New to meteor, new to crawlers. I am trying to build a small Meteor project which is able to crawl a website (multiple pages) and click elements before grabbing the data. From initial research all I find is a "headless browser" such as phantomjs. Not sure how well Phantom ties into Meteor. Has anyone used Meteor in this manner?
Related
I'm using Firebase dynamic links for referrals in an Expo app (managed workflow), but the link data isn't surviving the app installation process. The links worked fine when testing with a standalone app: when the user didn't have the app, they got directed to the App Store/Play Store* and when they did have the app, the link was handled using Expo's Linking.getInitialUrl().
I've actually gone ahead and published a native binary that displays a modal with the link data upon signing up to the app** (again, using Linking.getInitialUrl()). The modal only displays the app's custom url scheme myapp://. This same test displayed the link as expected when testing with a standalone app.
Note that the links work as expected if the app is already installed.
*This answer seems to suggest that this is enough to ensure the data will pass
**Fortunately we are very early stage with no users lol
I have an RShiny app that I had published online without any issues until this week. The app uses data stored in a Google Drive account, thus making use of the googlesheets package. When the app was working fine, I used gs_auth() to initially authorize and cache my Google credentials. Because my credentials were cached in a .httr_oauth in my working directory, it meant that when I ran the app's code, it wouldn't take me to a page in my browser that would ask me to authorize the account. When publishing the app to shiny.io, this cached .httr_oauth meant that the app would work online.
This week I made some changes to the app's code and when I ran the app it took me to the Google account authorisation page. There, I would click on the account and it would then say: "Sign in with Google temporarily disabled for this app".
After some looking I found that the googlesheets package is now outdated and I should instead be using the googlesheets4 package. However, after trying to authorize with that package (and the googledrive package) I'm able to get the app to run locally without it taking me to the Google account authorization page, but when I try to publish it still doesn't work. It always times out. Does anyone know what is going on? Anyone else having a similar issue?
I ran into this problem as well.
This outline worked really well for me.
https://gargle.r-lib.org/articles/non-interactive-auth.html
I would focus on Project Level OAthu part of the vignette.
The biggest issues I had for switching from "googlesheet" to "googlesheets4" was I forgot to install the "googledrive" package. The googledrive will allow you to use (drive_auth) and then build out from there as you can cache the authentication token.
I have created a project (and invested significant time & effort on it :-) ) .
When I try to add add-ons it complaints about some missing bits and pieces. Looking around, the root cause seems to be that I created this with the default barebones boilerplate, instead of the richer bootstrap3 boilerplate.
However, there is no option I can find to change the boilerplate. Is it possible to change the boilerplate it is built on? Is there a magic concoction...?
An Aldryn employee helped me out in their chatroom with this problem -- if you're trying to make changes on your templates and whatnot locally, and they aren't being reflected on the site when you deploy the test server, you can get things re-synched by totally deleting your local files, opening the test server with the desktop client, and clicking on the "auto-sync this project" button. This takes what's on the test server, and rebuilds the local files on your computer with that. Then, your local changes (remember to save your changes first!) should be reflected when you deploy the test server (after clicking on the "Refresh Projects" button in the lower righthand corner of the desktop client).
If you have a ton of work that you'd rather not delete and rewrite, you can move your local files elsewhere on your hard drive, and then replace the files that are written from the test server with them.
Cheers to Angelo for helping me with that. It's kind of a quirky solution, and I'd have never figured it out on my own.
You can also add custom boilerplates here (I think that'll make you log in to your account to see it), but I haven't attempted a custom boilerplate, yet, so I can't vouch for how well one would jive with the rest of the Aldryn platform.
EDIT: I forgot to mention that, while the title is unlikely to grab the interest of somebody who's making their own boilerplates, the first two or three entries in this series could be quite helpful in implementing your custom work into Aldryn's framework.
I'm creating a package where I allow to connect multiple same service accounts and I need to force the user to login to the account at the popup window every time (do not authenticate automatically) twitter has the option for "force_login" parameter when using Meteor.loginWithTwittertwitter docs
I can't find any way of using it with meteor, I also looked into creating my own twitter/accounts-twitter package but it seems like it's even deeper some where inside oauth packages and it gets really messy to figure out where or how I can do that. Anyone have any ideas?
It was not possible yesterday, but it's possible today. I opened this pull request in the Meteor repository adding this functionality: https://github.com/meteor/meteor/pull/5693
Meteor.loginWithTwitter({ force_login: true }, ...);
Note it's not merged yet, but hopefully it will be soon.
I'm making a web app. Part of it includes the automatic generation of word documents (letters). Those documents need to be opened by the end user, then saved back to the server once they've finished editing.
I tried using webdav for this but the only browser I could actually launch word from (using active-x) was IE. I also found a plugin for firefox, but for Chrome I couldn't find a way that worked.
Next I thought about making the clients map a drive to webdav (or similar), then offering up the files as file://... links, but that only works if the webpage is on the local machine.
Before I resort to ditching word (the clients won't like this) and using CKEditor or TinyMCE, is there another way?
In short, I'd like to have links to a document on the page, which when clicked are opened in word, but the file should stay remote - and then when saving, it's the remote file that gets updated.
I've also looked at Zoho but it could be very expensive for this project, plus I don't think it can be white-labelled and also looks a bit old fashioned, UI wise.
Maybe Sharepoint can do what I need? Haven't looked at that much. Also thought of making a client app to run in the system tray and deal with things.
If there was a decent way of editing Word docs from within the browser with something like CKEditor/TinyMCE and once finished, conversion back to Word format actually worked 100%, that would suffice.
I'm open to totally fresh ideas if anyone has any...
Currently Chrome, Firefox and Safari support MS Office plugin. They can open and save documents directly to server. I have tested this with MS Office 2007 and MS Office 2007 just about a month ago.
Ideally, your users would be able to use Word natively. Is there any chance you could create an Office Add-In that hooks into the BeforeDocumentSave event, looks for some indicator that the file is associated with your application, and save the updated file to your server?
Saving to the server via the Word Add-In would probably need to include some unique identifier (in addition to file name), so you could overwrite the previous version server-side. Then, if you were using something like SignalR, you could trigger a refresh on the web page when the file was saved successfully (assuming they were still on that web page) on the server (via FileSystemWatcher).
Had same problem myself.
I solved it by setting up a webdav share on the server with digest authentication (SabreDAV), and tied it into the users table on my app backend.
In relation to the client end, I solved accessing this by creating a small java applet which uses the java Desktop class (getDesktop().open()) to open the file. You will need to make sure the path is handled right for the client machine type (Windows, OS X or Linux)
You will also need to get your end users to permanently mount or map the webdav share locally.