[Solved] How to fix broken 'Statistical prediction' (Thunderbird v91.x, Nostalgy++ v3.0.11) after migrating to new IMAP-service provider?
See original GitHub issueIs there some way I can fix/rebuild/etc my Thunderbid-Nostalgy++ Statistical prediction
predictive performance?
Symptoms
I’ve used Nostalgy/Nostalgy++ for years/decades. Its Statistical prediction
for “moving/copying emails to a different folder” has worked marvelously, until I recently upgraded to v91.4.1. (At least, that’s what I presume the Statistical prediction
feature does; if wrong, pls advise.)
Now the only folder reference(s) that come up for an email move/copy is the last folder that was moved/copied to (via a Nostalgy++ move/copy).
More details
Important note: I switched IMAP servers, porting the email-folder data from the old to the new, deleting the old-imap-server’s Thunderbird-local-my-macOS account (named Personal
), and creating a new account (also named Personal
, purposely substituting for the old account with the new server and the exact same email-folder data).
Does the above maneuver mess things up for Nostalgy/Nostalgy++? If so, is there a way to recover/rebuild/export+import the previously-working-on-previous-IMAP-server/account Nostalgy data so that the Statistical prediction
can be reused on the new account (that has the same email-folder data)?
I have years/decades of past moves/copies and therefore (theoretically) lots of stats data to draw from for Nostalgy’s Statistical prediction
stuff. I also have several clean, historical copies of my Thunderbird data profile folder (which is where I presume Nostalgy keeps it’s prior stats data) saved, including but limited to prior to the Thunderbird upgrade (which includes v91.4.1 and v68.12.1 profile-folder saves).
If needed: is it possible to rebuild the Statistical prediction
“database” from one of these past, “clean” Thunderbird profile folders?
My current versions: Thunderbird v91.4.1 Nostalgy++ v3.0.11 macOS 10.14.6, build 18G9323
Issue Analytics
- State:
- Created 2 years ago
- Comments:7
Top GitHub Comments
I fixed this. And I was able to recover and reuse all the years of historical, predictive data I had built up (in Nostalgy’s
nfpredict.sqlite
database file) saving me tremendous effort of not having to manually recreate all the predictive-move data for hundreds (maybe thousands?) of email folders. (Woo hoo!)If the following procedure/directions do not make total sense and you’d like me to clarify/fix these directions, pls signal me with a reply (on this issue) with the details of what’s not clear to you, and I’ll try to fix the text. If I can, I’ll turn this into a git-Asciidoctor-markup-pullable file so others can edit it (or just a wiki file/page someplace)… but no time to do that now.
The following procedure is more strictly-text-based and less “Github-issue-markup-text based,” because they (the procedure) was borrowed from a git commit (of my Thunderbird app-profile folder).
Anybody who wants to convert this into a proper markup (asciidoctor/markdown/etc) text file directions, preferably in a git repo, are more than welcome to do so. (I tried to use some markup-language stuff in the content to make this easier.) I’ll even review it for you. 😃
~Johnny
I have not. But it seems like a SQL-based statement (and therefore an extension-feature addition) might work just fine; I have no reason to think that it would not.