andrscyv / sql-to-mongo Goto Github PK

Writer and reader builders should assume that sql and mongo configs are already validated

• user can write all cli params and sql and mongo connection params to a json file

For simplicity there is s2m.config.js in test folder, should be mocked instead

Abstract mongo driver to WriterClass that has methods:

• writer.init()
• -writer.write(collectionName, opts, rows)
• -wirter.destroy()
  and implement MongoWriter ConsoleWriter

load writer in Exporter class

Right now everything runs in serial, maybe theres oportunity for paralel procesing

To inform progress of export and the like

Handle files that are not export def

• -d --dir : path to folder with export defs
• -f --files : list of file paths of export def
• -a -after-all: path to afterAll.js file
• --dry-run: instead of writing to mongo, console.log what whould be written ( in format that could be piped to .json)
• -c --cursor-size: default 2000

Theres a lot of interfaces in that file, break it up into multple files

Hook to run arbitrary mongo commands after all exports have run succesfully

If exports specified from dir (--dir) then afterAll should be default export of afterAll.js

the path the the file can also be specifiec with -a --afterAll arg

default export should have signature : afterAll(db) , where db is mongo driver pointing to specified dbname

Hook to run arbitrary mongo queries before running exports. Mainly to clean collections if necesarry

it should be specified as AfterAll hook

Abstart sql into ExportReader class and move al opts logic to main function, there all objects should get initialized and opts shoudlnt leak into other logic

Export defs in js class should export:

• sqlQuery
• beforeEach(rows):void which transforms arbitrarily the rows