#cofacts

2021-06-01
booco1999 00:47:38
@booco1999 has joined the channel
2021-06-02
github 12:59:31
⚠️ *Action required - upgrade app for g0v* ⚠️
github 12:59:32
*Pull Request Test Coverage Report for <https://coveralls.io/builds/40227976|Build 1360>* • *1* of *5* *(20.0%)* changed or added relevant lines in *2* files are covered. • No unchanged relevant lines lost coverage. • Overall coverage decreased (*-0.3%*) to *86.691%* * * * * * * *:yellow_heart: - <https://coveralls.io|Coveralls>*
rebuild 兩次才成功 😾

```Pulling mongo-container (mongo:3.6)...

ERROR: toomanyrequests: You have reached your pull rate limit. You may increase the limit by authenticating and upgrading: https://www.docker.com/increase-rate-limit

Waiting for Redis ... Error: No such container: redis

Error: No such container: redis```
QQ
關於這個 PR
我想改成 mongodb
但那個檢查 context 還在不在的 webhook change 還是要進
Github action 放好了。
Travis 一直卡在 docker-compose up 所以我先關掉了,用 github action 取代~
所以下面這兩個 PR 可以 review 囉(其實也放了 2X 天了⋯⋯) - https://github.com/cofacts/rumors-line-bot/pull/253 - https://github.com/cofacts/rumors-deploy/pull/20
travis 裡面 docker-compose up 似乎特別容易遇到 `toomanyrequests`
其實我們應該用 travis 的 redis services 而不是自己寫 docker-compose script 自己拉 XD
mrorz 13:00:08
請問一下,Cofacts的Open Data,什麼時候會更新一次呢?
時間不固定唷,不過確實有一陣子沒有更新了
請問一下最近有可能更新嗎?
@dz352 今天更新囉~時間應該是到 6/6 清晨定期備份時間
請再到 Cofacts 的 Dataset 資料夾(google drive)查看
這邊則是有用 Data set 做的一些基礎數字
可以看到 5/13 之後每天送進來的訊息數暴增
https://datastudio.google.com/u/0/reporting/18J8jZYumsoaCPBk9bdRd97GKvi_W5v-r/page/ckUQ
非常感謝!
nonumpa 13:03:28
rebuild 兩次才成功 😾

```Pulling mongo-container (mongo:3.6)...

ERROR: toomanyrequests: You have reached your pull rate limit. You may increase the limit by authenticating and upgrading: https://www.docker.com/increase-rate-limit

Waiting for Redis ... Error: No such container: redis

Error: No such container: redis```
mrorz 13:29:18
QQ
關於這個 PR
我想改成 mongodb
但那個檢查 context 還在不在的 webhook change 還是要進
2021-06-03
Andy Zhao 06:38:29
請問一下,Cofacts的Open Data,什麼時候會更新一次呢?
mrorz 12:11:58
時間不固定唷,不過確實有一陣子沒有更新了
mrorz 20:01:41
Replied to a thread: 2021-06-02 12:59:32
Github action 放好了。
Travis 一直卡在 docker-compose up 所以我先關掉了,用 github action 取代~
所以下面這兩個 PR 可以 review 囉(其實也放了 2X 天了⋯⋯) - https://github.com/cofacts/rumors-line-bot/pull/253 - https://github.com/cofacts/rumors-deploy/pull/20
mrorz 20:02:43
travis 裡面 docker-compose up 似乎特別容易遇到 `toomanyrequests`
其實我們應該用 travis 的 redis services 而不是自己寫 docker-compose script 自己拉 XD
2021-06-04
github 03:32:05
⚠️ *Action required - upgrade app for g0v* ⚠️
github 03:34:11
Just curious. Why do you pack graphql schema in the bundle pipeline, but not in the development? In my opinion, the schema is part of API structure code, which should be part of the codebase.
2021-06-05
github 03:33:03
⚠️ *Action required - upgrade app for g0v* ⚠️
github 03:33:04
&gt; Just curious. Why do you pack graphql schema in the bundle pipeline, but not in the development? In my opinion, the schema is part of API structure code, which should be part of the codebase. Good question. I was also thinking the same thing. Including the remote GraphQL schema in the codebase helps developer understand the *expected* API. It also helps keep track of the API change in LINE bot code base, after API server is changed. I should include the remote GraphQL schema directly in the codebase in future PRs.
github 03:34:27
Default branch changed back to `dev`.
2021-06-06
mrorz 01:22:25
等等 1:30AM 我要來重新啟動 production 的 docker engine 讓 cloud watch 的設定生效
希望 downtime 可以越短越好
mrorz 01:30:17
go
mrorz 01:32:38
mrorz 01:34:17
done
mrorz 01:38:16
Github action 前一個 step 用 curl 抓下來的檔案,在下一個 step 不會保留的樣子⋯⋯
`.env` 完全沒抓到 囧
image.png
沒設定 artifacts ?
確實沒設定 XD
但 ci.yml 裡 `npm ci` 之後,也沒針對 `node_modules` 設定 artifacts<
artifact 看起來是跨 job 用的,但我是同一個 job 下的不同 step。同一個 job 的 step 都會在同一台 runner 跑
我在想是不是更低級的錯誤,我打錯設錯之類的囧
重設 secret 就有抓到了
果然是低級錯誤囧
kiang 01:39:55
我的情況是排程 10 分鐘執行一次但實際上 30-60 分鐘才會跑一次,所以放棄了 XD
mrorz 01:43:49
為了 npm 7 使用 `node:16` 結果 pm2 好像會吐怪東西 囧
image.png
Andy Zhao 05:21:14
請問一下最近有可能更新嗎?
yhsiang 11:45:31
沒設定 artifacts ?
mrorz 12:43:31
確實沒設定 XD
但 ci.yml 裡 `npm ci` 之後,也沒針對 `node_modules` 設定 artifacts<
mrorz 13:31:32
Replied to a thread: 2021-06-02 13:00:08
@dz352 今天更新囉~時間應該是到 6/6 清晨定期備份時間
請再到 Cofacts 的 Dataset 資料夾(google drive)查看
  • 🦒1
mrorz 13:34:47
這邊則是有用 Data set 做的一些基礎數字
可以看到 5/13 之後每天送進來的訊息數暴增
https://datastudio.google.com/u/0/reporting/18J8jZYumsoaCPBk9bdRd97GKvi_W5v-r/page/ckUQ
mrorz 15:33:32
artifact 看起來是跨 job 用的,但我是同一個 job 下的不同 step。同一個 job 的 step 都會在同一台 runner 跑
我在想是不是更低級的錯誤,我打錯設錯之類的囧
mrorz 15:41:01
重設 secret 就有抓到了
果然是低級錯誤囧
github 17:01:16
⚠️ *Action required - upgrade app for g0v* ⚠️
github 17:01:18
Previously, docker images built by Github actions are always in English. `console.log` in `babel.config.js` always detects `en_US` locale: <https://user-images.githubusercontent.com/108608/120918743-8ae66780-c6e8-11eb-80a3-b51a29b8cdf6.png|圖片> The root cause is that `actions/checkout@v2` action will remove `.env` file. This PR moves `.env` file download to after `actions/checkout@v2` and the build process can correctly pick up the env file: <https://user-images.githubusercontent.com/108608/120918792-cbde7c00-c6e8-11eb-992b-375cec4fe3ba.png|圖片>
mrorz 17:51:08
github 18:20:07
When starting up the server under Node.js 16, it will emit the following warnings: &gt; PM2 error: TypeError [ERR_INVALID_ARG_TYPE]: The "data" argument must be of type string or an instance of Buffer, TypedArray, or DataView. Received type number (18) This is fixed by <https://github.com/Unitech/pm2/issues/4685|Unitech/pm2#4685>. &gt; (node:29) Warning: Accessing non-existent property 'findOne' of module exports inside circular dependency This is fixed by <https://developer.mongodb.com/community/forums/t/warning-accessing-non-existent-property-count-of-module-exports-inside-circular-dependency/3183|https://developer.mongodb.com/community/forums/t/warning-accessing-non-existent-property-count-of-module-exports-inside-circular-dependency/3183> Also fixes npm audit warnings.
github 18:37:07
Deployed to staging, working fine. Will merge as there is no feature added.
mrorz 19:14:04
line-bot.cofacts.tw (on Linode)上線了。

