analyze: changed chunks per directory

[ThomasWaldmann]

Fixes #71.

• relies on the usual archive selection via -a pattern to select archives that belong together / that should be considered for analysis.
• goes over all selected archives, over all contained files and collects informations about chunks stored in all directories it encountered.
• considers chunk IDs and their plaintext sizes (we don't have the compressed size in the repo easily available).
• adds up added/removed chunks' sizes per direct parent directory (not per filename, that would consume lots of memory and also maybe is not really necessary).
• outputs a list of "directory: size" in reverse order by size.

[codecov]

Codecov Report

Attention: Patch coverage is 95.23810% with 4 lines in your changes missing coverage. Please review.

Project coverage is 81.63%. Comparing base (2e37b1a) to head (7533f23).
Report is 3 commits behind head on master.

Files with missing lines	Patch %	Lines
src/borg/archiver/analyze_cmd.py	95.12%	2 Missing and 2 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #8436      +/-   ##
==========================================
+ Coverage   81.52%   81.63%   +0.10%     
==========================================
  Files          70       71       +1     
  Lines       12797    12883      +86     
  Branches     2325     2336      +11     
==========================================
+ Hits        10433    10517      +84     
  Misses       1705     1705              
- Partials      659      661       +2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[ThomasWaldmann]

Hmm, unsure about counting added AND removed chunks:

• it might be a good score for the activity inside a directory
• otoh, people are usually rather interested in what's eating their repo space and for that, only added chunks are interesting.