Update predict_str is boot is None: bot = []

.ruff.toml

@@ -2,3 +2,16 @@
 target-version = "py310"
 # Decrease the maximum line length to 79 characters.
 line-length = 300
+
+# pyflakes, pycodestyle, isort
+# flake8 YTT, pydocstyle D, pylint PLC
+select = ["F", "E", "W", "I001", "YTT", "D", "PLC"]
+# select = ["ALL"]
+
+# D103 Missing docstring in public function
+# D101 Missing docstring in public class
+# `multi-line-summary-first-line` (D212)
+# `one-blank-line-before-class` (D203)
+extend-ignore = ["D103", "D101", "D212", "D203"]
+
+exclude = [".venv"]

app.py

@@ -1,4 +1,4 @@
-"""Run codes"""
+"""Run codes."""
 # pylint: disable=line-too-long, broad-exception-caught, invalid-name, missing-function-docstring, too-many-instance-attributes, missing-class-docstring
 # ruff: noqa: E501
 import os
@@ -35,6 +35,10 @@ assistant_prefix = "[assistant]: "
 
 def predict_str(prompt, bot):  # bot is in fact bot_history
     # logger.debug(f"{prompt=}, {bot=}, {timeout=}")
+
+    if bot is None:
+        bot = []
+
     logger.debug(f"{prompt=}, {bot=}")
 
     try:
@@ -150,7 +154,7 @@ def predict_api(prompt):
             stop=["<|im_end|>", "|<"],
         )
 
-        # TODO stream does not make sense in api?
+        # TODO: stream does not make sense in api?
         generator = generate(
             LLM, _, system_prompt=default_system_prompt, user_prompt=prompt.strip()
         )
@@ -215,7 +219,7 @@ def generate(
     system_prompt: str = default_system_prompt,
     user_prompt: str = "",
 ):
-    """Run model inference, will return a Generator if streaming is true"""
+    """Run model inference, will return a Generator if streaming is true."""
     # if not user_prompt.strip():
     return llm(
         format_prompt(
@@ -443,7 +447,7 @@ with gr.Blocks(
         fn=predict_str,
         inputs=[msg, chatbot],
         outputs=[msg, chatbot],
-        # queue=True,
+        queue=True,
         show_progress="full",
         api_name="predict",
     ).then(bot_str, chatbot, chatbot)