預計凌晨來執行 LINE bot server 切換到 line-bot.cofacts.tw 的作業,切換 SOP: https://g0v.hackmd.io/8VGFKx-cR8yEdKlAQ3vv5A?view#Deploy-SOP
  • 🦒1
目前進度:
> Update `master` to latest branch
撞到這個鳥 bug 所以 downtime 多了一點
https://github.com/go-redis/redis/issues/1343
現在要來切換 URL 了
切好了
設定通通完成。
就看這幾天 run 下去 load 是否撐得住,沒問題的話 heroku production 就可以關成 free dyno 囉~
搬家後目前還沒收到 `Messaging API: Your server did not return [200 OK]` 的信

凌晨有發生幾次 502,那是因為我當時在改 env var 後重開
LINE bot 的 resident memory 大概是 100 多 MB,比之前在 heroku 上觀測到得還少
今天一整天還是有幾次 request timeout 的狀況
目前 web concurrency (pm2 跑的 node.js 數量) 是 2
而我們機器是 6 core,要跑 API, elasticsearch, 網站
我來把 concurrency 調成 3,明天再來觀察
@info.cchsu has joined the channel
2021-06-07
mrorz 00:52:54
目前進度:
> Update `master` to latest branch
撞到這個鳥 bug 所以 downtime 多了一點
https://github.com/go-redis/redis/issues/1343
mrorz 00:53:30
現在要來切換 URL 了
mrorz 01:06:39
切好了
mrorz 01:06:47
發現這個怪 log
image.png
目前群組功能故障中 囧
在群組裡面發一次這個,就會觸發此 error
網路上都說這是因為 `.process` 定義了 named processor 但 add 的時候沒有附上 name 所致

但我們 `.process()` 也沒附上名字呀怪
啊我想我知道為啥了⋯⋯
https://github.com/cofacts/rumors-line-bot/blob/dev/src/webhook/handlers/groupHandler.js#L22
這裏 `process.env.JOBQUEUE_CONCURRENCY` 沒有轉數字,所以如果設定了 `JOBQUEUE_CONCURRENCY` ,`concurrency` 會變成字串。

所以 L27 的時候,bull 就會把那個字串 `concurrency` 當成 processor name 然後定義了一個 named processor⋯⋯
mrorz 01:23:46
目前群組功能故障中 囧
mrorz 01:37:01
在群組裡面發一次這個,就會觸發此 error
mrorz 01:38:00
網路上都說這是因為 `.process` 定義了 named processor 但 add 的時候沒有附上 name 所致

但我們 `.process()` 也沒附上名字呀怪
mrorz 01:39:37
啊我想我知道為啥了⋯⋯
mrorz 01:42:10
https://github.com/cofacts/rumors-line-bot/blob/dev/src/webhook/handlers/groupHandler.js#L22
這裏 `process.env.JOBQUEUE_CONCURRENCY` 沒有轉數字,所以如果設定了 `JOBQUEUE_CONCURRENCY` ,`concurrency` 會變成字串。

