Capture compression and open binary if present.

[delucchi-cmu]

The compression kwarg, when provided, should be used when opening the file, but should NOT be passed along to pandas.read_csv. This captures the value in a named argument so it can be used in the file opening.

Further, if some compression is provided, we change the read mode to "rb", as compression implies that the underlying file is not plain text.

Closes #407
Closes astronomy-commons/hats-import#425

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 93.02%. Comparing base (17ec4ac) to head (b3c1d38).
Report is 20 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #419      +/-   ##
==========================================
+ Coverage   92.99%   93.02%   +0.03%     
==========================================
  Files          49       47       -2     
  Lines        2127     2123       -4     
==========================================
- Hits         1978     1975       -3     
+ Misses        149      148       -1     

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

[github-actions]

Before [e47eab1]	After [9c40ded]	Ratio	Benchmark (Parameter)
12.7±0.2ms	15.0±1ms	~1.18	benchmarks.Suite.time_inner_pixel_alignment
385±2ms	389±5ms	1.01	benchmarks.Suite.time_outer_pixel_alignment
124±0.6ms	125±1ms	1.01	benchmarks.time_test_alignment_even_sky
1.04±0.02ms	1.03±0.02ms	0.99	benchmarks.time_test_cone_filter_multiple_order
46.3±0.6ms	45.3±0.4ms	0.98	benchmarks.Suite.time_pixel_tree_creation
117±0.9ms	111±0.6ms	0.95	benchmarks.Suite.time_paths_creation
97.9±0.5ms	92.1±0.6ms	0.94	benchmarks.MetadataSuite.time_load_partition_info_order7
98.4±0.6ms	92.5±0.5ms	0.94	benchmarks.MetadataSuite.time_load_partition_join_info
23.9±0.4ms	22.1±0.2ms	0.92	benchmarks.MetadataSuite.time_load_partition_info_order6

Click here to view all benchmarks.

[hombit]

Thank you! I have few questions

1. I don't really understand where decompression is happening - could you point on that code?
2. Could you please add a high-level test for this?