cofacts

Month: 2018-07

2018-07-04

mrorz 12:10:02
本日會議記錄
https://hackmd.io/s/rJBkfpKMQ

HackMD

20180704 會議記錄 - HackMD

20180704 會議記錄 ====== ## Dev update ### Release tonight! #### Version - API, site supporting 送出理由

mrorz 14:36:38
http://amp.poynter.org/news/wechat-rogue-fact-checkers-are-tackling-apps-fake-news-problem

Poynter

On WeChat, rogue fact-checkers are tackling the app's fake news problem

A handful of “guerrilla” debunking projects exist on WeChat, which has about 1 billion monthly active users worldwide.

2018-07-06

changhc84 22:18:53
@changhc84 has joined the channel

2018-07-07

Yi-Hua 11:28:18
@yihua1218 has joined the channel
yuhsuan 11:34:15
@yuhsuan has joined the channel
Sting 14:14:28
@sny831212 has joined the channel
Matt 14:32:13
@ycj1665 has joined the channel
Alan 15:56:39
@alanchanghsnu has joined the channel
bil 21:44:38
今天非常謝謝來寫扣扣發PR還有闢謠整理文章的大家:heart::heart::heart:

2018-07-09

Lulu Keng 17:32:17
@lulu523 has joined the channel
ttcat (not_staff) 17:32:38
要問 @hazelwei 翻譯進度以及我們要看最後要怎麼簽領據
mrorz 18:04:59
要問大家本週三晚上要不要開會
颱風看起來週三晚上就走了,也沒西南氣流

2018-07-10

mrorz 09:54:55
update: 會開會唷
chihao 09:57:11
是小聚嗎?
mrorz 09:57:41
是一般的開會呢
@chihao 要來嗎~~
mrorz 09:58:04
算是有畫面的
mrorz 09:58:34
待 3 小時取 10 秒的影片
mrorz 09:59:33
每週三 8pm ~ 11pm 在 Workis ( https://workis.space/ )
遲到早退可

Workis 工作是

Workis 工作是 – 勇者聚落 | 精神時光屋 | 工作空間

Workis是座落大安區巷內專注於「工作」的工作空間。是共同空間、咖啡廳、會議室、秘密基地、精神時光屋,有隱密的獨立包廂也有友善交流的開放空間,有策展空間也有不一定最專業但最用心煮的咖啡。在這裡我們希望你可以卸下平時的包袱,可以遇到一些可愛的人,可以重新找回自己。 Workis - 最硬也最溫暖的工作空間。

chihao 10:27:46
可以去取材嗎?g0v日常/g0v一句話 XD
chihao 10:48:47
啊明天晚上是OCF美國行分享會,如果取消或早結束我就衝一發cofact :muscle:
mrorz 10:52:31
應該不會取消也不會早結束 XD
可以下週來唷
chihao 17:15:25
所以是固定週三晚上開會的意思嗎?
mrorz 22:11:58
mrorz 10:52:46
但下週沒有 @bil 會比較寂寞 (?)
ttcat (not_staff) 20:12:58
上次工作坊的人在 signal 群組中問有沒有 cofacts 的人會加入群~
mrorz 22:12:17
在哪在哪
ttcat (not_staff) 14:20:02
你有 signal 嗎
mrorz 14:42:43
nope
ttcat (not_staff) 14:47:22
qq

2018-07-11

chihao 13:19:35
@han.kirstenly Cofact is here :slightly_smiling_face:
kirsten 13:19:39
@han.kirstenly has joined the channel
kirsten 13:20:30
Thank you!
kirsten 13:21:24
I actually already have a lot of the background info thanks to the workshop, but I’ll pop questions that I have here. :smile:
Yumi 13:24:18
@monorhapsody has joined the channel
mrorz 13:52:34
Welcome @han.kirstenly! good to see you here

2018-07-14

mrorz 23:48:00
抱歉發了個巨大的 PR,因為 branch 太久了途中一直 merge,如果 rebase 的話 package-lock.json 會超亂,
所以沒辦法拆 PR⋯⋯

https://github.com/cofacts/rumors-api/pull/76

總之這是 URL preview 的 API 部分,求 review :pray:

GitHub

Scrap website content from URL & leverage it in article search APIs by MrOrz · Pull Request #76 · cofacts/rumors-api

Sorry this PR has become so huge and the merge is so complex (merges 2 package.json upgrades) so that I cannot rebase & split PR at all. The major changes are: implement scrapUrls in src/util/...

wildjcrt 22:56:19
我本來想說我天天在用 rebase,可以幫忙弄一下,結果完全搞不定 package-lock.json :sweat_smile::upside_down_face:
mrorz 10:54:15
沒錯,PR 放太久 + 每個月會有 bot 自動發 package-lock 更新,反而讓 rebase 非常崩潰 QQ
wildjcrt 10:55:47
還有 submodule 那些 commit 也可以去掉 XD
通常 master 都比較新
wildjcrt 10:56:23
package-lock.json 因為是機器生出來的檔案,真的很難做 rebase
mrorz 11:00:51
總之我覺得比較好的解法就是直接 review 這個 PR ,然後盡快 merge⋯⋯ XD
wildjcrt 11:03:51
我在公司作法是,不大或可以獨立出的 commits,我會用 cherry-pick 拉成小 PR 去 merge。
wildjcrt 11:04:41
package-lock.json 每月更新的話,那就看 PR 是每月 rebase 一次,或是針對 package 先送 PR 進 master。
mrorz 11:21:28
有道理,merge package-lock 之後應該要把所有開著的 branch / PR 都 rebase 一次,才不會造成無法挽回的狀況 Orz
wildjcrt 14:10:38
rebase early, rebase often :joy:

2018-07-15

2018-07-16

mrorz 23:07:37
我們成功地提升了 Cancel submit article 的頻率,效果十分顯著。

希望大家不要再無腦地把 Cofacts 當 siri 啦~~
現在送出訊息時會問使用者「為什麼你會想送出這篇」
@null 23:08:51
@mrorz commented on @mrorz’s file https://g0v-tw.slack.com/files/U039CG5S7/FBQH3T8N4/_____________2018-07-16_______11.07.14.png|現在送出訊息時會問使用者「為什麼你會想送出這篇」: 「送出文章時要填寫理由」是2018/7/9 deploy 的,然後下面的 +183% 是指跟前面相同時間(8天)比起來
現在送出訊息時會問使用者「為什麼你會想送出這篇」

2018-07-17

mrorz 03:20:40
觀察先放在週三會議記錄:
https://hackmd.io/lGzzR6lcQjGgxjmIUMVPpQ

hackmd.io

20180718 會議記錄 - HackMD

2018-07-18

fly 09:41:59
https://www.poynter.org/news/meet-next-misinformation-format-fake-audio-messages

Poynter

Meet the next misinformation format: Fake audio messages

Over the past year, fake audio messages have been slowly making the rounds on WhatsApp.

2018-07-19

hcchien 20:53:21
@hcchien407 has joined the channel

2018-07-20

agameofprivacy 12:26:42
:face_with_monocle::thinking_face:

https://www.recode.net/2018/7/19/17594156/whatsapp-limit-forwarding-fake-news-violence-india-myanmar

Recode

WhatsApp will drastically limit forwarding across the globe to stop the spread of fake news, following violence in India and Myanmar

The social messaging service has been blamed for its role in spreading dangerous and inaccurate information.

agameofprivacy 12:28:21
在印度的五轉限制官方說是個Test...因為印度用戶是全世界最會轉的。
mrorz 13:54:22
> “We’re horrified by the violence in India, and we’ve announced number of different product changes to help address these issues,” a company spokesperson told Recode. “It’s a challenge which requires an action by civil society, government and tech companies.”

Civil society 表示:⋯⋯ 請問 whatsapp 有 API 嗎。

https://stackoverflow.com/questions/37078108/does-whatsapp-offer-an-open-api
mrorz 13:58:13
我覺得 whatsapp 的回應還滿荒唐的 (攤手
mrorz 14:07:07
Hmm 我讀了這個,感覺就算 whatsapp 開放 chatbot,好像也無法解決這個現象。
http://www.cna.com.tw/news/aopl/201807070091-1.aspx

> 調查說,上述私刑致死案中,多數被捕嫌犯未接受過教育或僅受過小學教育,多數是每天賭博或失業的年輕人。
> 馬哈拉什特拉省警局負責法律與秩序的警察總監畢哈利(Bipin Bihari)認為,當暴民是一群文盲、失業者和貧困人士時,常以暴力方式表現他們對政府、機構,或對整個社會的憤怒,這些被壓抑的憤怒在沒有身份識別的匿名群體中找到宣洩管道。

fact-check 對這樣的對象沒用呀⋯⋯
agameofprivacy 15:58:25
已經被flagged是謠言的資訊被平台以中立、言論自由之名一再地被傳播真的有一種在高速公路酒駕沒人攔的感覺。
ggm 16:47:07
五轉限制好像在玩遊戲,可以轉生或是轉職五次 … :video_game:

2018-07-23

canadian_aaron 23:37:09
@canadian_aaron has joined the channel

2018-07-24

mrorz 00:43:11
請問 @darkbtf 有沒有之前處理 s3 上 cofacts LINE bot raw log 的 code 呢
還是我自己寫一份會比較好?
mrorz 00:44:01
希望可以是吃一堆 txt --> 輸出一張 excel 之類的,欄位有 user id, input, output, context, timestamp 之類的
kirsten 12:35:45
Hi! I’m popping up again — thank you @mrorz for the answers on the Google docs! Unfortunately my word count was limited from my editor, or I would have liked to write heaps more about Cofacts. :slightly_smiling_face:

Checking in to find out if there are good photos that perhaps we could share of Cofacts and your work? I’ve looked in the Hackfoldr but thought it would be better to check with people first.
mrorz 13:38:01
Thanks for writing for us! :jump-g0v:

We have some photos in the official website https://cofacts.g0v.tw/ and the grant site https://grants.g0v.tw/power/en/ . You can pick some and I can send you the high-res version.
mrorz 14:22:53
@han.kirstenly As for the notes in Google doc, we still have some unanswered questions. Would you like us to finish them all?
kirsten 14:31:16
I think it should be okay! I’ll check with my editor, but I think he might actually be cutting things down—will let you know!

2018-07-25

mrorz 14:14:21
@ggm 我發現送出理由有一個 bug:
https://github.com/cofacts/rumors-line-bot/issues/86

我對此發了一個 PR,求 review https://github.com/cofacts/rumors-line-bot/pull/87

GitHub

Cannot submit reply request reason · Issue #86 · cofacts/rumors-line-bot

I have seen this conversation in the log: :bust_in_silhouette:User: Kalinda Grabar Kitarovic克羅埃西亞總統,善良、高雅、氣質... :robot_face:Bot: 【跟編輯說您的疑惑】↵抱歉這篇訊息還沒有人回應過唷!請指出您有疑惑之處,說服編輯這是一份應該被闢謠的訊息。 :bust_in_silhouette:User: 以英文關鍵字反查國外新聞,從未見到有這樣的內容。 :robot_face:Bot: 找不到關於「...

GitHub

Allow users to submit reply request resason by MrOrz · Pull Request #87 · cofacts/rumors-line-bot

Fixes #86. ASKING_REPLY_REQUEST_REASON should allow users to input > 3 letters.

ggm 15:16:40
噢噢有道理

2018-07-27

w3n9j 16:04:50
@wengj has joined the channel

2018-07-31

ggm 15:12:29
對吼我好像應該貼這邊 XD