prepare release

CHANGELOG.md

@@ -8,6 +8,21 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/).
 
 Nothing is unreleased!
 
+## [1.0.400] - 2023-03-06
+
+### Added
+
+- OpenAI integration
+
+### Fixed
+
+- Langchain final answer streaming should work again
+- Elements with public URLs should be correctly persisted by the data layer
+
+### Changed
+
+- Enforce UTC DateTimes
+
 ## [1.0.300] - 2023-02-19
 
 ### Added

backend/chainlit/element.py

@@ -123,14 +123,14 @@ def from_dict(self, _dict: FileDict):
  )
 
  async def _create(self) -> bool:
- if (self.persisted or self.url) and not self.updatable:
+ if self.persisted and not self.updatable:
  return True
  if data_layer := get_data_layer():
  try:
  asyncio.create_task(data_layer.create_element(self))
  except Exception as e:
  logger.error(f"Failed to create element: {str(e)}")
- if not self.chainlit_key or self.updatable:
+ if not self.url and (not self.chainlit_key or self.updatable):
  file_dict = await context.session.persist_file(
  name=self.name,
  path=self.path,
@@ -203,6 +203,7 @@ class Text(Element):
 class Pdf(Element):
  """Useful to send a pdf to the UI."""
 
+ mime: str = "application/pdf"
  page: Optional[int] = None
  type: ClassVar[ElementType] = "pdf"
 

backend/chainlit/emitter.py

@@ -1,6 +1,5 @@
 import asyncio
 import uuid
-from datetime import datetime
 from typing import Any, Dict, List, Literal, Optional, Union, cast
 
 from chainlit.data import get_data_layer
@@ -18,6 +17,7 @@
  UIMessagePayload,
 )
 from chainlit.user import PersistedUser
+from literalai.helper import utc_now
 from socketio.exceptions import TimeoutError
 
 
@@ -196,7 +196,7 @@ async def process_user_message(self, payload: UIMessagePayload):
 
  message = Message.from_dict(step_dict)
  # Overwrite the created_at timestamp with the current time
- message.created_at = datetime.utcnow().isoformat()
+ message.created_at = utc_now()
 
  asyncio.create_task(message._create())
 

backend/chainlit/haystack/callbacks.py

@@ -1,13 +1,14 @@
-from datetime import datetime
-from typing import Any, Generic, List, Optional, TypeVar
 import re
+from typing import Any, Generic, List, Optional, TypeVar
 
 from chainlit.context import context
 from chainlit.step import Step
 from chainlit.sync import run_sync
-from chainlit import Message
 from haystack.agents import Agent, Tool
 from haystack.agents.agent_step import AgentStep
+from literalai.helper import utc_now
+
+from chainlit import Message
 
 T = TypeVar("T")
 
@@ -36,7 +37,12 @@ class HaystackAgentCallbackHandler:
  stack: Stack[Step]
  last_step: Optional[Step]
 
- def __init__(self, agent: Agent, stream_final_answer: bool = False, stream_final_answer_agent_name: str = 'Agent'):
+ def __init__(
+ self,
+ agent: Agent,
+ stream_final_answer: bool = False,
+ stream_final_answer_agent_name: str = "Agent",
+ ):
  agent.callback_manager.on_agent_start += self.on_agent_start
  agent.callback_manager.on_agent_step += self.on_agent_step
  agent.callback_manager.on_agent_finish += self.on_agent_finish
@@ -56,14 +62,16 @@ def on_agent_start(self, **kwargs: Any) -> None:
  self.stack = Stack[Step]()
 
  if self.stream_final_answer:
- self.final_stream = Message(author=self.stream_final_answer_agent_name, content="")
+ self.final_stream = Message(
+ author=self.stream_final_answer_agent_name, content=""
+ )
  self.last_tokens: List[str] = []
  self.answer_reached = False
 
  root_message = context.session.root_message
  parent_id = root_message.id if root_message else None
  run_step = Step(name=self.agent_name, type="run", parent_id=parent_id)
- run_step.start = datetime.utcnow().isoformat()
+ run_step.start = utc_now()
  run_step.input = kwargs
 
  run_sync(run_step.send())
@@ -73,7 +81,7 @@ def on_agent_start(self, **kwargs: Any) -> None:
  def on_agent_finish(self, agent_step: AgentStep, **kwargs: Any) -> None:
  if self.last_step:
  run_step = self.last_step
- run_step.end = datetime.utcnow().isoformat()
+ run_step.end = utc_now()
  run_step.output = agent_step.prompt_node_response
  run_sync(run_step.update())
 
