workflows/unit_tests/assistants/recommend_test_context.py

117 lines
3.4 KiB
Python

import json
from typing import List, Optional
from devchat.llm.openai import chat_completion_no_stream_return_json
from llm_conf import (
CONTEXT_SIZE,
DEFAULT_CONTEXT_SIZE,
DEFAULT_ENCODING,
USE_USER_MODEL,
USER_LLM_MODEL,
)
from model import FuncToTest
from openai_util import create_chat_completion_content
from tools.tiktoken_util import get_encoding
MODEL = USER_LLM_MODEL if USE_USER_MODEL else "gpt-4-turbo-preview"
ENCODING = (
get_encoding(DEFAULT_ENCODING) # Use default encoding as an approximation
if USE_USER_MODEL
else get_encoding("cl100k_base")
)
TOKEN_BUDGET = int(CONTEXT_SIZE.get(MODEL, DEFAULT_CONTEXT_SIZE) * 0.9)
# ruff: noqa: E501
recommend_symbol_context_prompt = """
You're an advanced AI test generator.
You're about to write test cases for the function `{function_name}` in the file `{file_path}`.
Before you start, you need to check if you have enough context information to write the test cases.
Here is the source code of the function:
```
{function_content}
```
And here are some context information that might help you write the test cases:
{context_content}
Do you think the context information is enough?
If the information is insufficient, recommend which symbols or types you need to know more about.
Return a JSON object with a single key "key_symbols" whose value is a list of strings.
- If the context information is enough, return an empty list.
- Each string is the name of a symbol or type appearing in the function that lacks context information for writing test.
- The list should contain the most important symbols and should not exceed 10 items.
JSON Format Example:
{{
"key_symbols": ["<symbol 1>", "<symbol 2>", "<symbol 3>",...]
}}
"""
def _mk_user_msg(func_to_test: FuncToTest, contexts: List) -> str:
"""
Create a user message to be sent to the model within the token budget.
"""
msg = None
while msg is None:
context_content = "\n\n".join([str(c) for c in contexts])
msg = recommend_symbol_context_prompt.format(
function_content=func_to_test.func_content,
context_content=context_content,
function_name=func_to_test.func_name,
file_path=func_to_test.file_path,
)
token_count = len(ENCODING.encode(msg, disallowed_special=()))
if contexts and token_count > TOKEN_BUDGET:
# Remove the last context and try again
contexts.pop()
msg = None
return msg
def get_recommended_symbols(
func_to_test: FuncToTest, known_context: Optional[List] = None
) -> List[str]:
known_context = known_context or []
msg = _mk_user_msg(func_to_test, known_context)
json_res = {}
if USE_USER_MODEL:
# Use the wrapped api parameters
json_res = (
chat_completion_no_stream_return_json(
messages=[{"role": "user", "content": msg}],
llm_config={
"model": MODEL,
"temperature": 0.1,
},
)
or {}
)
else:
response = create_chat_completion_content(
model=MODEL,
messages=[{"role": "user", "content": msg}],
response_format={"type": "json_object"},
temperature=0.1,
)
json_res = json.loads(response)
key_symbols = json_res.get("key_symbols", [])
return key_symbols