Friendly exceptions (#73)

* Re-merging master

Updating requirements, where possible, to match AzureNotebooks
Some fixes to comp_reqs.py to show missing packages and handle version comparison operators.

# Conflicts:
#	conda/conda-reqs.txt
#	requirements.txt

* Adding statsmodels and scipy to mypy.ini

* Fixing CSS for logon display table.

Miscellaneous linting and other fixes

* Fixing broken test

* Implementing friendly exceptions for msticpy.

* Adding unit tests for nbinit.py

* Updating version for release

* Updating version

* Bug when comparing times from a series - if min_time is created without timezone info

* Allow nbwidgets output action functions return tuples or lists

* Seems like the problem is that the data has no timezone data - redoing logic so that

it only uses the data values. This should work as long as *all* or *none* of the data sets
have timezone info.

* Prevent default output from queries (Kqlmagic post 0.1.111)

* Fixed problem with test cases not being able to do absolute imports of msticpy modules

(deleting rogue __init__.py at root of package
Updating Readme to make images clearer.
Add check_version to root __init__.py to check version against latest on PyPI
Also added check version code to tools.
Minor change to nbinit._check_config() to make testing easier.

* Tweaking images in README

* Test and linting fixes

* Error in fix fixed in kql_driver.pyl

* More test and linting fixes

* Still battling with test error in some environments

Added better exception handling in nbinit

* Removing experimental json schema files

* Making pkginstall and notebook init errors more friendly

* Fixing duplication of function from merge

Adding descriptive exception for Browshot data
Removing some deprecated terms.
Adding back test_nbinit.py

* Added User exception to Browshot

Added exception handling to ip_utils.def_whois_info
Fixed merge duplication of line in kql_driver
Fixed Output widgets where action function returns single dataframe
Fixed syslog_utils.cluster_syslog_logons_df so that it uses only source data to generate datetime objects - this should avoid TypeErrors from pandas when comparing datetimes where one dt has TZ info and the other doesn't. As long as the source is consistent this should avoid this error.

* Fixing timestamp Timezone issue in process_tree_utils and eventcluster

* bug fixes

* Excluding tests from package

* test fix

* Fixing review comments

* Test error in EventClustering notebook

* Left in an unwanted cell from testing

* Incorrect capitalization in Readme

* Fixing mypy error in wsconfig

Fixing odd test break in ti_providers (didn't have default msticpyconfig in current directory)
Removing redundant test in test_nb_tools
Fixing bug in unit_test_lib - setting ENV VAR to "" instead of removing it.

Co-authored-by: petebryan <pebryan@microsoft.com>
Co-authored-by: Pete Bryan <peter.bryan@microsoft.com>

MANIFEST.in

@@ -1,4 +1,5 @@
 include msticpy/data/queries/*.yaml
 include msticpy/msticpyconfig.yaml
 include msticpy/sectools/tld_seed.txt
-include msticpy/resources/*
+include msticpy/resources/*
+recursive-exclude tests *

conda/conda-reqs-dev-pip.txt

@@ -1,4 +1,5 @@
 flake8-mypy>=17.8.0
 mypy-extensions>=0.4.1
 prospector>=1.1.7
-pyroma>=2.5
+pyroma>=2.5
+pytest-check>=0.3.9

conda/conda-reqs-dev.txt

@@ -1,3 +1,4 @@
+aiohttp>=3.0.0
 bandit>=1.6.2
 black>=19.3b0
 coverage>=4.5.4