`DELAY_TIME` was an undefined constant previously. This patch makes
delay timeout configurable using a config param.

Default of 100 seconds as API request quota limits are guaranteed to be
refilled after said time interval accoring to the documentation at:
https://cloud.google.com/compute/docs/api-rate-limits

* Load service location from ENV

* Add Dashboards and Metrics

* Fix default ES names

* Make tag names always lowercase

* Update sample_data_loader.py

* Update the dep and move connector to extra_require

* update test deps

* not use all as it is keyword

When transformer returned null no more records is processed, this means a transformer cannot filter out data. It also mean the task silently skip process rest of the records.

Use configuration parameter for Google Cloud project_id instead of hardcoded.

* Initial working version of Watermark extractor with some tests

* Implement handling for sharded tables (with YYYYMMDD suffix)

* Extract BaseBigQueryExtractor with shared code

* ISSUE-186
* prep work, moving model objects to top level

* ISSUE-186
* added csv extractor

* ISSUE-186
* minor fix, to finish the work

* * added csv extractor test
* renamed standalone column model
* fixed sample data loader and sample data

* * parameterized the sample loader connections
* fixed table owner sample data file

* * fixed linting errors

* * added some missing load calls in new data loader
* fixed table stats data problem (quoting)

postgres metadata extractor logger fixed to only log sql stmt after getting the sqlalchemy scoped config (#172)

* add cassandra extractors, tests and docs

* add cassandra extractors, tests and docs

* code refactor based on pr and code linted

In practice, when requesting logs going back more than a few days, the
StackDriver API sometimes responds with an empty page with a next page
cursor. After polling the next page a few times you eventually start
getting entries. The empty pages have no `entries` key, which previously
resulted in a KeyError.

This change allows the extractor to continue polling instead of blowing up.

* glue_extractor scope refactor to glue and filters option add

* README refactor to explain glue configuration

* Mark sort_order col so Publisher leaves unquoted.

* Add UNQUOTED for sort_order to unit tests.

* Shorter lines for the flake8.

PyCharm's highlights were very confusing.

* Fixed Snowflake views being captured as tables

Updated snowflake_metadata_extractor.py and fixed that Snowflake views were captured as tables.

* Fixed unit tests for snowflake metadata extraction

* Fix lint error

This involved adding schema_name to table_stats model as well as float and int columns to sample data

* add aws glue extractor

* add sample code for glue extractor on README

* patch whole glue_extractor test to avoid boto3 error if not configured

* boto3 add on requirements

* Prevent empty table tags appearance

* Prevent empty table tags appearance

* Add sample table tag data to quickstart

* Add sample table tag data to quickstart

* Add sample table owner data to quick start (#139)

* Update example/sample_data/sample_table_owner.csv
Co-Authored-By: jornh <jornhansen@gmail.com>

* Add sample table owner data to quick start

Will scroll past GitHub repo folders (specially more handy on small screens like phones/tablets)

* Add sample high/low watermark data to quick start (#62)

* Update setup.py

* Relax dependency to make it range

* Update requirements.txt

* set field_size_limit to max

* fix lint

* increase version

* update

Removed `SQUASH COMMIT` from the PR template

* Make some unittest work with py3

* Make some unit tests work with py3

* Make snowflake metadata extractor py3 test green

* feat(user): add Neo4j to ES quickstart sample job

* lint

* Fix (user): Whitespace glitch giving quotes in UI

* Correct ES mapping, thanks Tao!

* lint

* Trailing Linefeed

Check for a valid string on the regex transformer.  If not a valid string, then the value is ignored and the original record returned. (#136)