repent / acromatic Goto Github PK

Add a button to the dictionary listing that would convert an acronym from, say, "Total Allowable Catch" to "Total allowable catch".

Logged-in users could see an "bulk add to [current dictionary]" button underneath the main text area of acronyms after uploading a document, to speed up the learning loop.

If acronyms are exactly duplicated, perhaps they should be automatically merged on batch upload, with only conflicting definitions appearing in red as now.

E.g. "Authorised Economic Operators" (AEO) is not understood.

Separate category?

e.g. S&DT TAF2+

May be able to pick up a definition in the form compared to the North Atlantic Treaty Organisation (NATO), it has not been very successful.

Managed to crash it with "PICs-" -- not sure what type of dash, I think it was a hyphen.

Error:

D, [2021-04-02T21:15:39.598782 #15535] DEBUG -- : (1.3ms) commit transaction
D, [2021-04-02T21:15:39.598926 #15535] DEBUG -- : (0.0ms) begin transaction
D, [2021-04-02T21:15:39.600109 #15535] DEBUG -- : SQL (0.2ms) UPDATE "acronyms" SET "document_id" = ?, "updated_at" = ? WHERE "acronyms"."id" = ? [["document_id", 1347], ["updated_at", "2021-04-02 21:15:39.599245"], ["id", 119326]]
D, [2021-04-02T21:15:39.601668 #15535] DEBUG -- : (1.2ms) commit transaction
D, [2021-04-02T21:15:39.602444 #15535] DEBUG -- : Acronym Load (0.1ms) SELECT "acronyms".* FROM "acronyms" WHERE "acronyms"."document_id" = ? AND "acronyms"."initialism" = ? ORDER BY "acronyms"."id" ASC LIMIT 1 [["document_id", 1347], ["initialism", "STCW"]]
D, [2021-04-02T21:15:39.603524 #15535] DEBUG -- : (0.1ms) begin transaction
D, [2021-04-02T21:15:39.604163 #15535] DEBUG -- : (0.0ms) commit transaction
D, [2021-04-02T21:15:39.604689 #15535] DEBUG -- : CACHE (0.0ms) SELECT "acronyms".* FROM "acronyms" WHERE "acronyms"."document_id" = ? AND "acronyms"."initialism" = ? ORDER BY "acronyms"."id" ASC LIMIT 1 [["document_id", 1347], ["initialism", "STCW"]]
D, [2021-04-02T21:15:39.605734 #15535] DEBUG -- : (0.1ms) begin transaction
D, [2021-04-02T21:15:39.606255 #15535] DEBUG -- : (0.0ms) commit transaction
D, [2021-04-02T21:15:39.606908 #15535] DEBUG -- : Acronym Load (0.1ms) SELECT "acronyms".* FROM "acronyms" WHERE "acronyms"."document_id" = ? AND "acronyms"."initialism" = ? ORDER BY "acronyms"."id" ASC LIMIT 1 [["document_id", 1347], ["initialism", "PICs-"]]
I, [2021-04-02T21:15:39.608110 #15535] INFO -- : Completed 500 Internal Server Error in 1492ms (ActiveRecord: 366.6ms)
F, [2021-04-02T21:15:39.609079 #15535] FATAL -- :
RuntimeError (Acronym PICs- not re-found in the text document):
app/models/document.rb:310:in get_context' app/models/document.rb:186:in block in trawl'
app/models/document.rb:100:in scan' app/models/document.rb:100:in trawl'
app/controllers/documents_controller.rb:42:in block (2 levels) in create' app/controllers/documents_controller.rb:39:in create'

Handle the case of a user wanting to create an acronym list from multiple files, or (equivalently-ish) a user wanting to add the acronyms from a new file to an existing acronym list (delete duplicates, realphebetise).

The first solution is probably easier to implement?

Currently, if the plural appears and only the singular is in the dictionary, the acronym is treated as unknown.

Probably the best resolution is a "singularize" option that only lists the singular form (stripping Ss where necessary), which probably fits with most style guides.

Add a "singularize" button to the dictionary list view, which (tries to) operate on both acronym and meaning.

Think about trying to always list the singular in the acronym list, even if it doesn't appear in the text?

A plural definition does not match a plural acronym (only a singular dictionary definition will match a plural acronym) -- this is wrong (but would be less bad with a singularize button).

App broken by CS^2.

Error: premature end of char-class: /(CS[2)/

docx2txt spits out CS^2 as CS[2] which is converted to /\CS[2)/ as a regexp, which is invalid due to the unterminated [.

PACER+, TAF2+

E.g. MOU (memorandum of understanding).

I, [2017-04-12T09:34:41.695744 #12690] INFO -- : Completed 500 Internal Server Error in 102ms (ActiveRecord: 3.0ms)
F, [2017-04-12T09:34:41.696723 #12690] FATAL -- :
ArgumentError (invalid byte sequence in UTF-8):
app/models/document.rb:40:in scan' app/models/document.rb:40:in trawl'
app/controllers/documents_controller.rb:32:in block (2 levels) in create' app/controllers/documents_controller.rb:29:in create'

Probably best to disallow internal hyphens/+s and remove the interface option.

850MB of uploads collected by 2/18, need some periodic cleanup of /public/uploads/document/file.

The quick add screen is very cumbersome to type new defs into directly.

Allow a dictionary to blacklist certain acronyms so they are ignored completedy and not added to the list

E.g. if the text features '"skills for competitiveness" (SFC)' then the definition won't be picked up.

Add a link for each acronym found to look it up in a search engine.

Quickadd makes it easy to add duplicate definitions (same or different).

Easyest fix would be to have a "kill duplicates" button, highlighting but leaving conflicts?

See Google contacts merge.

Update get_meaning() to work if incidentals are present.

Add testing for this at least!