所以 L27 的時候,bull 就會把那個字串 `concurrency` 當成 processor name 然後定義了一個 named processor⋯⋯
github 01:46:43
Currently when `JOBQUEUE_CONCURRENCY` env is explicitly set, group chat handler will not work. Every time a message is send to group, it emits errors as below: <https://user-images.githubusercontent.com/108608/120934669-d6713380-c731-11eb-9763-2249a453fe7c.png|image> The root cause is that when `JOBQUEUE_CONCURRENCY` env var is given, it will be a string , which will be interpreted as a <https://github.com/OptimalBits/bull/blob/develop/REFERENCE.md#queueprocess|processer name> in <https://github.com/cofacts/rumors-line-bot/blob/dev/src/webhook/handlers/groupHandler.js#L22|https://github.com/cofacts/rumors-line-bot/blob/dev/src/webhook/handlers/groupHandler.js#L22> . When the processor is registered as a named processor, calling `.add()` without "name" will <https://github.com/OptimalBits/bull/issues/1529#issuecomment-548402563|cause the "missing handler for *default*" error>. Current workaround: do not set `JOBQUEUE_CONCURRENCY` in env var.
mrorz 01:52:40
Replied to a thread: 2021-06-06 19:14:04
設定通通完成。
就看這幾天 run 下去 load 是否撐得住,沒問題的話 heroku production 就可以關成 free dyno 囉~
  • 🙏2
Andy Zhao 03:31:57
非常感謝!
mrorz 10:19:09
搬家後目前還沒收到 `Messaging API: Your server did not return [200 OK]` 的信

凌晨有發生幾次 502,那是因為我當時在改 env var 後重開
mrorz 10:21:19
LINE bot 的 resident memory 大概是 100 多 MB,比之前在 heroku 上觀測到得還少
2021-06-08
mrorz 00:52:44
Replied to a thread: 2021-06-06 19:14:04
今天一整天還是有幾次 request timeout 的狀況
目前 web concurrency (pm2 跑的 node.js 數量) 是 2
而我們機器是 6 core,要跑 API, elasticsearch, 網站
我來把 concurrency 調成 3,明天再來觀察
image.png
  • 🦒1
JaneW 02:33:47
@jnes41023 has joined the channel
2021-06-09
Sephora Sephora 11:16:46
@kinkojiang has joined the channel
goodjack 小克 11:16:51
@goodjack has joined the channel
Gin Liu 14:26:14
@gingirlliu has joined the channel
Gin Liu 14:47:08
大家好,我是 Gin, 我們團隊正在開發一個在地社群的服務,裡面會有疫情相關的用戶發文、留言。由於 cdc 說建議我們最好可以針對用戶發布的內容做事實查核,因此我們有幾個問題想問:
1. 是否有 api 可以讓我們回傳用戶發布內容到 cofacts,然後根據 cofacts 回應的結果,決定是否移除內容
2. 如果上述成立,在資料互傳的時候,是否有需要注意的事項,比如待查證內容的篩選方式、品質、數量
3. 待查證的內容,通常需要多少時間才會完成查證?(可能影響我們是否需要增加其他把關方法)
希望能夠得到回覆!感謝各位大大!
想問一下這個「疫情相關的用戶發文、留言」具體是在什麼樣的情境發生的呢?Cofacts 資料庫這裡希望搜集的會是實際在社群網站流傳的可疑訊息唷,因為這樣才有後續查詢的效益,也比較不會浪費寶貴的編輯資源在未來不會有人再詢問的訊息上。

Cofacts 有 API 可以查詢已知的訊息。若有新的訊息想要回報,目前比較方便的作法可能是,人工確認這是「在網路上有在轉發的訊息」後,再手動傳進 LINE bot 傳進資料庫建檔。若查詢數累積了一定程度,也比較容易有編輯看到並且進行查證。

Cofacts 目前其實最缺的會是撰寫闢謠回應人手,所以我想「通常需要多少時間完成」這件事情無法給任何數字,而且也無法確定群眾的回應是不是符合你們的期待。

如果查證訊息的方面可以一起投入人力來做的話就更好了唷!闢謠教學: https://cofacts.tw/tutorial
「疫情相關的用戶發文、留言」這是用戶主動在我們的社群服務中發布的,有可能是用戶從其他地方分享過來的連結,也有可能是用戶自己撰寫的內容
目前爬了 cofacts 的紀錄,也有發現最缺的是一起撰寫闢謠回應的人手,真希望我們也有能力可以在這部分幫忙一些忙!!!
mrorz 14:58:36
想問一下這個「疫情相關的用戶發文、留言」具體是在什麼樣的情境發生的呢?Cofacts 資料庫這裡希望搜集的會是實際在社群網站流傳的可疑訊息唷,因為這樣才有後續查詢的效益,也比較不會浪費寶貴的編輯資源在未來不會有人再詢問的訊息上。

Cofacts 有 API 可以查詢已知的訊息。若有新的訊息想要回報,目前比較方便的作法可能是,人工確認這是「在網路上有在轉發的訊息」後,再手動傳進 LINE bot 傳進資料庫建檔。若查詢數累積了一定程度,也比較容易有編輯看到並且進行查證。

