Collections and temporary directory parameters

[benoit74]

Rationale

• Fix Zimit does not really support setting the --collection CLI argument #252
• Add ability to dissociate output from temporary directories

Changes

• --collection parameter is now used correctly to find WARC files
• more logs are displayed explaining the search for WARC files location
• new --build parameter (optional) to place Browsertrix files in a different directory that ZIM files
  • ZIMs are always placed in --output directory
  • Browsertrix files are placed in a "temporary" subdir of --build (if supplied) or --output by default
• display warc2zim arguments

[benoit74]

@rgaudin @kelson42 I finally slightly changed my opinion compared to Tuesday discussion, I fixed the --collection parameter logic, it was not that complex to do (and a bit sad to remove that even if not really useful) ; and it made me realized that it was not possible to set a distinct dir for WARC files location so I added this as well.

[rgaudin]

LGTM but there are some naming to fix

[benoit74]

Thank you, I also fixed a "bug" where I did not created one temporary subdir per run, which I found problematic in terms of data segregation and very/risky for the cleanup which was deleting the whole build dir while it might have contained other files before the run.

[rgaudin]

👍