add Numpy + PIL IMAGE return type, ability to specify response mime-type

[baturayo]

1. Following return types for exposed python function will be accepted;

• PIL.Image
• numpy.ndarray
• numpy.matrix
• int8, int 16, int32, int 64
• uint8, unit16, uint32, uint64
• float16, float32, float64

2. Add the option to manually select response mime type by using # @dploy response_mime_type annotation.

3. Fix ScriptImportError & UnsupportedEntrypoint, return more meaningful errors instead of raising

dploy_kickstart.errors.UnsupportedEntrypoint: (UnsupportedEntrypoint(...), UnsupportedEntrypoint(...))

or

dploy_kickstart.errors.ScriptImportError: (ScriptImportError(...), ScriptImportError(...))

4. Add additional unit tests to increase coverage for transform.py, error.py and deps.py

[codecov]

Codecov Report

Merging #13 (a68eef2) into master (415a57b) will increase coverage by 1.46%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #13      +/-   ##
==========================================
+ Coverage   96.47%   97.94%   +1.46%     
==========================================
  Files           8        8              
  Lines         312      340      +28     
==========================================
+ Hits          301      333      +32     
+ Misses         11        7       -4     

Impacted Files	Coverage Δ
dploy_kickstart/errors.py	96.15% <ø> (+6.86%)	⬆️
dploy_kickstart/server.py	100.00% <ø> (ø)
dploy_kickstart/annotations.py	98.75% <100.00%> (ø)
dploy_kickstart/transformers.py	100.00% <100.00%> (ø)
dploy_kickstart/wrapper.py	97.14% <100.00%> (ø)
dploy_kickstart/deps.py	96.29% <0.00%> (+7.40%)	⬆️

---

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 415a57b...a68eef2. Read the comment docs.