Cofacts 目前其實最缺的會是撰寫闢謠回應人手,所以我想「通常需要多少時間完成」這件事情無法給任何數字,而且也無法確定群眾的回應是不是符合你們的期待。

如果查證訊息的方面可以一起投入人力來做的話就更好了唷!闢謠教學: https://cofacts.tw/tutorial
Gin Liu 15:01:34
「疫情相關的用戶發文、留言」這是用戶主動在我們的社群服務中發布的,有可能是用戶從其他地方分享過來的連結,也有可能是用戶自己撰寫的內容
Gin Liu 15:02:24
目前爬了 cofacts 的紀錄,也有發現最缺的是一起撰寫闢謠回應的人手,真希望我們也有能力可以在這部分幫忙一些忙!!!
mrorz 16:37:08
@mrorz set the channel description: g0v Cofacts 真的假的專案討論區;開發訊息回報與訊息查證之 LINE 聊天機器人。
官網:https://cofacts.tw

Cofacts crowd-sourced fact-checking chatbot.
2021-06-10
github 03:33:47
⚠️ *Action required - upgrade app for g0v* ⚠️
github 03:33:49
Currently, both staging LINE bot and production LINE bot are deployed using docker. • Staging: build &amp; push to `cofacts/rumors-line-bot:dev-tw` after `master` branch merge • Production: build &amp; push to `cofacts/rumors-line-bot:latest-tw` after `release/XXX` tag is pushed Therefore, we no longer need `dev` and `master` branch. We can just use `master` branch only. This PR: • replaces `dev` branch existence in code base • connects CI test to Coveralls
mrorz 03:34:10
To LINE bot developers:

