Википедия:Заявки на статус бота/Michaello

Здесь находятся завершившиеся обсуждения. Просьба не вносить изменений.

Michaello

Botmaster: Михајло Анђелковић

Цель / Purpose

I request bot flag for my bot Michaello since I intend to let it do more intense work on this Wikipedia for the time coming. The flag would allow me to run the bot with higher frequency (an edit per 6 - 20s) without expecting each moment to get a warning to stop flooding the recent changes list. :) In the near future, on the ru.wp I would perform repairing of titles that contain words written by mixing Cyrillic and Latin letters. This activity at some point will extend to the article content since the needed mechanisms already exist (see below) in the bot. The critical element for this is my free time, needed to teach Michaello to behave properly on this project.

Michaello is already present on sr.wp and bg.wp where it completed lots of tasks, as follows.

Технические подробности / Technical details

• What exactly your bot do
  • Michaello is a general purpose bot, intended for a wide specter of tasks, from most simple ones to moderate ones. His late activities included Wikipedias on Serbian and Bulgarian language, where it was performing so-called chained filters. Particular tasks were:
    • Cleansing occurrences of cyrlat words (words written by using Latin and Cyrillic letters)
    • Formatting roman numbers (regarding to abusive usage of Cyrillic or cyrlats to write them)
    • Formatting measuring units and numbers bound to them (needed in sr.wp only)
    • Doing grammar corrections
  • The longs for the chained filters you can find here.
  • Other activities included:
    • Observing the recent changes on both Wikipedias via irc channel.
    • Marking improperly sent images and warning the senders
    • Moving categories, refreshing (purge or blank edit) all articles from some category
    • Performing various searches on Wikipedia XML dumps and Wikipedia itself.
  • At present, the bot has over 48000 edits on sr.wp and bg.wp together, and over 49000 edits in total. (visible via the SUL util tool linked on the top of the request)

• which language does it written in?
  • Michaello has two parts. The first part is used for searching XML dumps and it is written in C++ to provide maximal efficiency (at least on my mech :) ). The second part that appears on Wikipedia and IRC channels is written in Java. I am the author of the complete source.

• does it use pywikipedia framework?
  • No.

• Does it operate automatically or with human assistance?
  • Bot can mess up the things sometimes (yet, by following my instructions). But in that cases I use logs to track the problems down and be the first to fix them. Sometimes fixing requires manual intervention.

• Link to source code is encouraged.
  • The very core (without numerous add-ons) of the C++ part of the bot is available online. The Java-part of the bot isn't available online yet, for it is still under development.

Флаг бота в других виках? / Already has a bot flag?

Michaello has bot flag on sr.wp and bg.wp. (visible via the SUL util tool that has been linked on the top of the request).

Обсуждение / Discussion

Cleansing occurrences of cyrlat words is a very useful job. Please do some (20-50) edits to show us, how the bot works. — Obersachse 19:08, 23 ноября 2008 (UTC)

• «Cleaning occurrences of cyrlat words (words written by using Latin and Cyrillic letters)» — that’s great; «Doing grammar corrections» — do you know common russian mistakes? «Marking improperly sent images and warning the senders» — we already have a special bot working on this task. But cyrlat words — that’s really good. rubin16 19:24, 23 ноября 2008 (UTC)

Oh, I see that I missed to write more particularly what I intend to do here (fixed right away). My bad, in the answer on the question "What exactly your bot do" however, I gave a bot's portfolio in general. I will try to answer the risen questions:

1. For now I can only afford to fix the cyrlat titles here, which will take yet a not-small number of edits, having on mind that bot will have to de-reference cyrlat redirects locally before they can be deleted.
2. I intend to extend bot's activity here on the article content in the future, but let me explain how it works first. Every time bot starts work on a new Wikipedia (I experienced that on sr.wp and bg.wp) it requires some time to be taught "what is right and what is wrong on the new Wikipedia". According to this, it builds a database that prevents it repeat the mistakes from the past. Right now I lack in time to do this properly for ru.wp, so it would have to wait.
3. Examples of past cleanings you can see in the provided logs. Let me present some moderate edits: example 1, example 2 and example 3.
4. Grammar corrections on bg.wp have been done in correspondence with the staff there. They've gathered quite a collection of grammar corrections. For the sr.wp I build my own collection with the time.

項 Михајло ^{_[ talk _]} 20:07, 23 ноября 2008 (UTC)

• Nice examples, I think we really need such kind of a bot. rubin16 20:13, 23 ноября 2008 (UTC)

• За --Latitude 19:30, 25 ноября 2008 (UTC)
• За It would be great, I think.rlu 23:19, 25 ноября 2008 (UTC)

Итог / Summary

Bot flag granted. — Obersachse 13:42, 26 ноября 2008 (UTC)

Спасибо. :) 項 Михајло ^{_[ talk _]} 16:15, 26 ноября 2008 (UTC)