💡 chore(loading.py): improve code readability and maintainability by simplifying the instantiation of the prompt object and assignment of format_kwargs dictionary.
The code has been refactored to use a dictionary comprehension to create the format_kwargs dictionary. This simplifies the logic and improves code readability. The commented out code for the prompt.partial() method has been removed as it is no longer necessary.
✨ feat(utils.py): add support for setting input keys values from artifacts in build_input_keys_response function
The build_input_keys_response function now takes an additional parameter, artifacts, which is a dictionary containing key-value pairs. The function sets the values of the input keys in the input_keys_response dictionary based on the corresponding keys in the artifacts dictionary. This allows for more flexibility in setting the input keys values dynamically based on the provided artifacts.
✨ feat(chat.py): update build_input_keys_response function to accept artifacts parameter to set input_keys values
The stream_build function now supports the artifacts parameter, which allows passing prompt variables to the build_input_keys_response function. This ensures that the input_keys values are correctly set based on the provided artifacts. The build_input_keys_response function has been updated to accept the artifacts parameter and use it to set the input_keys values. This improves the functionality of the chat module by allowing more flexibility in handling prompt variables.
🐛 fix(vertex/base.py): update instantiation logic to handle tuple result from loading.instantiate_class()
🐛 fix(loading.py): update return value of instantiate_prompt() to return a tuple of prompt and format_kwargs
The Vertex class now has a new attribute called artifacts, which is a dictionary used to store additional data related to the vertex. The instantiation logic in the Vertex class has been updated to handle the case where loading.instantiate_class() returns a tuple containing the built object and additional artifacts. The loading.instantiate_prompt() function now returns a tuple containing the prompt and format_kwargs. These changes fix issues related to storing and handling additional data in the vertex and loading modules.
The build_input_keys_response function now correctly filters out memory variables from the input keys response. Previously, it was using a list comprehension to filter the keys, but it should have been using a dictionary comprehension to preserve the key-value pairs. This fix ensures that only the keys that are not present in the langchain_object's memory variables are included in the input keys response.
The chat history is now reset when building a flow to ensure that the chat starts with a clean state. This helps prevent any potential issues or conflicts that may arise from previous chat interactions.
🐛 fix(loading.py): set input variable values if they are present in params
The "input_variables" field in the prompt's frontend_node template is now updated with the new list of variables, if it exists. This ensures that the input variables are correctly set when validating the prompt. In the loading module, the input variable values are now set if they are present in the params dictionary. This ensures that the prompt is correctly instantiated with the provided input variable values.
🐛 fix(utils.py): rename variable in process_graph function to improve clarity and consistency
In the `ChatManager` class, the variable `chat_message` has been renamed to `chat_inputs` to better reflect its purpose of storing the input messages from the chat payload. This improves clarity and consistency within the class.
In the `process_graph` function in `utils.py`, the variable `chat_message` has been renamed to `chat_inputs` to better reflect its purpose of storing the input messages from the chat payload. This improves clarity and consistency within the function.
The build_input_keys_response function has been refactored to return an empty string for each input key in the langchain_object. This change ensures consistency in the response structure and provides a default value for each input key.
The input_types attribute is added to the template_field in the post_validate_prompt function. This change allows for specifying the input types that are accepted for the template field, providing more flexibility and control over the input data.
🔧 fix(index.ts): add input_types property to TemplateVariableType interface
The `input_types` property is added to the `TemplateFieldCreator` class in `base.py` and the `TemplateVariableType` interface in `index.ts`. This change allows for specifying the input types for template fields and template variables, respectively.
The message field of the ChatMessage schema now supports a dictionary type in addition to the previous string and None types. This change was made to allow for more complex messages to be sent and received in the chat.
✨ feat(validate.py): add support for dynamic template fields in prompt validation
The changes in this commit remove unused variables from the prompt.frontend_node.template. The commit also adds support for dynamic template fields in prompt validation. The new variables are added to the template and the custom_fields list. The old custom_fields list is copied and then updated with the new variables. The variables that are not in the template anymore are removed from the prompt.frontend_node.template.
The `dict()` method was causing a TypeError when trying to serialize the `template_field` object. The `to_dict()` method is the correct method to use to serialize the object.
🔥 chore(base.py): remove unused FrontendNodeRequest class
🔨 refactor(validate.py): modify post_validate_prompt function to update the template field in the frontend_node object
The ValidatePromptRequest class was using the FrontendNode class instead of the FrontendNodeRequest class, which caused an error. The unused FrontendNodeRequest class was removed. The post_validate_prompt function was modified to update the template field in the frontend_node object instead of appending to the fields list.
The FrontendNode class now has a new attribute called custom_fields, which is a list of strings. This attribute can be used to store any custom fields that are specific to a particular frontend node. This change allows for more flexibility in the data that can be stored in a FrontendNode object.
✨ feat(base.py): add frontend_node parameter to ValidatePromptRequest and PromptValidationResponse models
The validate_prompt endpoint now accepts a frontend_node parameter in the ValidatePromptRequest model. This parameter is used to add input variables to the frontend_node's template fields and custom fields. The PromptValidationResponse model now includes the frontend_node parameter to return the updated frontend_node object.
