otieno-juma
commented
2 weeks ago I've created a script using groq and incorporated some security measures, such as masking github tokens using os environs.I have documented the script to explain what each item does. If anyone has suggestions for improvements or deletions, it would be greatly appreciated! I'm in the process of setting up my test environment. If you have any lack with it please share your findings.
"""
This script creates pull requests on a GitHub repository to add examples for NumPy functions that have an 'axis' parameter.
The script first sets up a GitHub instance using an environment variable for the GitHub token. It then retrieves all the NumPy functions that have an 'axis' parameter, and creates a dictionary to store the examples and explanations for each function.
The script then creates a pull request template that includes the function name, three examples, and their descriptions. It then creates a pull request for each function, with the title 'Add examples for {function_name}' and the pull request body filled in with the examples and descriptions.
The script waits for 10 seconds between creating each pull request to avoid rate limiting.
Finally, the script prints the number of pull requests that were created.
"""
import importlib
import numpy as np
import difflib
from github import Github
from github.GithubException import UnknownObjectException
import os
import time
import inspect
import github
from github.GithubException import UnknownObjectException
import git
# Set up environment variable for GitHub token
GITHUB_TOKEN = os.getenv("GITHUB_TOKEN")
# Check if the environment variable is set
if GITHUB_TOKEN is None:
print("Error: GITHUB_TOKEN environment variable is not set")
exit()
# Create a GitHub instance with the token
g = Github(GITHUB_TOKEN)
# Set up the repository and branch you want to create pull requests for
repo = g.get_repo("your-repo-name")
branch = repo.get_branch("main")
# Get all numpy functions with axis parameters
axis_params = []
for name in dir(np):
obj = getattr(np, name)
if hasattr(obj, "__call__") and "axis" in inspect.signature(obj).parameters:
axis_params.append((name, obj))
# Create a dictionary to store the examples and explanations for each function
examples = {}
# Create the pull request template
pr_template = """
# Examples for {function_name}
{function_name} is used to perform various operations on arrays. Here are some examples of how to use it:
**Example 1:**
{example_code}
This code does {example_description}
**Example 2:**
{example_code}
This code does {example_description}
**Example 3:**
{example_code}
This code does {example_description}
Please let me know if you have any questions or need further clarification.
"""
# Create the pull requests
pr_count = 0
for func, obj in axis_params:
# Create the examples and explanation for the function
examples[func] = []
for i in range(3):
example_code = f"{func}(np.array([1, 2, 3]), axis={i})"
example_description = f"Performing operation on axis {i}"
examples[func].append((example_code, example_description))
# Create the pull request
pr_title = f"Add examples for {func}"
pr_body = pr_template.format(function_name=func, **examples[func])
pr = repo.create_pull(title=pr_title, body=pr_body, head=branch.name, base=branch.name)
pr_count += 1
# Wait for 10 seconds before creating the next pull request to avoid rate limiting
time.sleep(10)
print(f"Created {pr_count} pull requests!")
bmwoodruff
Description:
At the triage meeting, one place where additional examples could be welcome is functions that have an axis element. The axis parameter can be tricky (and many of the AI examples generated so far have completely gotten it wrong).
I bet we could design a better prompt that helps generate these types of examples, and includes an example to help understand the axis parameter.
Acceptance Criteria: