fix: turn off deprecated correction by default (#196)

* mention Claude in docs

* switch correction default off

biochatter/llm_connect.py

@@ -77,7 +77,7 @@ def __init__(
         self,
         model_name: str,
         prompts: dict,
-        correct: bool = True,
+        correct: bool = False,
         split_correction: bool = False,
         use_ragagent_selector: bool = False,
     ):
@@ -458,7 +458,7 @@ def __init__(
         self,
         model_name: str,
         prompts: dict,
-        correct: bool = True,
+        correct: bool = False,
         split_correction: bool = False,
     ):
         """
@@ -523,7 +523,7 @@ def __init__(
         base_url: str,
         prompts: dict,
         model_name: str = "auto",
-        correct: bool = True,
+        correct: bool = False,
         split_correction: bool = False,
     ):
         """
@@ -891,7 +891,7 @@ def __init__(
         base_url: str,
         prompts: dict,
         model_name: str = "llama3",
-        correct: bool = True,
+        correct: bool = False,
         split_correction: bool = False,
     ):
         """
@@ -1095,7 +1095,7 @@ def __init__(
         self,
         model_name: str,
         prompts: dict,
-        correct: bool = True,
+        correct: bool = False,
         split_correction: bool = False,
     ):
         """
@@ -1317,7 +1317,7 @@ def __init__(
         self,
         model_name: str,
         prompts: dict,
-        correct: bool = True,
+        correct: bool = False,
         split_correction: bool = False,
     ):
         """
@@ -1478,7 +1478,7 @@ def __init__(
         deployment_name: str,
         model_name: str,
         prompts: dict,
-        correct: bool = True,
+        correct: bool = False,
         split_correction: bool = False,
         version: Optional[str] = None,
         base_url: Optional[str] = None,

docs/features/chat.md

@@ -60,6 +60,20 @@ designate the `OPENAI_API_KEY` variable in your environment directly (`export
 OPENAI_API_KEY=sk-...`) by adding it to your shell configuration (e.g., the
 `zshrc`).
 
+## Using Anthropic models (Claude)
+
+Similarly, to use an Anthropic model, you need a billable account with Anthropic
+API access, and to set the `ANTHROPIC_API_KEY` variable in your environment.
+
+```python
+from biochatter.llm_connect import AnthropicConversation
+
+conversation = AnthropicConversation(
+    model_name="claude-3-5-sonnet-20240620",
+    prompts={},
+)
+```
+
 ## Multimodal models - Text and image
 
 We support multimodal queries in models that offer these capabilities after the