chatlog

chatlog#

Attributes#

`logger`
`log`
`mypath`
`json_keyfile_path`
`file`

Classes#

`ChatLogItem`	Represents a chat log entry for a request and response
`ChatLogger`	Handles chat logging functionality
`ChatSocketAPI`	An instance of this class is created for each chat request.
`ChatSocketIOHandler`	A handler class which writes logging records, appropriately formatted,
`LangchainLoggingHandler`

Functions#

`initialize_json_log`()
`log_to_json`(context, prompt, prompt_template, response)
`log_chat_item`(project, question, output, ...[, error])	Log a chat interaction to the chat log file.

Module Contents#

chatlog.logger[source]#

chatlog.log[source]#

class chatlog.ChatLogItem[source]#

Represents a chat log entry for a request and response

context: str[source]#

query: str[source]#

prompt_template: str[source]#

response: str[source]#

timestamp: str[source]#

conversation_id: str | None = None[source]#

view_name: str | None = None[source]#

error: bool | None = None[source]#

to_dict() → Dict[str, Any][source]#: Convert to dictionary for JSON serialization

classmethod from_dict(data: Dict[str, Any]) → ChatLogItem[source]#: Create from dictionary (e.g. from JSON)

class chatlog.ChatLogger(log_file_path: str)[source]#

Handles chat logging functionality

log_file_path[source]#

_ensure_log_file_exists()[source]#: Ensure the log file exists and is properly initialized

log_chat(item: ChatLogItem)[source]#: Log a chat item to the JSON file

get_logs() → List[ChatLogItem][source]#: Get all chat logs

get_conversation_logs(conversation_id: str) → List[ChatLogItem][source]#: Get logs for a specific conversation

class chatlog.ChatSocketAPI(project: mdvtools.mdvproject.MDVProject, id: str, room: str, conversation_id: str)[source]#

An instance of this class is created for each chat request. It will instantiate a Logger instance & SocketIOHandler which should be GCed when the request is finished.

project[source]#

socketio = None[source]#

progress_name = 'chat_progress'[source]#

project_namespace = '/project/Uninferable'[source]#

room[source]#

logger[source]#

log(msg: str)[source]#

update_chat_progress(message: str, id: str, progress: int, delta: int)[source]#

Send a message to the chat log and also update the progress bar.

Parameters:

message (str) – the message to send to the chat log
id (str) – the id of the associated chat request
progress (int) – the progress value (%) to update the progress bar with
delta (int) – the expected cost of the current operation (%)

class chatlog.ChatSocketIOHandler(socketio: flask_socketio.SocketIO, event_name: str, namespace: str, id: str, room: str)[source]#

Bases: logging.StreamHandler

A handler class which writes logging records, appropriately formatted, to a stream. Note that this class does not close the stream, as sys.stdout or sys.stderr may be used.

socketio[source]#

event_name[source]#

namespace[source]#

id[source]#

room[source]#

emit(record)[source]#: Emit a record - send it via socketio & also print it to the console. subject to change.

chatlog.mypath[source]#

chatlog.json_keyfile_path[source]#

chatlog.initialize_json_log()[source]#

chatlog.log_to_json(context, prompt, prompt_template, response)[source]#

chatlog.file = False[source]#

class chatlog.LangchainLoggingHandler(logger: logging.Logger)[source]#

Bases: langchain_core.callbacks.BaseCallbackHandler

logger[source]#

log[source]#

on_chat_model_start(serialized: Dict[str, Any], messages: List[List[langchain_core.messages.BaseMessage]], **kwargs) → None[source]#

on_llm_end(response: langchain_core.outputs.LLMResult, **kwargs) → None[source]#

on_chain_start(serialized: Dict[str, Any], inputs: Dict[str, Any], **kwargs) → None[source]#

on_chain_end(outputs: Dict[str, Any], **kwargs) → None[source]#

chatlog.log_chat_item(project, question, output, prompt_template, response, conversation_id, view_name: str | None, error: bool = False)[source]#: Log a chat interaction to the chat log file. :param project: The MDVProject instance :param output: Result of invoke ‘from langchain.chains import RetrievalQA’ (can be None for errors) :param prompt_template: The template used for the prompt (can be empty for errors) :param response: The response generated (error message if error) :param conversation_id: ID to group messages from the same conversation :param error: Whether this log is for an error