我把 rumors-line-bot 的 main brain 從 dev 設回 master 囉!請見此 PR 的說明:
https://github.com/cofacts/rumors-line-bot/pull/261
Currently, both staging LINE bot and production LINE bot are deployed using docker. • Staging: build &amp; push to `cofacts/rumors-line-bot:dev-tw` after `master` branch merge • Production: build &amp; push to `cofacts/rumors-line-bot:latest-tw` after `release/XXX` tag is pushed Therefore, we no longer need `dev` and `master` branch. We can just use `master` branch only. This PR: • replaces `dev` branch existence in code base • connects CI test to Coveralls
Ada 09:11:48
@slack1087 has joined the channel
2021-06-11
8loser 01:09:05
@8loser has joined the channel
Anges 11:39:03
@weidesign00 has joined the channel
2021-06-13
github 01:29:06
⚠️ *Action required - upgrade app for g0v* ⚠️
github 01:29:08
As a prep work for <https://www.figma.com/file/DvmAQjMJCncuPORWKnljM1/Cofacts-website-MrOrz?node-id=3005%3A0|Article LIFF>, we are setting up storybook to demonstrate components. Built storybook: <https://cofacts.github.io/rumors-line-bot/|https://cofacts.github.io/rumors-line-bot/> <https://user-images.githubusercontent.com/108608/121784414-76bdd100-cbe6-11eb-85b4-fc453c684471.png|圖片> <https://user-images.githubusercontent.com/108608/121784477-b4baf500-cbe6-11eb-9d4c-ff71bbb70dd9.png|圖片>
mrorz 16:25:55
LIFF「看過的訊息」新外觀
使用之前 LIFF 的元素
字變大
image.png
  • 1
github 18:18:43
As discussed in <https://github.com/cofacts/rumors-line-bot/pull/253#issuecomment-854124426|previous code review comment>, we should consider Cofacts API graphql schema as part of the LINE bot code base.
github 18:33:17
As we move to larger redis instance, we no longer need this fix. 1. In the future we plan to pass context along with actions (only store full text in redis), which should further reduce access of redis keys. 2. As `lastScannedAt` should never expire, we should put it in mongodb instead of redis. Thus the solution proposed in this pull request is suboptimal, should be rejected.
github 18:42:04
*Pull Request Test Coverage Report for <https://coveralls.io/builds/40539150|Build 933010003>* • *0* of *0* changed or added relevant lines in *0* files are covered. • No unchanged relevant lines lost coverage. • Overall coverage remained the same at *87.013%* * * * * * * *:yellow_heart: - <https://coveralls.io|Coveralls>*
github 18:47:34
*Pull Request Test Coverage Report for <https://coveralls.io/builds/40539169|Build 933018419>* • *0* of *0* changed or added relevant lines in *0* files are covered. • No unchanged relevant lines lost coverage. • Overall coverage remained the same at *87.013%* * * * * * * *:yellow_heart: - <https://coveralls.io|Coveralls>*
github 21:24:46
• new svelte-loader only works with Webpack 5: svelte-loader#178 • upgrades all loaders
github 21:27:49
*Pull Request Test Coverage Report for <https://coveralls.io/builds/40539906|Build 933244265>* • *0* of *0* changed or added relevant lines in *0* files are covered. • No unchanged relevant lines lost coverage. • Overall coverage remained the same at *87.013%* * * * * * * *:yellow_heart: - <https://coveralls.io|Coveralls>*
2021-06-14
github 01:58:20
⚠️ *Action required - upgrade app for g0v* ⚠️
github 02:00:28
Upgrading storybook's webpack to v5 is also having the following issues: • Storybook 6.2's webpack v5 is still rc version, not stable yet • After making storybook support webpack v5, stories will dissapeear, which requires <https://github.com/storybookjs/addon-svelte-csf/issues/23|more configuration> to fix I decided that we should wait until storybook officially support webpack v5 before we jump into webpack v5. We are upgrading svelte-loader to `master` branch's version to mitigate the <https://github.com/sveltejs/svelte-loader/issues/178|original issue>. Closing this PR.
chihao 17:23:06
有問題 ✋ 想請問 cofacts article 網頁上寫「Line 詢問 n 次」這是指什麼呢?例如這則的網頁上寫「Line 詢問 657 次」
使用 line 上的chatbot來查詢這篇可疑訊息的次數
lucien ++ `line 上的chatbot` 指的是 cofacts?還是還有別的 chatbot 😅
是cofacts
了解 👍 lucien ++
要注意 UI 上的數字只顯示最近 31 天ㄛ
lucien 17:42:43
使用 line 上的chatbot來查詢這篇可疑訊息的次數
chihao 18:26:21
lucien ++ `line 上的chatbot` 指的是 cofacts?還是還有別的 chatbot 😅
lucien 18:49:56
是cofacts
chihao 18:51:51
了解 👍 lucien ++
2021-06-15
tinlans 02:50:11
@tinlans has joined the channel
mrorz 11:25:40
要注意 UI 上的數字只顯示最近 31 天ㄛ
2021-06-16
github 13:39:03
⚠️ *Action required - upgrade app for g0v* ⚠️
github 13:39:05
Discussion: <https://g0v.hackmd.io/giYWfM8yRkKghtPOG1EBig#%E4%BA%82%E5%9B%9E%E8%88%87%E9%80%9A%E7%9F%A5|https://g0v.hackmd.io/giYWfM8yRkKghtPOG1EBig#%E4%BA%82%E5%9B%9E%E8%88%87%E9%80%9A%E7%9F%A5> We should not notify LINE user of replies with negative feedbacks. `getArticlesInBatch()` in `src/scripts/lib.js` should only yield articles when there are new replies having positive or zero feedbacks (positive - negative score &gt;= 0).
mrorz 13:56:24
今日會議延到 10pm 開始唷
github 17:02:50
View details in Rollbar: <https://rollbar.com/mrorz/rumors-line-bot/items/431/|https://rollbar.com/mrorz/rumors-line-bot/items/431/> ``` TypeError: Cannot read property 'split' of undefined File "/srv/www/build/webhook/handlers/utils.js", line 500, in createHighlightContents for (let highlightPair of text.split('&lt;/HIGHLIGHT&gt;')) { File "/srv/www/build/webhook/handlers/initState.js", line 166, in &lt;unknown&gt; contents: (0, _utils.createHighlightContents)(highlight, text), File "&lt;anonymous&gt;", line unknown, in Array.map File "/srv/www/build/webhook/handlers/initState.js", line 141, in initState const articleOptions = edgesSortedWithSimilarity.map(({ File "&lt;anonymous&gt;", line unknown, in runMicrotasks File "node:internal/process/task_queues", line 96, in processTicksAndRejections File "/srv/www/build/webhook/handleInput.js", line 117, in async handleInput params = await (0, _initState.default)(params); File "/srv/www/build/webhook/index.js", line 241, in async processText result = await (0, _handleInput.default)(context, { File "/srv/www/build/webhook/index.js", line 160, in async singleUserHandler result = await processText(context, type, input, otherFields, userId, req); ``` <https://user-images.githubusercontent.com/108608/122190796-ba555b00-cec4-11eb-9825-55f3131b53aa.png|image> Seems that `text` can be `undefined` and break code below. *Text* &gt; 陳培哲說這1年多來參加了約50多次疫情會議,如果篩檢出新冠肺炎的高齡患者、糖尿病、腎臟病等慢性病者,立刻打新冠單株抗體,可以減少5到6成死亡率。「這我很早就講過了,他們充耳不聞,那些生命是可以挽救的,你就知道我的挫折有多深。這代表這個指揮中心的科學判斷與能力不足。」 &gt; &gt; 陳培哲指出,另一個疫情中心不夠科學專業的地方是「核酸檢測方式」。美國食藥局(FDA)有一套檢測系統,但台灣的病管局(CDC)卻自己弄另外一套,沒自動化、量能少又不凖,專家建議應採FDA的系充,但指揮中心根本不聽。「這次中國廣州疫情爆發,1000多萬人1個禮拜做完採檢,用我們手動方式要做到哪一年?」 &gt; <https://www.storm.mg/new7/article/3740414?fbclid=IwAR3KtouSJVdWjFM0VVA3dj7J-0kWBxxObEw0jNjpy87LpQil0fmMYrm0OzE|https://www.storm.mg/new7/article/3740414?fbclid=IwAR3KtouSJVdWjFM0VVA3dj7J-0kWBxxObEw0jNjpy87LpQil0fmMYrm0OzE> *Output* <https://user-images.githubusercontent.com/108608/122191184-14562080-cec5-11eb-84a2-b897d91f51fa.png|image>
github 19:02:19
Same as <https://github.com/cofacts/rumors-line-bot/issues/220|#220> ?
github 20:12:01
You are right. let's just use <https://github.com/cofacts/rumors-line-bot/issues/220|#220>.
mrorz 20:12:14
Replied to a thread: 2021-06-16 13:56:24
今日會議延到 10pm 開始唷
2021-06-17
github 08:54:35
GitHub app is successfully upgraded in your workspace 🎉
To receive notifications in your private channels, you need to invite the GitHub app `/invite @GitHub`
2021-06-18
mrorz 02:29:09
週三開會覺得系統穩定
結果凌晨:
image.png
  • 🤔1
nginx access log 與 error log 相對應的時間也都沒有 502 log
mrorz 03:34:27
nginx access log 與 error log 相對應的時間也都沒有 502 log
2021-06-19
mrorz 01:13:08
使用者一多,計算上次 notification 送到哪的 lastScannedAt 就被洗掉 囧
image.png
北農的其中一個版本也害 #220 error 一天內噴了400 多次..

