromptTemplate` field type has been commented out as it is currently unused. This is to prevent confusion and to keep the codebase clean.
The create_app function now accepts a static_path parameter that defaults to "static". The setup_static_files function is created to mount the static files directory to the app. A custom 404 handler is added to the app to return the index.html file when a 404 error occurs. This allows the app to serve static files such as HTML, CSS, and JavaScript files.
🚀 feat(__main__.py, main.py): add support for serving static files
✨ feat(__main__.py): create a setup_static_files function to mount the static files directory to the app
The `save_api_keys` variable has been renamed to `remove_api_keys` to improve the semantics of the code. The new variable name better reflects the functionality of the code, which is to remove API keys from the projects saved in the database.
🔧 chore(__main__.py): change save_api_keys to remove_api_keys to improve semantics
✨ feat(__main__.py): add save_api_keys parameter to serve command to allow users to save API keys in their projects
The update_settings function now accepts a save_api_keys parameter, which allows the user to specify whether or not to save API keys in their projects. The serve command now has a save_api_keys parameter that defaults to True, allowing users to save API keys in their projects. This feature improves the user experience by allowing them to save API keys for future use.
This commit adds a new function called `remove_api_keys` to the `utils.py` file. The function takes in a dictionary representing a flow and removes any API keys from the flow data. The function iterates through each node in the flow and checks if the node contains any API keys. If an API key is found, the function sets the value of the key to `None`. This function is useful for removing sensitive information from flow data before it is stored or transmitted.
The `remove_api_keys` function is now called on the `flow_data` dictionary if the `save_api_keys` setting is False. This ensures that sensitive information is not saved in the database.
The condition to show the password field was not working for the 'tokens' field name. The condition has been updated to include 'tokens' in the field name and show the password field.
🔨 refactor(parameterComponent): remove unused imports and refactor onChange function to handleOnNewValue
✨ feat(tabsContext): add tabsState and setTabsState to TabsContextType and TabsProvider
🔨 refactor(flowSettingsModal): refactor handleSaveFlow function to update flow and setTabsState with isPending false
The update_flow endpoint now returns a FlowRead response model instead of FlowReadWithStyle. The parameterComponent file has been refactored to remove unused imports and to use a handleOnNewValue function to handle onChange events. The TabsContextType and TabsProvider have been updated to include tabsState and setTabsState. The flowSettingsModal has been refactored to update the flow and setTabsState with isPending false.
🔨 refactor(extraSidebarComponent): add tabsState and setTabsState to TabsContextType
🐛 fix(extraSidebarComponent): disable save button when flow is not pending
🐛 fix(extraSidebarComponent): update flow state after saving
The TabsContextType now includes tabsState and setTabsState to allow for the management of the state of each tab. The save button is now disabled when the flow is not pending. The flow state is now updated after saving to reflect the changes made.
The description of CombineDocsChainNode was updated to reflect the actual functionality of the node. The node is now used to load a question answering chain instead of constructing a chain from combined documents.
The openai_api_version field value is no longer set as it is not needed. The password field is now set to False for fields not containing 'key' or 'token' in their name to improve security.
🐛 fix(llms.py): remove setting of openai_api_version field value
🔧 chore(base.py): fix typo in docstring
The fix in chat.py ensures that an error message is yielded when an exception occurs during stream_build. This helps to provide more information to the client-side when an error occurs. The typo in base.py's docstring is fixed to improve readability.
- Rename `VectorStoreAgent` to `Vector store agent` in `custom.py`
- Replace "Construct a sql agent from an LLM and tools." with "Construct an SQL agent from an LLM and tools." in `custom.py`
- Update descriptions of `SQLAgentNode`, `TimeTravelGuideChainNode`, `CombineDocsChainNode`, and `ToolNode` in related files
- Increase width of `ExtraSidebarComponent` container in `index.tsx` from 52 to 56
The build_template function is now outside of the DocumentLoaderFrontNode class to improve code organization and make it more modular. This change also makes it easier to reuse the function in other parts of the codebase.
Added new OpenAI chat models to the CHAT_OPENAI_MODELS list to improve the quality of the chatbot responses. The new models are "gpt-3.5-turbo-0613", "gpt-3.5-turbo-16k-0613", "gpt-4-0613", and "gpt-4-32k-0613". The existing models "gpt-3.5-turbo" and "gpt-4-32k" were also updated to their latest versions.
The status codes have been added to the API endpoints to improve the readability of the code and to make it more explicit what the expected response codes are. The status codes have been added to the following endpoints: create_flow, read_flows, read_flow, update_flow, delete_flow, create_flows, upload_file, and download_file.
✨ feat(endpoints.py): add flow_id parameter to predict_flow function to allow for running a flow by ID
The predict_flow function now includes a flow_id parameter and a session dependency to allow for running a flow by ID. The flow object is retrieved from the session using the flow_id parameter. If the flow is not found, a ValueError is raised.
