Download ISBN Report from Alma and create ISBN Holdings at PUL file for GOBI

[kevinreiss]

User story

As a Library selector I don't want to purchase items the Library already has in the collection

Concrete example

Acceptance criteria

• [x] Add Gobi SFTP Credentials to Princeton Ansible
• [x] I can download the items added in the last five years report from Alma (/alma/isbns/received_items_published_last_5_years_[datestamp].csv) from lib-sftp
• [x] Create the ISBN export file in the format recommended by Gobi (see Library Holdings Load documentation)
• [ ] Create a weekly task that posts a new version of the export to GOBI. See the alma-config ticket below for details on the connection. Current Alma report is scheduled to run at 6 a.m. Mondays, we can set it to be processed on Tuesday morning.

Implementation notes, if any

See Alma-config issue for more details

[maxkadel]

@mzelesky How should we treat "Begin Publication Dates" like the following?

• 202u (sometime in the 2020s)
• 9999 (unknown date?)
• publication dates in the future (e.g. 2025)

[mzelesky]

• If there is a u in the Begin Publication Date, treat it as a 0.
• If it is 9999, it should not be included
• If the date is in the future, include it

[maxkadel]

How much flexibility do we have in the report we're creating from Alma?

Right now, the ISBN column is delimited with semicolon space ('; '), but when there is only one value, it ends with a colon (:). We can deal with it if there's not an easy way to change it, but it would be lovely if it didn't have that colon.

[mzelesky]

We have very little flexibility in the report. The colon is most likely due to incorrect formatting of a cataloger's work (the colon often separates the ISBN in $a from the manifestation information in $b). Semicolons and spaces are how Alma concatenates fields.

[maxkadel]

@pdiskin is in touch with Gobi to see whether they can give us a connection via SFTP, rather than FTP.

[mzelesky]

YBP has enabled SFTP on the standard SFTP port.