Refactor the ChatOutput and InterfaceVertex classes to improve message handling. In ChatOutput, remove the unnecessary storage of the message response and update the text_response method to only store the message if a session ID is present. In InterfaceVertex, update the text_output assignment to handle different scenarios and ensure the correct text is used. Also, update the message assignment to handle instances of Message and str. These changes improve the code logic and maintain consistency in message handling across the components.
Update the VertexAIEmbeddingsComponent to use the `Embeddings` field type instead of importing it from `langchain_google_vertexai`. This ensures compatibility with the latest version of the langflow library and improves code readability.
Update the langflow components to use the `Prompt` field type instead of the deprecated `type` attribute. This ensures compatibility with the latest version of the langflow library and improves code readability.
Refactor the ChatInput and ChatOutput components to store the message response in the session. This ensures that the message is properly stored and can be accessed later.
This pull request includes two changes. First, the `.gitignore` file is
updated to include `*.db-shm` and `*.db-wal` files. Second, the
`ChatInput` and `ChatOutput` classes are refactored to include a new
`return_message` parameter. This parameter allows the caller to specify
whether they want the message to be returned as a `Message` object or
just the message text. These changes improve the functionality and
maintainability of the code.
🐛 (schemas.py): Fix issue with logs field type in ResultDataResponse class
🐛 (Prompt.py): Fix issue with kwargs assignment in build_prompt method
🐛 (component.py): Fix issue with results and arguments assignment in Component class
🐛 (schema.py): Fix issue with logs field type in ResultData class
📝 (types.py): Add ResultData import and build_logs_from_artifacts function for better logging and message extraction
📝 (setup.py): Set DEFAULT_PROMPT_INTUT_TYPES for missing template keys to provide default input types
📝 (loading.py): Update build_component function to return artifacts along with build_results
📝 (schema.py): Add build_logs_from_artifacts function to generate logs from artifacts
📝 (utils.py): Update log_transaction function to use source instead of vertex and handle source result for outputs
📝 (prompt.py): Define DEFAULT_PROMPT_INTUT_TYPES constant and use it for missing template input types
* Create PythonStructuredTool
This draft involves receiving two 'Code' types as inputs and creating a structured tool.
* Delete src/backend/base/langflow/components/experimental/PythonStructuredTool
* Create PythonCodeStructuredTool.py
* feat: Update Couchbase component to handle import errors
* chore: Update package versions in pyproject.toml and poetry.lock
* refactor: Fix encoding issue in read_text_file function
* 🐛 (parse.py): fix issue where method was always set to "post" when data or data binary was provided in curl command
* refactor: Fix string formatting in MonitorService delete_messages method
* fix: Update Couchbase import instructions in CouchbaseComponent
* refactor: Fix encoding issue in read_text_file function
The code changes introduce two new components, `MessageComponent` and `StoreMessageComponent`, to the `langflow` project.
The `MessageComponent` creates a `Message` object given a session ID, allowing for the creation of chat messages with customizable sender type, sender name, and text.
The `StoreMessageComponent` stores a chat message, utilizing the `store_message` function and the `get_messages` function to retrieve the stored messages.
These changes aim to enhance the chat functionality of the project.
