add basic text file imports #172
Conversation
this adds basic support for importing files as blog posts. .txt and .md are supported at this time and the collection is selectable, defaulting to draft. if a collection is specified the post is federated.
|
I got a big sidetracked down the rabbit hole of #96, once I figured out what was getting in my way I remembered seeing something about the so there may be things I missed migrating to a fresh branch, but quick manual tests were successful. |
different template action for partial or complete import success
|
It's looking pretty good. I would prefer if the errors could be in a similar style box as the success alert, it would look a bit more polished. Other than that I might change the form to allow all plain text files, not just .txt and .md. Then verify they are plain text again on the server. Not everyone will use files extensions. |
|
Also, good idea on allowing all plain text files, and not just those with a file extension. |
|
Awesome, I'll fix these up shortly. I think I tried using |
handler for post request to import is now under /api/me/import form target updated also allow all plaintext files in form
this changes the import handler to use CreatePost instead of CreateOwnedPost which required generation of non expiring access tokens
|
@thebaer all set, tested on my end and still working |
in favor of library side generation to support zip files
this updates to parse the time from the imported file, using v0.1.1 of the wfimport library
now checking for and returning invalid content type errors
- Changes Import link location in dropdown menu - Makes design consistent with Invite People page (and extracts some common CSS into core.less) - Selects the user's first blog by default in the dropdown - Changes the copy a bit Ref T609
- Only retrieve a collection from database if an alias is submitted - Only call GetCollection() once (previously, it was inside the loop) - Return error if user doesn't own the collection Ref T609
This moves file operations inside the `for` loop into an anonymous func, so the `defer` calls don't wait until the end of the handler call to actually execute. Ref T609
|
Thanks for getting those things! After testing, I made a few changes. I tweaked the design and copy on the Import page to be consistent with the rest of the site. I also ran into problems when importing Draft posts -- whether from bad local data or something else, the I also reduced the number of database calls (previously, this was repeatedly making the same Next I'll probably tweak the stuff around temporary files a bit. |
ghost
self-assigned this
|
I had to use the last modified time as the File API does not provide created |
|
Cool, that's not a problem (and maybe better, anyway). Testing now. |
File API gives timestamp in milliseconds, not seconds, so this converts it correctly. Ref T609
This reverts commit 2b06699.
File API gives timestamp in milliseconds, not seconds, so this converts it on the client-side and sends it the correct time to the server. Ref T609
|
What information is supported with imports (authors, tags, etc.)? Is there official documentation on this somewhere? |
|
@heyakyra, built-in text file imports only support:
For more comprehensive imports, the API is probably the best way to migrate (documentation here) |
this adds basic support for importing files as blog posts.
.txt and .md are supported at this time and the
collection is selectable, defaulting to draft.
if a collection is specified the post is federated.
T609