看起來真的該修了⋯⋯
@acerxp511 最近有空嗎 QQ
我可以先讓 linebot 不壞掉
針對 `lastScannedAt` 一直被刪掉
我做了個爛招
就是加一筆 cronjob 每一小時去讀一次 `lastScannedAt`
希望這樣他就不會被 LRU 掃到⋯⋯囧
nonumpa 01:36:01
北農的其中一個版本也害 #220 error 一天內噴了400 多次..
2021-06-20
github2 02:32:47
[cofacts/rumors-line-bot] Pull request opened by MrOrz
• Remove global material design styles, which is not currently in use • Add Cofacts palette as CSS custom properties • Implement `Button` and `TextArea` • Add SVG Icons
:white_check_mark: No checks have passed
github2 02:34:46
[cofacts/rumors-line-bot] Pull request opened by MrOrz
• Reduces bundle size from XXX to OOO • Apply common style to match <https://www.figma.com/file/DvmAQjMJCncuPORWKnljM1/Cofacts-website-MrOrz?node-id=3005%3A0|new article page>
:white_check_mark: No checks have passed
github2 23:52:04
[cofacts/rumors-line-bot] Comment on #267
*Pull Request Test Coverage Report for <https://coveralls.io/builds/40725306|Build 954714778>* • *0* of *0* changed or added relevant lines in *0* files are covered. • No unchanged relevant lines lost coverage. • Overall coverage remained the same at *87.013%* * * * * * * *:yellow_heart: - <https://coveralls.io|Coveralls>*
2021-06-21
github2 01:22:45
[cofacts/rumors-line-bot] Comment on #268
*Pull Request Test Coverage Report for <https://coveralls.io/builds/40725634|Build 954836834>* • *0* of *0* changed or added relevant lines in *0* files are covered. • No unchanged relevant lines lost coverage. • Overall coverage remained the same at *87.013%* * * * * * * *:yellow_heart: - <https://coveralls.io|Coveralls>*
mrorz 09:51:02

看起來真的該修了⋯⋯
@acerxp511 最近有空嗎 QQ
nonumpa 10:03:00
我可以先讓 linebot 不壞掉
github2 11:19:54
[cofacts/rumors-line-bot] Pull request opened by nonumpa
Changes 1. Fixes <https://github.com/cofacts/rumors-line-bot/issues/220|#220>, return original text if highlight text and hyperlinks are null We still need to fix this issue in rumors-api 2. Update `package-lock.json`, `lockfileVersion`: 1 -&gt; 2 Snapshot <https://user-images.githubusercontent.com/6376572/122703219-a7a3a300-d283-11eb-821e-2571c5f6b73f.png|截圖 2021-06-21 上午11 22 10>
:white_check_mark: All checks have passed
github2 11:20:24
[cofacts/rumors-line-bot] Comment on #220
Two more cases: ``` "6/18 立法院實錄~ 民進黨立委於立法院否決4000萬劑進口國際認疫苗的法案!全民(包括1450+817)如何看這個黨?<https://youtu.be/nSbLL5ZtDbo> ``` ``` 小心囉,北農18名新冠確診,四樓營業部重災區,每天包20-30萬包蔬菜給全聯,前幾天就有了,但因為蔬果是民生必需品,故沒公佈,今天爆18人,大概是瞞不住了,暫時不要去全聯! <https://udn.com/news/story/120940/5537424?from=udnamp_storysns_line> ```
github2 11:22:31
[cofacts/rumors-line-bot] Comment on #269
*Pull Request Test Coverage Report for <https://coveralls.io/builds/40730347|Build 955810650>* • *2* of *2* *(100.0%)* changed or added relevant lines in *1* file are covered. • No unchanged relevant lines lost coverage. • Overall coverage increased (+*0.09%*) to *87.107%* * * * * * * *:yellow_heart: - <https://coveralls.io|Coveralls>*
github2 11:36:39
[cofacts/rumors-line-bot] MrOrz commented on nonumpa's pull request
Thanks a million on the fix! I wonder why package-lock.json would change in the PR, it seems that we did not change `package.json`. Suggest exclude that from the PR and we are good to go!
github2 11:45:27
[cofacts/rumors-line-bot] Comment on #269
<https://github.blog/2021-02-02-npm-7-is-now-generally-available/#changes-to-the-lockfile|https://github.blog/2021-02-02-npm-7-is-now-generally-available/#changes-to-the-lockfile> npm changes its lockfile version after v7. If you use npm@7 to run `npm install` you'll get a version change in your lockfile unless you run `npm install --no-save`.
github2 11:46:55
[cofacts/rumors-line-bot] Comment on #269
&gt; Thanks a million on the fix! &gt; &gt; I wonder why package-lock.json would change in the PR, it seems that we did not change `package.json`. Suggest exclude that from the PR and we are good to go! I just switched to master from a old branch and run `npm i`. I thought it would have dependency error if I don't commit the change. (seems not)
github2 13:16:11
[cofacts/rumors-line-bot] MrOrz approved nonumpa's pull request
Thanks! Let's ship it <https://github.githubassets.com/images/icons/emoji/shipit.png|:shipit:>
github2 13:16:21
[cofacts/rumors-line-bot] Pull request merged by MrOrz
github2 13:20:07
[cofacts/rumors-line-bot] Comment on #269
Seems that we ~are also changing `lockVersion` to 2 in <https://github.com/cofacts/rumors-line-bot/pull/262|this PR>. After its merge we should have a stable lockfile.~ have already changed `lockfileVersion` to 2 in <https://github.com/cofacts/rumors-line-bot/pull/259|a previously merged PR>.
github2 19:45:43
[cofacts/rumors-line-bot] Comment on #269
Sorry, my `lockfileVersion` was from 2 to 1. It seems that running `npm install` through npm@6 (my local version) will change `lockfileVersion` back to 1.
mglee 22:30:21
Cofacts 的朋友大家好!我是研究g0v的人類學家 MG,去年有在這裡跟大家分享我論文中關於 Cofacts 的一章。我最近將它改為期刊論文,大幅度地修改了前言與結論,準備七月投稿,同時也會在延期至暑假尾巴的台灣 STS 年會中發表。草稿在此,歡迎大家上來評論、給意見~
https://docs.google.com/document/d/1Sqpdt_iS8-f41Cz11uGkr8I1dZpHtpsVC440MktXeH4/edit?usp=sharing
  • 👍9
  • 1
  • 1
This is cool
我也在做Cofacts相關的研究,之後會仔細拜讀這篇論文
Andy Zhao 22:38:02
This is cool
Andy Zhao 22:38:36
我也在做Cofacts相關的研究,之後會仔細拜讀這篇論文
2021-06-22
mrorz 10:55:12
針對 `lastScannedAt` 一直被刪掉
我做了個爛招
就是加一筆 cronjob 每一小時去讀一次 `lastScannedAt`
希望這樣他就不會被 LRU 掃到⋯⋯囧
mrorz 16:56:27
有使用者吵起來了
幸好 reply request 一人只能發一個,佔據版面有限
https://cofacts.tw/article/14qezlzjtt9ej
是不是使用者變多了,人多的地方就會吵架 😛
應該是的
我想定調為非公開發言 + 撰文者個人意見
我覺得這種討論是好事
新的查核我覺得清楚
每一篇新查核都是因為有舊查核先探路收真實回饋之後,才有辦法寫出來的呢
tmonk 18:21:05
是不是使用者變多了,人多的地方就會吵架 😛
2021-06-23
mrorz 03:08:26
應該是的
我想定調為非公開發言 + 撰文者個人意見
lucien 03:25:59
我覺得這種討論是好事
mrorz 10:45:47
請問 @acerxp511 當時修 https://github.com/cofacts/rumors-line-bot/pull/269 的時候呀,staging 上有什麼可以觸發這個 error 的訊息嗎
我傳了一個測試,發現 staging 沒有那個訊息,就改 production 了:p
也許把我截圖上 match 到的訊息都丟到staging 就可以觸發了,我試試
Try this
```小心囉,北農18名新冠確診,四樓營業部重災區,每天包20-30萬包蔬菜給全聯,前幾天就有了,但因為蔬果是民生必需品,故沒公佈,今天爆18人,大概是瞞不住了,暫時不要去全聯!

https://udn.com/news/story/120940/5537424?from=udnamp_storysns_line```
有了!
感謝感謝
我來試試送其他訊息
看起來 trigger 條件是
1. 先送個不含 fbclid 不含奇怪 query 字串的 URL 進去 cofacts 資料庫
2. 查詢有 fbclid 的 URL + 一些文字
上面這些是還沒更新 staging 的樣子。
我要來 deploy 修正版到 staging 上囉~~
fly 12:32:43
++
mrorz 12:54:27
請問 @acerxp511 當時修 https://github.com/cofacts/rumors-line-bot/pull/269 的時候呀,staging 上有什麼可以觸發這個 error 的訊息嗎
nonumpa 12:56:49
我傳了一個測試,發現 staging 沒有那個訊息,就改 production 了:p
nonumpa 12:59:43
也許把我截圖上 match 到的訊息都丟到staging 就可以觸發了,我試試
nonumpa 13:11:24
Try this
```小心囉,北農18名新冠確診,四樓營業部重災區,每天包20-30萬包蔬菜給全聯,前幾天就有了,但因為蔬果是民生必需品,故沒公佈,今天爆18人,大概是瞞不住了,暫時不要去全聯!