@@ -85,7 +93,7 @@ def on_agent_step(self, agent_step: AgentStep, **kwargs: Any) -> None:
  # If token streaming is disabled
  if self.last_step.output == "":
  self.last_step.output = agent_step.prompt_node_response
- self.last_step.end = datetime.utcnow().isoformat()
+ self.last_step.end = utc_now()
  run_sync(self.last_step.update())
 
  if not agent_step.is_last():
@@ -101,12 +109,16 @@ def on_new_token(self, token, **kwargs: Any) -> None:
  else:
  self.last_tokens.append(token)
 
- last_tokens_concat = ''.join(self.last_tokens)
- final_answer_match = re.search(self.final_answer_pattern, last_tokens_concat)
+ last_tokens_concat = "".join(self.last_tokens)
+ final_answer_match = re.search(
+ self.final_answer_pattern, last_tokens_concat
+ )
 
  if final_answer_match:
  self.answer_reached = True
- run_sync(self.final_stream.stream_token(final_answer_match.group(1)))
+ run_sync(
+ self.final_stream.stream_token(final_answer_match.group(1))
+ )
 
  run_sync(self.stack.peek().stream_token(token))
 
@@ -115,7 +127,7 @@ def on_tool_start(self, tool_input: str, tool: Tool, **kwargs: Any) -> None:
  parent_id = self.stack.items[0].id if self.stack.items[0] else None
  tool_step = Step(name=tool.name, type="tool", parent_id=parent_id)
  tool_step.input = tool_input
- tool_step.start = datetime.utcnow().isoformat()
+ tool_step.start = utc_now()
  self.stack.push(tool_step)
 
  def on_tool_finish(
@@ -128,13 +140,13 @@ def on_tool_finish(
  # Tool finished, send step with tool_result
  tool_step = self.stack.pop()
  tool_step.output = tool_result
- tool_step.end = datetime.utcnow().isoformat()
+ tool_step.end = utc_now()
  run_sync(tool_step.update())
 
  def on_tool_error(self, exception: Exception, tool: Tool, **kwargs: Any) -> None:
  # Tool error, send error message
  error_step = self.stack.pop()
  error_step.is_error = True
  error_step.output = str(exception)
- error_step.end = datetime.utcnow().isoformat()
+ error_step.end = utc_now()
  run_sync(error_step.update())

backend/chainlit/langchain/callbacks.py

@@ -1,6 +1,5 @@
 import json
 import time
-from datetime import datetime
 from typing import Any, Dict, List, Optional, TypedDict, Union
 from uuid import UUID
 
@@ -13,6 +12,7 @@
 from langchain.schema.output import ChatGenerationChunk, GenerationChunk
 from langchain_core.outputs import ChatGenerationChunk, GenerationChunk
 from literalai import ChatGeneration, CompletionGeneration, GenerationMessage
+from literalai.helper import utc_now
 from literalai.step import TrueStepType
 
 DEFAULT_ANSWER_PREFIX_TOKENS = ["Final", "Answer", ":"]
@@ -180,6 +180,10 @@ def _convert_message(
  content="",
  )
 
+ if literal_uuid := message.additional_kwargs.get("uuid"):
+ msg["uuid"] = literal_uuid
+ msg["templated"] = True
+
  if name := getattr(message, "name", None):
  msg["name"] = name
 
@@ -353,6 +357,18 @@ def on_llm_new_token(
  if start["tt_first_token"] is None:
  start["tt_first_token"] = (time.time() - start["start"]) * 1000
 
+ if self.stream_final_answer:
+ self._append_to_last_tokens(token)
+
+ if self.answer_reached:
+ if not self.final_stream:
+ self.final_stream = Message(content="")
+ self._run_sync(self.final_stream.send())
+ self._run_sync(self.final_stream.stream_token(token))
+ self.has_streamed_final_answer = True
+ else:
+ self.answer_reached = self._check_if_answer_reached()
+
  return super().on_llm_new_token(
  token,
  chunk=chunk,
@@ -458,7 +474,7 @@ def _start_trace(self, run: Run) -> None:
  parent_id=parent_id,
  disable_feedback=disable_feedback,
  )
- step.start = datetime.utcnow().isoformat()
+ step.start = utc_now()
  step.input = run.inputs
 
  self.steps[str(run.id)] = step
@@ -505,6 +521,15 @@ def _on_run_update(self, run: Run) -> None:
  ],
  message_completion=message_completion,
  )
+
+ # find first message with prompt_id
+ for m in chat_start["input_messages"]:
+ if m.additional_kwargs.get("prompt_id"):
+ current_step.generation.prompt_id = m.additional_kwargs["prompt_id"]
+ if custom_variables := m.additional_kwargs.get("variables"):
+ current_step.generation.variables = custom_variables
+ break
+
  current_step.language = "json"
  current_step.output = json.dumps(message_completion)
  else:
@@ -529,7 +554,7 @@ def _on_run_update(self, run: Run) -> None:
  current_step.output = completion
 
  if current_step:
- current_step.end = datetime.utcnow().isoformat()
+ current_step.end = utc_now()
  self._run_sync(current_step.update())
 
  if self.final_stream and self.has_streamed_final_answer:
@@ -547,7 +572,7 @@ def _on_run_update(self, run: Run) -> None:
 
  if current_step:
  current_step.output = output
- current_step.end = datetime.utcnow().isoformat()
+ current_step.end = utc_now()
  self._run_sync(current_step.update())
 
  def _on_error(self, error: BaseException, *, run_id: UUID, **kwargs: Any):
@@ -556,7 +581,7 @@ def _on_error(self, error: BaseException, *, run_id: UUID, **kwargs: Any):
  if current_step := self.steps.get(str(run_id), None):
  current_step.is_error = True
  current_step.output = str(error)
- current_step.end = datetime.utcnow().isoformat()
+ current_step.end = utc_now()
  self._run_sync(current_step.update())
 
  on_llm_error = _on_error

backend/chainlit/llama_index/callbacks.py

@@ -1,10 +1,10 @@
-from datetime import datetime
 from typing import Any, Dict, List, Optional
 
 from chainlit.context import context_var
 from chainlit.element import Text
 from chainlit.step import Step, StepType
 from literalai import ChatGeneration, CompletionGeneration, GenerationMessage
+from literalai.helper import utc_now
 from llama_index.callbacks import TokenCountingHandler
 from llama_index.callbacks.schema import CBEventType, EventPayload
 from llama_index.llms.base import ChatMessage, ChatResponse, CompletionResponse
@@ -87,7 +87,7 @@ def on_event_start(
  disable_feedback=False,
  )
  self.steps[event_id] = step
- step.start = datetime.utcnow().isoformat()
+ step.start = utc_now()
  step.input = payload or {}
  self.context.loop.create_task(step.send())
  return event_id
@@ -107,7 +107,7 @@ def on_event_end(
 
  self._restore_context()
 
- step.end = datetime.utcnow().isoformat()
+ step.end = utc_now()
 
  if event_type == CBEventType.RETRIEVE:
  sources = payload.get(EventPayload.NODES)

backend/chainlit/message.py

@@ -3,7 +3,6 @@
 import time
 import uuid
 from abc import ABC
-from datetime import datetime
 from typing import Dict, List, Optional, Union, cast
 
 from chainlit.action import Action
@@ -23,6 +22,7 @@
  FileDict,
 )
 from literalai import BaseGeneration
+from literalai.helper import utc_now
 from literalai.step import MessageStepType
 
 
@@ -149,7 +149,7 @@ async def _create(self):
 
  async def send(self):
  if not self.created_at:
- self.created_at = datetime.utcnow().isoformat()
+ self.created_at = utc_now()
  if self.content is None:
  self.content = ""
 
@@ -367,7 +367,7 @@ async def send(self) -> Union[StepDict, None]:
  """
  trace_event("send_ask_user")
  if not self.created_at:
- self.created_at = datetime.utcnow().isoformat()
+ self.created_at = utc_now()
 
  if config.code.author_rename:
  self.author = await config.code.author_rename(self.author)
@@ -439,7 +439,7 @@ async def send(self) -> Union[List[AskFileResponse], None]:
  trace_event("send_ask_file")
 
  if not self.created_at:
- self.created_at = datetime.utcnow().isoformat()
+ self.created_at = utc_now()
 
  if self.streaming:
  self.streaming = False
@@ -512,7 +512,7 @@ async def send(self) -> Union[AskActionResponse, None]:
  trace_event("send_ask_action")
 
  if not self.created_at:
- self.created_at = datetime.utcnow().isoformat()
+ self.created_at = utc_now()
 
  if self.streaming:
  self.streaming = False

backend/chainlit/openai/__init__.py

@@ -1,10 +1,11 @@
-from typing import TYPE_CHECKING, Union
+from typing import Union
 
 from chainlit.context import get_context
 from chainlit.step import Step
 from chainlit.sync import run_sync
 from chainlit.utils import check_module_version
 from literalai import ChatGeneration, CompletionGeneration
+from literalai.helper import timestamp_utc
 
 
 def instrument_openai():
@@ -34,12 +35,14 @@ async def on_new_generation(
  step.generation = generation
  # Convert start/end time from seconds to milliseconds
  step.start = (
- timing.get("start") * 1000
+ timestamp_utc(timing.get("start"))
  if timing.get("start", None) is not None
  else None
  )
  step.end = (
- timing.get("end") * 1000 if timing.get("end", None) is not None else None
+ timestamp_utc(timing.get("end"))
+ if timing.get("end", None) is not None
+ else None
  )
 
  if isinstance(generation, ChatGeneration):