Fix support for TIFF documents in Reader/eFolder #14193
Labels
Product: caseflow-eFolder-Express
Product: caseflow-reader
Source: Bat Team
Stakeholder: BVA
Functionality associated with the Board of Veterans' Appeals workflows/feature requests
Comments
nanotone
added
Product: caseflow-reader
Product: caseflow-eFolder-Express
Stakeholder: BVA
|
The Caseflow efolder app has a TIFF-to-PDF converter already. Maybe we just need to expose that as an API? Earlier thought by @enriquemanuel and I was making that into a lambda so it was not married to either app. |
|
That's good news, although it does raise the question why Reader was getting TIFFs when Perhaps this ticket should be reincarnated as a straight-up bug report in the efolder repo? |
nanotone
changed the title
|
I believe that's because Reader does not reader from eFolder Express. It reads directly from VBMS efolder. |
|
For the document that I was investigating in the Bat Team thread above, Reader was hitting eFolder Express's |
|
Researching related efolder INC we discovered imagemagick error example: We are wondering whether the tiff-to-pdf service is broken. Investigation in https://dsva.slack.com/archives/CAM9FJ85P/p1589485314027300 |
|
Update: yes, TIFF service in EE was broken. Still not the case that Caseflow Reader uses it, afaik. |
|
Support ticket came up about this issue again: https://dsva.slack.com/archives/CHX8FMP28/p1590085212373400 |
|
I have a sequence of actions that enable Reader to display TIFF as PDF (almost all the time) based on this code: # Refresh Reader at https://appeals.cf.ds.va.gov/reader/appeal/4025589/documents/13466417
# Get "Unable to load document" error
# In Certification console
doc=Document.find(13466417)
vbms_doc_id=doc.vbms_document_id
RequestStore.store[:application]="reader"
doc.content_url
=> "https://efolder.cf.ds.va.gov/api/v2/records/7F45E2D6-6060-46F3-AFAA-041D666694AF"
# Go to that doc.content_url in the browser, and it downloads the file as a TIFF
# doc.content_url is used by Reader's PDF.js
# In eFolder Express console
vbms_doc_id="{7F45E2D6-6060-46F3-AFAA-041D666694AF}"
record=Record.find_by(version_id: vbms_doc_id)
# check if conversion worked in the past
record.conversion_status
content=record.service.v2_fetch_document_file(record)
content=ImageConverterService.new(image: content, record: record).process
# If "conversion_success", then store file in S3 for Reader to retrieve.
S3Service.store_file(record.s3_filename, content) if record.conversion_status=="conversion_success"
# I don't know why this is necessary:
# Refresh the browser at doc.content_url; browser downloads file as a PDF
# Refresh Reader and it shows the pdfNote the download button (near the top-right corner) within Reader may still download the file as TIFF. To do a mass conversion, may want to query for record.manifest_source.records.count
record.manifest_source.records.where(mime_type: "image/tiff", conversion_status: "not_converted").count
retryRecords=record.manifest_source.records.where(mime_type: "image/tiff", conversion_status: "not_converted")
retryRecords.map{|record|
content=record.service.v2_fetch_document_file(record)
content=ImageConverterService.new(image: content, record: record).process
S3Service.store_file(record.s3_filename, content) if record.conversion_status=="conversion_success"
} |
|
@yoomlam that's good stuff. I would suggest working it into this https://github.com/department-of-veterans-affairs/appeals-deployment/issues/2718 |
|
Some more info as I'm digging into a related ticket #14298.
So if the document is not in S3 and comes from VVA, then Reader won't be able to show it. A RetrieveDocumentsForReaderJob caches documents in S3
When developing a solution, we should also consider that these S3 documents are auto-deleted after 5 days -- Slack convo. |
2 tasks
8 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Description
TIFF documents are sometimes included in the Reader documents for a case, and cannot currently be displayed by Reader's PDF viewer,
PDF.js. See technical notes for some thoughts on implementation.Background/context/resources
Bat Team thread
Related ticket on "corrupted" files in eFolder: #10504
Technical notes
There is some internet literature on getting PDF.js to display TIFF images embedded within the PDF, but displaying plain TIFF images (mimetype
image/tiff) is out of scope for PDF.js and will likely never be implemented.Rather than detecting the file type in Reader and using yet another frontend library to display TIFFs, a more straightforward approach may be to implement a PDF-to-TIFF converter in eFolder.
Because PDF-to-TIFF conversion is likely to be fraught and full of exotic edge cases, here is the output of /usr/bin/file on the header of one such problematic TIFF file seen in Reader:
TIFF image data, little-endian, direntries=19, height=3367, bps=1, compression=bi-level group 4, PhotometricIntepretation=WhiteIsZero, orientation=upper-left, width=2541The text was updated successfully, but these errors were encountered: