update_stored_issues.py

"""
Module which updates any of the issues to reflect changes in the issue state
"""
import json
import datetime
from defaults import TOKEN, OWNER, REPO

GITHUB_API_VERSION = "2022-11-28"



# Get the issues that have been updated since the last update

import json

import argparse

import requests
import os
import numpy as np
import json
import datetime
import logging

logging.basicConfig(level=logging.INFO)

logger = logging.getLogger(__name__)

today = datetime.datetime.now().strftime("%Y_%m_%d_%H_%M_%S")

OWNER = "huggingface"
REPO = "transformers"
GITHUB_API_VERSION = "2022-11-28"
TOKEN = os.environ.get("GITHUB_TOKEN")
JSON_FILE = f"issues.json"


def get_issues(
    input_filename=JSON_FILE,
    output_filename=JSON_FILE,
    github_api_version=GITHUB_API_VERSION,
    owner=OWNER,
    repo=REPO,
    token=TOKEN,
    n_pages=-1,
):
    """
    Function to get the issues from the transformers repo and save them to a json file
    """
    with open("issues_dict.json", "r") as f:
        issues = json.load(f)

    # Get most recent updated at information
    updated_at = [issue["updated_at"] for issue in issues.values()]
    most_recent = max(updated_at)

    # If file exists and we want to overwrite it, delete it
    if not os.path.exists(output_filename):
        raise ValueError(f"File {output_filename} does not exist")

    # Define the URL and headers
    url = f"https://api.github.com/repos/{owner}/{repo}/issues"
    headers = {
        "Accept": "application/vnd.github+json",
        f"Authorization": f"{token}",
        "X-GitHub-Api-Version": f"{github_api_version}",
        "User-Agent": "amyeroberts",
    }
    per_page = 100
    page = 1
    query_params = {
        "state": "all",
        "since": "2024-02-01T11:33:35Z",
        # "since": most_recent,
        "sort": "created",
        "direction": "asc",
        "page": page,
    }

    new_lines = []

    page_limit = (n_pages + page) if n_pages > 0 else np.inf
    while True:
        if page >= page_limit:
            break

        # Send the GET request
        response = requests.get(url, headers=headers, params=query_params)

        if not response.status_code == 200:
            raise ValueError(
                f"Request failed with status code {response.status_code} and message {response.text}"
            )

        json_response = response.json()
        logger.info(f"Page: {page}, number of issues: {len(json_response)}")

        # If we get an empty response, we've reached the end of the issues
        if len(json_response) == 0:
            break

        new_lines.extend(json_response)

        # If we get less than the number of issues per page, we've reached the end of the issues
        if len(json_response) < per_page:
            break

        page += 1
        query_params["page"] = page

    issue_lines_map = {issue["number"]: issue for issue in new_lines}

    with open(input_filename, "r") as f:
        with open("tmp_" + output_filename, "a") as g:
            for line in f:
                issue = json.loads(line)
                number = issue["number"]
                if number in issue_lines_map:
                    g.write(json.dumps(issue_lines_map[number]))
                    g.write("\n")
                else:
                    g.write(line)

    os.rename("tmp_" + output_filename, output_filename)

    with open("updated_issues.json", "w") as f:
        json.dump(issue_lines_map, f, indent=4, sort_keys=True)

    return output_filename


if __name__ == "__main__":
    parser = argparse.ArgumentParser()
    parser.add_argument("--input_filename", type=str, default=JSON_FILE)
    parser.add_argument("--output_filename", type=str, default=JSON_FILE)
    parser.add_argument("--github_api_version", type=str, default=GITHUB_API_VERSION)
    parser.add_argument("--owner", type=str, default=OWNER)
    parser.add_argument("--repo", type=str, default=REPO)
    parser.add_argument("--token", type=str, default=TOKEN)
    parser.add_argument("--n_pages", type=int, default=-1)
    args = parser.parse_args()
    get_issues(**vars(args))