Langroid

Latest version: v0.2.2

Safety actively analyzes 638466 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 20 of 41

0.1.150

Task control improvements; LanceDB examples

0.1.149

* DocChatAgent: able to ingest Pandas dataframe, by first converting to Documents
see [tests/main/test_doc_chat_agent.py](https://github.com/langroid/langroid/blob/main/tests/main/test_doc_chat_agent.py),
* LanceDocChatAgent: ingest Pandas dataframe directly (without needing to convert to Documents)
see [tests/main/test_lance_doc_chat_agent](https://github.com/langroid/langroid/blob/main/tests/main/test_lance_doc_chat_agent.py)

* Task: new control mechanisms, which should replace the need for `llm_delegate`, `single_round`, but also enables more complex workflows, and is more clear and intuitive:
- when initializing a task, you can set 2 new params: `done_if_response` = list of Entities (Entity class), meaning if ANY of these entities
produces a non-empty response, the task is done. E.g. if you set it to [Entity.LLM],
this means the task is done as soon as there is a non-empty response from the LLM of the Task's Agent.
Similarly, `done_if_no_response` is a list of Entities, meaning if ANY of them produce an empty response, the task is done.
- An Agent's responders (`llm_response`, `user_response`, `agent_response` or tools) can now insert a string __PASS__ in their response, which tells the Task orchestrator to "pass-thru" the current pending message and not update it.

These mechanisms are illustrated in these tests, showing various scenarios of multi-agent + tools: [tests/main/test_task.py](https://github.com/langroid/langroid/blob/main/tests/main/test_task.py),

0.1.148

* DocChatAgentConfig: new `filter` param to be used to constrain searched document-set.
This gets used inside `get_semantic_search_results` and is passed to `self.vecdb.similar_texts_with_scores()`

* LanceDocChatAgent: version of DocChatAgent that leverages LanceDB filtered vector-search and full-text search
(see [`tests/main/test_lance_doc_chat_agent.py`](https://github.com/langroid/langroid/blob/main/tests/main/test_lance_doc_chat_agent.py))

0.1.147

0.1.146

* Task: new param `max_stalled_steps` (default = 3): If stalled for more than this many steps, quit
* Agent: friendly warning on mis-named tool
* DocChatAgent: make relevance extractor optional

0.1.145

* OpenAI streaming: use retry-with-exp-backoff
* ChatAgent.llm_response_forget: check msg history length
* LanceDB update to latest, which natively supports nested Pydantic models (so no need to use our workaround)
* Document.metadata.id: change type from str|None -> str, with default = ""

Page 20 of 41

Links

Releases

Has known vulnerabilities

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.