Codeberg seems to start experiencing issues search not as many results as expected #9

Open Benjamin-Loison opened 9 months ago

For instance: prodezarts.com Benjamin_Loison/Improve_websites_thanks_to_open_source#18.

Same issue with https://codeberg.org/api/swagger#/issue/issueListIssues.

Establishing a workaround based on download everything seems to make sense.

As I do not find, even with DuckDuckGo and Google search engines, a solution to export repository issues. Let us try to import on my own Gitea instance. Maybe it is a Codeberg intended limitation.

Maybe incorrect scopes, as there are two interesting ones: issue and repository.

Is there a dedicated option in Gitea for importing from Forgejo instance? As it seems to be such one on Codeberg.

With above both scopes (they are both required), I am able to import from Codeberg to Codeberg and then there is no issue anymore searching prodezarts.com for instance.

Are issue ids kept? I do not have this feeling at least with Improve_websites_thanks_to_open_source_issues/issues/300.

Description

For instance: https://codeberg.org/Benjamin_Loison/Improve_websites_thanks_to_open_source/issues?type=all&state=open&labels=&milestone=0&project=0&assignee=0&poster=0&q=github.com

+69

Gitea Version

Not read.

Can you reproduce the bug on the Gitea demo site?

Yes

Log Gist

No response

Screenshots

No response

Git Version

No response

Operating System

No response

How are you running Gitea?

Not read.

Database

None

#!/usr/bin/python3

# https://github.com/Benjamin-Loison/gitea/issues/9

import requests
import json
from tqdm import tqdm
import math

GITEA_INSTANCE_URL = 'https://codeberg.org'
# With `read:{issue,repository}` scopes.
TOKEN = 'CENSORED'
REPOSITORY_FULL_NAME = 'Benjamin_Loison/Improve_websites_thanks_to_open_source'

headers = {
    'Authorization': f'token {TOKEN}'
}

def getApi(url, params = {}):
    response = requests.get(f'{GITEA_INSTANCE_URL}/api/v1/{url}', params, headers = headers)
    return response.json()

LIMIT = 50
repositoryIssueParams = {
    'limit': LIMIT,
}
repositoryIssues = []

repository = getApi(f'repos/{REPOSITORY_FULL_NAME}')
if repository.get('message') == 'The target couldn\'t be found.':
    print(f'Repository {GITEA_INSTANCE_URL}/{REPOSITORY_FULL_NAME} not found!')
    exit(1)

# See [Benjamin-Loison/gitea/issues/60](https://github.com/Benjamin-Loison/gitea/issues/60).
openIssuesCount = repository['open_issues_count']
for state, total in [('closed', None), ('open', math.ceil(openIssuesCount / LIMIT))]:
    repositoryIssueParams['state'] = state
    with tqdm(total = total, desc = f'{state.title()} issues page') as progressBar:
        while True:
            repositoryIssueParams['page'] = progressBar.n + 1
            repositoryIssuePage = getApi(f'repos/{REPOSITORY_FULL_NAME}/issues', repositoryIssueParams)
            if repositoryIssuePage == []:
                break
            repositoryIssues += repositoryIssuePage
            progressBar.update(1)

repositoryIssuesStr = json.dumps(repositoryIssues, indent = 4)
#print(repositoryIssuesStr)
import pyperclip
pyperclip.copy(repositoryIssuesStr)

Using .url is too long to fit the whole line on the screen.

Above script does not seem to copy to clipboard on Debian 12 GNOME (this is probably due to Wayland) at least in Pyzo but seems to work fine when run in a gnome-terminal. In fact it does not work in a terminal too, current workaround consists in uncommenting print and executing in terminal and piping to xclip -selection clipboard (clip alias in my case).

Would help Community_detection_in_graphs/issues/10.

Related to Improve_websites_thanks_to_open_source/issues/680.

Note the new enum Exact and Fuzzy:

What about attachments not mentioned in the body? Can we deduce to what issue comment they were uploaded to?

Benjamin-Loison / gitea