mirror of
https://github.com/OpenBMB/ChatDev.git
synced 2024-09-19 14:17:32 +03:00
62 lines
3.3 KiB
Python
62 lines
3.3 KiB
Python
import json
|
|
import re
|
|
import openai
|
|
class ProjectEvaluator:
|
|
def __init__(self):
|
|
self.prompt = """You are a professional start-up project judge. Please score the following open source software project based on the information provided, on a scale of 1 to 10. Your scoring should be divided into three dimensions: feasibility, usability, and innovativeness. Your return result should be a JSON format dictionary. An example is in the following line\n'{"feasibility": {"score": 8.5, "reason": "the idea of this project is simple but natural. tools and tech-schemes it requires are very mature so that it is easy to be implemented"}, "usability": {"score": 9.0, "reason": "the function it claims is very useful. it can help many people enhance efficiency"}, "novelty": {"score": 3.5, "reason": "main the idea of this project is not very frontier"}}'.\nNOTE: You should NOT copy the statement in the example above. You should write your reason independently."""
|
|
def evaluate_project(self, readme_file):
|
|
# Read the README.md file
|
|
readme_content = self.read_file(readme_file)
|
|
# Extract relevant information from the README.md file
|
|
project_name = self.extract_project_name(readme_content)
|
|
project_description = self.extract_project_description(readme_content)
|
|
# Add more evaluation criteria as needed
|
|
# Calculate the score based on the extracted information
|
|
score = self.calculate_score(project_name, project_description)
|
|
return score
|
|
def read_file(self, file_path):
|
|
with open(file_path, "r") as file:
|
|
content = file.read()
|
|
return content
|
|
def extract_project_name(self, readme_content):
|
|
# Extract project name from the README.md file
|
|
# Implement your logic here
|
|
project_name = ""
|
|
match = re.search(r"#\s*(.*)", readme_content)
|
|
if match:
|
|
project_name = match.group(1)
|
|
return project_name
|
|
def extract_project_description(self, readme_content):
|
|
# Extract project description from the README.md file
|
|
# Implement your logic here
|
|
project_description = ""
|
|
match = re.search(r"##\s*Description\n\n(.*)", readme_content)
|
|
if match:
|
|
project_description = match.group(1)
|
|
return project_description
|
|
def calculate_score(self, project_name, project_description):
|
|
# Calculate the score based on the project name and description
|
|
# Implement your logic here
|
|
score = 0
|
|
resp = "### NOT POST YET ###"
|
|
for i in range(10):
|
|
try:
|
|
print("post request ", i)
|
|
resp = openai.ChatCompletion.create(
|
|
model="gpt-3.5-turbo",
|
|
messages=[
|
|
{"role": "system", "content": self.prompt},
|
|
{"role": "user", "content": f"Project Name: {project_name}\nProject Description: {project_description}\n"}
|
|
]
|
|
)
|
|
print("response got", i)
|
|
content = resp.choices[0]["message"]["content"]
|
|
json_str = re.search(r'\{.+\}', content, re.S).group(0)
|
|
scores_dict = json.loads(json_str)
|
|
return scores_dict
|
|
except Exception as e:
|
|
print(e)
|
|
print(resp)
|
|
print('api calling failed')
|
|
return
|