🔨 refactor(constants.tsx): change API_URL constant to BASE_API_URL and add flow_id parameter to run_flow function
The API_URL constant has been renamed to BASE_API_URL to better reflect its purpose. The run_flow function now includes a flow_id parameter to allow for running a flow by ID. The flow_id parameter is used to construct the API URL.
The code now handles exceptions that may occur while building the flow. A final response is added to the event stream to indicate the end of the stream.
The try-except block was added to handle exceptions that may occur in the init_build function. If an exception occurs, it is logged and an HTTPException with a status code of 500 is returned. This ensures that the server does not crash and provides a more informative error message to the client.
The init_build endpoint now returns a status code of 201 to indicate that a new resource has been created. This improves the consistency of the API and makes it easier for clients to understand the response.
🚀 feat(chat.py): add response models to /build/init and /build/{flow_id}/status endpoints
The InitResponse and BuiltResponse models are now imported from the schemas module to improve code organization. The /build/init and /build/{flow_id}/status endpoints now have response models to provide a clear understanding of the expected response.
The docstring for the CombineDocsChain class has been updated to reflect the correct function name. In the loading.py file, the instantiate_vectorstore function has been updated to ensure that metadata values are not None for the Chroma class. This is because Chroma requires all metadata values to not be None, and this fix ensures that the application will not encounter errors when using Chroma.
🐛 fix(custom.py): update docstring to reflect the correct function name
The process_tweaks function has been refactored to improve readability and maintainability. The function now takes in two parameters, graph_data and tweaks, and returns the modified graph_data. The tweaks parameter is a dictionary of dictionaries, where the key is the node id and the value is a dictionary of the tweaks. The function processes the graph data to add the tweaks by iterating over the nodes and checking if the node id is in the tweaks dictionary. If it is, the function applies the tweaks to the node by updating the template data with the new values. The function also prints a message to the console to indicate that a tweak has been applied.
The PredictRequest schema now includes an optional tweaks field, which is a dictionary of dictionaries. This field allows for additional customization of the prediction request, such as specifying tool names or descriptions. The tweaks field is optional, and if not provided, the default value is an empty dictionary. The schema_extra attribute has also been updated to include an example of the new tweaks field.
🐛 fix(endpoints.py): change predict endpoint to use Flow object instead of flow_id
🐛 fix(endpoints.py): add support for processing tweaks in predict endpoint
The predict endpoint now requires authentication using HTTPBearer. The flow_id is now extracted from the bearer token instead of being passed as a parameter. This improves security as the flow_id is not exposed in the URL. The predict endpoint now uses the Flow object instead of the flow_id to retrieve the graph data. This improves code readability and reduces the number of database queries. The predict endpoint now supports processing tweaks, which allows for more flexibility in the processing of messages.
The function stream_build was returning a JSONResponse with an error message when an exception was raised. This is not the correct way to handle errors in FastAPI. Instead, we should raise an HTTPException with the appropriate status code and error message.
The unused imports and exception handling for WebSocketDisconnect and WebSocketException were removed from the chat.py file. The code was reformatted to improve readability.
🔥 refactor(chat.py): remove unused imports and exception handling
✨ feat(chat.py): add build_status endpoint to check if flow_id is in flow_data_store
The function name websocket_endpoint was renamed to chat to improve semantics. The new build_status endpoint was added to check if a flow_id is in the flow_data_store. This endpoint returns a JSON response with a boolean value indicating whether the flow_id is built or not. If the flow_id is not in the flow_data_store, a JSON response with a value of False is returned. If an exception occurs, a 500 HTTPException is raised with the exception message as the detail.
The build method call in the for loop of the build_vertices method is unnecessary as the vertices are already sorted and built in the topological_sort method. The yield from statement is used to return the sorted vertices.
The changes in this commit handle exceptions that may occur when creating a streaming response. The try-except block ensures that any exceptions are caught and an appropriate response is returned. Additionally, the SSE response has been improved by adding a valid flag and node id to the response. This provides more information to the client about the status of the node build and allows for better error handling.
🐛 fix(chat.py): handle exceptions when creating a streaming response
The commented out code for the /build/{client_id} endpoint has been removed as it is no longer needed. The new implementation uses the /build/init endpoint to initiate the build process and then establishes an SSE connection using EventSource to stream the build process. This allows for a more efficient and responsive build process as the client can receive updates in real-time.
🔥 chore(chat.py, index.tsx): remove commented out code for /build/{client_id} endpoint
✨ feat(chat.py): add support for storing graph data and returning a unique session ID for building langchain object
✨ feat(chat.py): add support for streaming the build process based on stored flow data
The fix adds a check for the client_id in the in_memory_cache before handling the websocket. This ensures that the flow has been built before sending messages.
The first feature adds support for storing graph data and returning a unique session ID for building the langchain object. This allows the user to build the flow and then send messages.
The second feature adds support for streaming the build process based on stored flow data. This allows the user to see the progress of the build process.