The config.yaml file has been updated to include new prompt templates. In types.py, the input_variables list was not being cleared when prompt or messages were present, which has been fixed. The instantiate_prompt function in loading.py has been updated to support the new prompt templates.
🔀 chore(config): add new prompt templates to config.yaml
🐛 fix(types.py): fix input_variables not being cleared when prompt or messages are present
The `AgentType` enum is added to the `langchain.agents.custom` module to improve readability and type safety. The `InitializeAgent` class now uses the `AgentType` enum to ensure that the `agent` parameter is a valid value from the enum.
🚀 feat(chat.py): use build_input_keys_response function to build input keys response
The build_input_keys_response function was extracted from the remove_api_keys function to improve code reusability. It builds the input keys response object, which contains the input keys and memory keys of a langchain object. The chat.py file now uses this function to build the input keys response object instead of building it manually. This improves code readability and reduces code duplication.
The commented out code was removed to improve code readability. Input validation was added to ensure that the template and input variables are valid before creating a PromptTemplate object. This helps to prevent errors that may occur when the object is created with invalid input.
🔒 chore(base.py): remove commented out code and add input validation for PromptTemplate
The stream_build function has been refactored to send input_keys to the client if the langchain_object has the attribute "input_keys". This change improves the user experience by providing the client with the necessary input keys to continue the conversation.
The `update_settings` function now accepts a `cache` parameter that allows the user to specify the type of cache to use. The `cache` parameter is set to a default value of `SQLiteCache` and can be overridden by setting the `LANGCHAIN_CACHE` environment variable. This feature improves the flexibility of the application as it allows the user to choose the type of cache that best suits their needs.
The cache configuration option has been added to the settings file with a default value of "InMemoryCache". This allows the user to choose the cache implementation they want to use.
This commit adds support for configurable LLM caching. The `setup_llm_caching` function now imports the cache class from the `langchain.cache` module based on the `settings.cache` value. If the import is successful, the `langchain.llm_cache` is set to an instance of the cache class. If the import fails, a warning is logged. If an exception is raised during the setup, a warning is logged with the error message.
✨ feat(main.py): call setup_llm_caching function on app startup
The `setup_llm_caching` function is added to `utils.py` to set up LLM caching. The function is then called on app startup in `main.py` using the `app.on_event("startup")` method. This improves the performance of the application by caching LLM objects.
The SlackDirectoryLoader is added to the list of document loaders in the DocumentLoaderFrontNode class. This allows users to load zip files from Slack into the application.
The GitLoader template now has four new fields: repo_path, clone_url, branch, and file_filter. These fields allow the user to specify the repository path, clone URL, branch, and file extensions to be loaded. This improves the flexibility of the GitLoader template and allows it to be used in a wider range of scenarios. Additionally, a minor change was made to the add_extra_fields method to ensure that the field.show attribute is set to True for all fields.
The `instantiate_documentloader` function now supports filtering files by extension using a `file_filter` parameter. The parameter is a string of comma-separated extensions, and the function now converts it into a lambda function that filters files based on whether their name contains any of the specified extensions. This improves the flexibility of the document loader by allowing it to load only specific types of files.
The fields in the Template class were previously sorted by DIRECT_TYPES, which caused issues when fields had the same field_type. Sorting alphabetically first ensures that fields are sorted in a consistent manner before sorting by DIRECT_TYPES.
The `instantiate_vectorstore` function now uses a dictionary to initialize vector stores instead of a series of if-else statements. This improves the readability and maintainability of the code. A new dictionary `vecstore_initializer` is added to `vector_store.py` to map the class names of vector stores to their respective initialization functions.
The `instantiate_vectorstore` function now supports the `MongoDBAtlasVectorSearch` vector store. This allows for the use of MongoDB Atlas as a vector store for Langflow. The `search_kwargs` parameter is now supported for all vector stores that have a `as_retriever` method. This allows for the configuration of the vector store's search parameters.
The hardcoded values for db_name, collection_name, and index_name have been removed from the initialize_mongodb function and are now required parameters. This makes the function more flexible and allows it to be used with different databases and collections. The support for the index_name parameter has been added to the MongoDBAtlasVectorSearch template in vectorstores.py, which allows the user to specify the name of the index to be used in the search.
🐛 fix(vector_store.py): remove hardcoded values for db_name, collection_name, and index_name and make them required parameters
✨ feat(langflow): add support for search_kwargs field in VectorStoreFrontendNode
The changes add support for MongoDB Atlas Vector Search in the vectorstores. The `MongoDBAtlasVectorSearch` class is now imported and initialized in `vector_store.py`. The `initialize_mongodb` function is added to initialize the MongoDB Atlas Vector Search class. The `VectorStoreFrontendNode` class is updated to add the `mongodb_atlas_cluster_uri`, `collection_name`, and `db_name` fields. The `search_kwargs` field is also added to the `VectorStoreFrontendNode` class to allow users to pass additional search parameters to the vector store.
The type hinting for allowed_tools variable is unnecessary as it is already defined in the previous line. Removing the type hinting improves the readability of the code.
The display name for the SupabaseVectorStore is now set to "Supabase". This improves the user experience by providing a more descriptive name for the vector store.
