[SVCS-528] [Blocked] Improve gdoc rendering - WB Part #304
Conversation
| else: | ||
|
|
||
| # TODO figure out metadata.raw.get('downloadUrl') | ||
| download_url = metadata.raw.get('downloadUrl') or drive_utils.get_export_link(metadata.raw) # type: ignore |
There was a problem hiding this comment.
This assumes that metadata.raw.get('downloadUrl') is always None. Why?
There was a problem hiding this comment.
My understanding of this line is that downloadUrl is always present unless it's a file that that isn't exportable in it's native type, so a .txt will have a downloadUrl, but a .gdoc won't. Google files have frustratingly nonstandard metadata.
| @@ -234,9 +234,18 @@ def can_intra_copy(self, | |||
|
|
|||
| metadata = await self.metadata(path, revision=revision) | |||
|
|
|||
| if kwargs.get('mfr', None) and kwargs['mfr'].lower() == 'true': | |||
| download_url = drive_utils.get_alt_export_link(metadata.raw) # type: ignore | |||
There was a problem hiding this comment.
download_url = metadata.raw.get('downloadUrl') or drive_utils.get_alt_export_link(metadata.raw) 
cslzchen
force-pushed
the
feature/svcs-528-improve-gdoc
branch
from
4768410
to
38b4e1a
Compare
We can export .gdoc as a pdf if we know its coming from mfr. Added `alt_export` functions and properties to Gdrive utils and metadata. Gdrive download now looks for 'mfr' in its kwargs and will request file as pdf if used properly.
raises key error and update gdrive utils to use `dict.get()` instead of `dict[]`.
cslzchen
force-pushed
the
feature/svcs-528-improve-gdoc
branch
from
38b4e1a
to
a7c70ef
Compare
| @@ -234,9 +234,16 @@ def can_intra_copy(self, | |||
|
|
|||
| metadata = await self.metadata(path, revision=revision) | |||
|
|
|||
| if kwargs.get('mfr', None) and kwargs['mfr'].lower() == 'true': | |||
| download_url = metadata.raw.get('downloadUrl') or drive_utils.get_alt_export_link(metadata.raw) | |||
There was a problem hiding this comment.
FYI, fixed. Need to check metadata.raw.get('downloadUrl') first. There are file types on gdrive (e.g. .txt) that do not have export_links.
There was a problem hiding this comment.
what about
if kwargs.get('mfr', '').lower() == 'true':
Also I don't get the or here.
cslzchen
changed the title
cslzchen
changed the title
| @@ -234,9 +234,16 @@ def can_intra_copy(self, | |||
|
|
|||
| metadata = await self.metadata(path, revision=revision) | |||
|
|
|||
| if kwargs.get('mfr', None) and kwargs['mfr'].lower() == 'true': | |||
| download_url = metadata.raw.get('downloadUrl') or drive_utils.get_alt_export_link(metadata.raw) | |||
There was a problem hiding this comment.
what about
if kwargs.get('mfr', '').lower() == 'true':
Also I don't get the or here.
| if format['mime_type'] == metadata['mimeType']: | ||
| return format | ||
| for format_type in DOCS_FORMATS: | ||
| if format_type.get('mime_type') == metadata.get('mimeType'): |
There was a problem hiding this comment.
metadata.get('mimeType') should be outside the for loop so it doesn't need to be accessed each iteration. Given that this data structure is tiny so I doubt there will be any perf. difference.
| format = get_format(metadata) | ||
| return metadata['exportLinks'][format['type']] | ||
| format_type = get_format(metadata) | ||
| return metadata.get('exportLinks').get(format_type.get('type')) |
There was a problem hiding this comment.
What if metadata doesn't have an exportLinks?
|
I think some of the stuff here may be better passed as a query param that waterbutler passes through to gdrive transparently, this pr currently puts code that deals with exporting into mfr, I'd like to look at if theres a way to put that logic in mfr |
Ticket
https://openscience.atlassian.net/browse/SVCS-528
MFR counterpart: CenterForOpenScience/modular-file-renderer#303
Purpose
This PR replaces #290. Credit goes to @AddisonSchiller 🎆 🎆 .
Allow MFR to request
.gdocfiles in.pdfformat while keep.docxformat for downloading. This will allow MFR to cut out the expensive unoconv conversion for .gdoc filesWe can export .gdoc as a pdf if we know its coming from mfr. Added
alt_exportfunctions and properties to Gdrive utils and metadata. Gdrive download now looks for 'mfr' in its kwargs and will request file as pdf if used properly.Changes
WB V1 provider now looks for a
mfr=query param. Ifmfr=true, the google drive provider will usealt_export_nameandalt_export_link()instead of its normal counterparts. This will cause the file being sent to be a.pdfinstead of.docx.Side effects
None
QA Notes
Test that
.gdocrendering works as usual.Deployment Notes
This PR must be merged/deployed at the same time as its MFR counterpart.
This PR should be blocked by @TomBaxter 's v3 update: #276.