https://udn.com/news/story/120940/5537424?from=udnamp_storysns_line```
mrorz 13:23:41
有了!
感謝感謝
mrorz 13:23:53
我來試試送其他訊息
mrorz 13:34:09
看起來 trigger 條件是
1. 先送個不含 fbclid 不含奇怪 query 字串的 URL 進去 cofacts 資料庫
2. 查詢有 fbclid 的 URL + 一些文字
mrorz 13:34:48
上面這些是還沒更新 staging 的樣子。
我要來 deploy 修正版到 staging 上囉~~
mrorz 14:13:23
Johnson Liang
(時效已過)
Cofacts 的敘述已經更新了,大家可以看一下有沒有要改的~
mrorz 14:52:54
Cofacts 的敘述已經更新了,大家可以看一下有沒有要改的~
github2 15:52:46
[cofacts/rumors-line-bot] Pull request ready for review by MrOrz
github2 16:22:49
[cofacts/rumors-line-bot] bil4444 approved MrOrz's pull request
github2 19:57:05
[cofacts/rumors-line-bot] Pull request ready for review by MrOrz
github2 20:31:30
[cofacts/rumors-line-bot] bil4444 approved MrOrz's pull request
2021-06-24
JaneW 01:16:52
新的查核我覺得清楚
mrorz 01:45:49
每一篇新查核都是因為有舊查核先探路收真實回饋之後,才有辦法寫出來的呢
2021-06-26
nonumpa 09:45:23
怎麼把 cofacts logo 畫成長頸鹿了😆
chihao 09:52:54
可以從高處俯瞰地面上忙碌工作的謠言們
mrorz 09:56:48
今日共筆~~ cofacts.tw/hack
github2 11:05:36
[cofacts/rumors-site] Comment on #434
  • 1
github2 12:20:32
[cofacts/rumors-site] Comment on #434
感謝 <https://github.com/luoyu54321|@luoyu54321> ! Setup 有問題的話可以隨時在 Gather 發問唷~
fly 12:50:32
https://cofacts.tw/article/1vl6meo7cjmk4 有機器留言的樣子
fly 12:50:54
image.png
github2 13:41:04
[cofacts/takedowns] Pull request opened by MrOrz
感謝 @leo424y 回報!請大家 review 此篇刪除公告,沒問題我就執行囉。
已經刪除囉!感謝 @acerxp511 review
mrorz 13:41:51
感謝 @leo424y 回報!請大家 review 此篇刪除公告,沒問題我就執行囉。
github2 13:48:22
[cofacts/takedowns] nonumpa approved MrOrz's pull request
github2 13:48:49
[cofacts/takedowns] Pull request merged by MrOrz
mrorz 13:51:09
已經刪除囉!感謝 @acerxp511 review
github2 14:00:42
[cofacts/rumors-site] Pull request ready for review by luoyu54321
github2 14:33:31
[cofacts/rumors-site] MrOrz approved luoyu54321's pull request
<https://user-images.githubusercontent.com/108608/123504369-69dfb980-d68b-11eb-869b-05122096b707.png|image> Also confirmed on my laptop. Thanks a million!
github2 14:36:27
[cofacts/rumors-site] Comment on #434
Fixed in <https://github.com/cofacts/rumors-site/pull/437|#437>, kudos to <https://github.com/luoyu54321|@luoyu54321> !
irvin 16:32:48
@irvin has joined the channel
irvin 16:32:55
是 "DMCA" takedown XD
哎呀 XDDDD
github2 16:34:06
[cofacts/takedowns] Pull request opened by irvin
mrorz 16:38:05
哎呀 XDDDD
github2 16:47:22
[cofacts/takedowns] MrOrz approved irvin's pull request
Thanks for the fix! I have been wrong for years XD
github2 16:47:27
[cofacts/takedowns] Pull request merged by MrOrz
2021-06-27
pofeng (ocf) 18:04:01
感覺集團內部沒有溝通好xd
可以用 LINE 傳進來唷
pofeng (ocf) 2021-06-28 11:22:03
好, 我加一下 XD
是說那個蘋果報導好像 404 了
pofeng (ocf) 2021-06-28 11:32:15
pofeng (ocf) 2021-06-28 11:32:25
我剛剛送進去了
pofeng (ocf) 2021-06-28 11:35:06
( web 還查不到 )
pofeng (ocf) 2021-06-28 11:36:29
但是很多人在 FB 轉了, 我很少用 LINE , 不知道 LINE 的狀況
會不會是沒送成功呀,LINE bot 有回檢視訊息的連結嗎 @@
理論上有送進資料庫的話 web 一定搜尋得到才對,再不然就是在 cofacts.tw/articles 會看得到
pofeng (ocf) 2021-06-28 17:11:59
我看到的畫面
pofeng (ocf) 2021-06-28 17:12:41
我又再送一次, 還是沒在 web 介面看到
因為「在 LINE 外面看到」是我們的陷阱選項,把 LINE 以外的東西給導向到其他 bot XD
選項設計是為了讓 LINE bot 只收 LINE 上的內容。只要選前三個選項之一,就會送進資料庫了
pofeng (ocf) 2021-06-29 12:48:34
哦,了解,那我先來問問看有沒有人真的在 LINE 上面看過
感謝感謝
其實也是可以直接選前三個選項送進來 X
Peter 20:10:46
感覺集團內部沒有溝通好xd
2021-06-28
mrorz 10:18:21
可以用 LINE 傳進來唷
pofeng (ocf) 11:22:03
好, 我加一下 XD
mrorz 11:22:48
是說那個蘋果報導好像 404 了
pofeng (ocf) 11:32:15
pofeng (ocf) 11:32:25
我剛剛送進去了
pofeng (ocf) 11:35:06
( web 還查不到 )
pofeng (ocf) 11:36:29
但是很多人在 FB 轉了, 我很少用 LINE , 不知道 LINE 的狀況
mrorz 12:00:35
會不會是沒送成功呀,LINE bot 有回檢視訊息的連結嗎 @@
mrorz 12:04:09
理論上有送進資料庫的話 web 一定搜尋得到才對,再不然就是在 cofacts.tw/articles 會看得到
pofeng (ocf) 17:11:59
我看到的畫面
pofeng (ocf) 17:12:41
我又再送一次, 還是沒在 web 介面看到
mrorz 19:02:43
因為「在 LINE 外面看到」是我們的陷阱選項,把 LINE 以外的東西給導向到其他 bot XD
選項設計是為了讓 LINE bot 只收 LINE 上的內容。只要選前三個選項之一,就會送進資料庫了
2021-06-29
pofeng (ocf) 12:48:34
哦,了解,那我先來問問看有沒有人真的在 LINE 上面看過
mrorz 13:45:41
感謝感謝
其實也是可以直接選前三個選項送進來 X
github2 14:24:40
[cofacts/takedowns] Pull request opened by MrOrz
github2 14:25:58
[cofacts/takedowns] bil4444 approved MrOrz's pull request
github2 14:27:07
[cofacts/takedowns] bil4444 approved MrOrz's pull request
github2 14:35:34
[cofacts/takedowns] Pull request merged by MrOrz
irvin 16:11:49
@irvin has left the channel
林哲民 16:46:22
@calvin.lin has joined the channel
github2 20:29:54
[cofacts/rumors-line-bot] Comment on #267
github2 20:29:58
[cofacts/rumors-line-bot] Pull request merged by MrOrz
2021-06-30
github2 20:08:48
[cofacts/rumors-line-bot] Pull request merged by MrOrz
github2 23:06:52
[cofacts/rumors-line-bot] Pull request opened by MrOrz
Resolve the bug that users cannot submit reasons found in <https://g0v.hackmd.io/@mrorz/cofacts-meeting-notes/%2Fcjyi0ezPSzS2Dsgw1oBfmg|release test>. Refactor: change dev API URL to fix loading issue
:white_check_mark: No checks have passed