Fix issue 296

[khdesai]

Change invalid_types="" to invalid_types=[""]

[khdesai]

I assumed (correctly??) that for STIX 2.1, in _custom_observable_builder() props ending in *_ref(s) should be a ReferenceProperty.

I think we should remove test_custom_observable_object_invalid_valid_refs() since _valid_refs isn't really a thing anymore in STIX 2.1, due to ReferenceProperty (and the spec requirements that most (if not all) *_ref(s) properties need to have IDs as their values).

I changed all instances of parse_observable() to parse() in STIX 2.1-specific code only, except for test/v21/test_core.py's test_parse_observable_with_version() and test_parse_observable_with_no_version() since they specifically test the parse_observable function. This was done to highlight the fact that SCOs (STIX Cyber Observables) are now top-level objects in STIX 2.1.

I changed the CustomContentError (to a ParseError) in stix2/core.py's parse_observable(). For effectively the same error, parse() uses only ParseError, so I thought it fitting to make the errors consistent.

[khdesai]

This pull request also addresses issues #298 and #299, so please let me know here if those issues have been resolved.

[codecov-io]

Codecov Report

Merging #297 into stix2.1 will increase coverage by <.01%.
The diff coverage is 100%.

@@             Coverage Diff             @@
##           stix2.1     #297      +/-   ##
===========================================
+ Coverage    98.15%   98.15%   +<.01%     
===========================================
  Files          123      123              
  Lines        13845    13854       +9     
===========================================
+ Hits         13589    13598       +9     
  Misses         256      256

Impacted Files	Coverage Δ
stix2/v21/sdo.py	95.29% <ø> (ø)	⬆️
stix2/v20/sro.py	100% <ø> (ø)	⬆️
stix2/v21/sro.py	94.73% <ø> (ø)	⬆️
stix2/v21/common.py	100% <ø> (ø)	⬆️
stix2/v20/sdo.py	100% <ø> (ø)	⬆️
stix2/v21/observables.py	96.96% <ø> (ø)	⬆️
stix2/custom.py	97.29% <100%> (+0.23%)	⬆️
stix2/test/v21/test_observed_data.py	100% <100%> (ø)	⬆️
stix2/properties.py	98.37% <100%> (+0.04%)	⬆️
stix2/test/v21/test_base.py	100% <100%> (ø)	⬆️
... and 3 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d4c0115...e75d80d. Read the comment docs.

[clenk]

Looks good; rebased and merged in with 1d9ff5d.