Closed peetucket closed 3 weeks ago
I like the changes you made to the logic too, it makes it clearer what is happening with updates vs adds
The unit tests for file update pass for me, but I've not seen the file update work properly in QA yet.
Initially this faillure was because the file_manifest.csv
I was using to update the Item I had created did not have a column for sdrGeneratedText
, so Preassembly was setting it to false
which caused the initial OCR XML not to get overwritten.
Even with a fix deployed to Argo to include that sdr_generated_text
column in the file_manifest.csv
the value for sdrGeneratedText
appears to be still be getting set to false
by Preassembly. Perhaps there is something missing from Preassembly to set read and use it?
I've left some notes on how to replicate the problem here: https://github.com/sul-dlss/argo/pull/4475#pullrequestreview-2111518346
The unit tests for file update pass for me, but I've not seen the file update work properly in QA yet.
Initially this faillure was because the
file_manifest.csv
I was using to update the Item I had created did not have a column forsdrGeneratedText
, so Preassembly was setting it tofalse
which caused the initial OCR XML not to get overwritten.Even with a fix deployed to Argo to include that
sdr_generated_text
column in thefile_manifest.csv
the value forsdrGeneratedText
appears to be still be getting set tofalse
by Preassembly. Perhaps there is something missing from Preassembly to set read and use it?I've left some notes on how to replicate the problem here: sul-dlss/argo#4475 (review)
The unit tests for file update pass for me, but I've not seen the file update work properly in QA yet.
Initially this faillure was because the
file_manifest.csv
I was using to update the Item I had created did not have a column forsdrGeneratedText
, so Preassembly was setting it tofalse
which caused the initial OCR XML not to get overwritten.Even with a fix deployed to Argo to include that
sdr_generated_text
column in thefile_manifest.csv
the value forsdrGeneratedText
appears to be still be getting set tofalse
by Preassembly. Perhaps there is something missing from Preassembly to set read and use it?I've left some notes on how to replicate the problem here: sul-dlss/argo#4475 (review)
I think this may fix this problem if I am understanding it correctly: https://github.com/sul-dlss/pre-assembly/pull/1478
Suggest deploying that branch of pre-assembly and trying again. I am not sure we ever thought about using file manifests in pre-assembly with OCR type content since file manifest are almost always used for media content.
I think this may fix this problem if I am understanding it correctly: sul-dlss/pre-assembly#1478
Suggest deploying that branch of pre-assembly and trying again. I am not sure we ever thought about using file manifests in pre-assembly with OCR type content since file manifest are almost always used for media content.
Yes, that fixed it!
Why was this change made? 🤔
(hopefully) fixes #1281 -- when we generate new OCR files, we need to be sure they are replaced in cocina
How was this change tested? 🤨
Updated specs
System testing: